Here is the main position:

Here's yesterday's blotter:

Four fills on September 2: 40,000 shares of AIDR between $0.5700 and $0.5800. About $23,350 all-in including commission, or roughly $0.584 per share landed.

Where the dry powder comes from

I’ve left the sells in the same screenshot on purpose, because “where does this guy keep finding money to buy more” is a reasonable question and I’d rather answer it before someone has to ask.

That day, I sold 2,000 Berkshire CDRs at $36.57 and 1,500 Cardinal Energy at $12.13, about $91,000 in combined proceeds. That funded the AIDR buy, a 25,000-share position in Silver X Mining at $0.90, and left the balance in cash. No leverage, no margin, no outside money. I rotated out of two positions I’d held for a while and into two I’d rather own right now.

Worth noting before anyone else does: the Silver X buy was slightly larger in dollar terms than the AIDR buy that day. I’m not going to hide that to make the conviction story cleaner. AIDR remains the higher-weight position in my portfolio by a wide margin; a single day’s allocation isn’t the whole picture.

Why now

Nothing has changed in the past week that isn’t already public. The thesis is the same one I’ve been writing: Rocket Doctor has built payer access, roughly 21 million U.S. covered lives across about 20 insurers per the filed MD&A, well ahead of the revenue it currently recognizes against that footprint. The binding constraint is physician credentialing throughput, not patient demand.

I just covered Rocket Doctor (OTC: AIRDF): Q2 Is the Quarter the Skyscraper Came Above Ground here:

If that’s right, the market is still pricing the “before” picture while the structural work is largely done. If it’s wrong, credentialing conversion stays slow, U.S. revenue stays a small share of the total, and I’m early on something that doesn’t arrive. That’s the actual risk, and I’ve written about it before.

I’d rather add at $0.58 than write about it and do nothing.

One caveat on posting fills

I’ve historically disclosed my position as a share count and a portfolio weight rather than posting trade-by-trade screenshots, for a reason: showing buys and not showing sells is how people get misled. So the standard I’m setting here is symmetrical: if I ever materially reduce this position, I’ll say so with the same specificity.

This isn’t a trade alert service, and I’m not going to post every fill. I’m posting these because a lot of people have asked whether I’m actually in this, and screenshots answer that faster than another thousand words of thesis.

Disclosure: I am a long AIDR holder of over 1.5M shares, including the 40,000 shares shown above. I have no relationship with the company and am not compensated for writing about it. I also hold a position in Silver X Mining as shown. Prices are in CAD. Nothing here is investment advice. NFA. DYOR.

Operating figures attributed to Cherniak below are unaudited management commentary; the Q2 2026 MD&A has not yet been filed. Where I cite filed disclosure, I say so.

I’ve argued for a while that this market is pricing the before picture: Rocket Doctor spent two to three years building payer infrastructure the tape gave it zero credit for, and the gap between ~21 million covered lives and the revenue actually recognized against them was the whole opportunity.

Cherniak’s Q2 update is the first quarter where you can watch that gap start to close in real time. Here’s what stood out.

1. The supply side: 80+ physicians, and they’re paying to get in

Rocket Doctor has scaled to over 80 signed physicians across the United States (be mindful not all have been credentilized yet, therefore not all are seeing patints and/or generting revenue more that below).

The number I care about more is US$500, the upfront fee each physician pays Rocket Doctor to process their credentialing with insurance networks. That’s the tell. Doctors paying out of pocket to join a platform is a fundamentally different signal than doctors signing a free-tier agreement, because it filters the merely curious out of the count. And the company got there without expensive recruiting firms. Physicians are coming organically, looking for autonomy over their schedules and the ability to capture reimbursement rates they actually generate rather than handing them to a corporate employer.

Credentialing takes three to six months. That’s the constraint I’ve called the binding one on this business from the beginning not patient demand, not payer access, credentialing throughput. Which means a meaningful slice of those 80+ physicians is dry powder: capacity already signed and paid for that structurally unlocks over the coming quarters without any new sales effort.

That’s the part the market isn’t modelling. The doctors are already in the pipeline. The revenue from them just hasn’t arrived yet.

2. The demand side: ~4,000 visits, 200% growth

Roughly 4,000 U.S. patient visits in Q2, up about 200% over Q1.

For context on the slope: U.S. operations were essentially at zero last December. The filed Q1 MD&A disclosed about 1,200 completed U.S. visits in April 2026, up 69% month over month. Getting to 4,000 for the quarter off that base, while credentialing was still the gate on supply, suggests demand isn’t the problem. Patient routing and the booking experience are working.

Worth holding in mind for how you read the next print: under the company’s stated U.S. policy, revenue is recognized on actual cash receipt, not on service delivery. Medicare and Medicaid claims settle 45 to 90 days out. So most of Q2’s visit volume lands in Q3 and Q4 revenue, not Q2’s. I’ve said before I read that as accounting conservatism rather than a warning sign, but it does mean the reported revenue line will always trail the operating reality by a quarter or so. If you’re waiting for the visit curve and the revenue curve to move together, you’ll be early to worry and late to act.

3. CAC is falling, and the margin mix is telling you the U.S. business is real

Cherniak noted that customer acquisition cost has dropped considerably in every state they operate in, even through the 200% volume surge. CAC isn’t a disclosed metric, so I can’t independently check it, but it’s consistent with what you’d expect once brand recognition and payer in-network status start doing the work paid acquisition used to do.

Management is also being disciplined about how it spends. Rather than burning cash on conventional ad buys, the company secured a high-visibility partnership with Rick Ware Racing across NASCAR, motocross, and hot rods, with the Naval Base Coronado race described as the first of 50 planned motorsports activations, paid in equity rather than cash. That preserves the balance sheet at a moment when balance sheet matters. It’s worth being clear-eyed that equity isn’t free; it’s dilution with a deferred settlement date, landing on a structure that already carries ~7.4 million warrants at C$0.85 expiring January 22, 2027. But trading paper for national brand exposure while cash is the scarcer input is a defensible call.

Now the part I want to flag because I think it’s misread as bad news when it’s actually confirmation:

Filed gross margin has compressed for four straight quarters, 89.2% in Q2 2025, 88.0% in Q3, 83.7% in Q4, and 75.0% in Q1 2026. Direct costs more than tripled year over year, while revenue grew roughly 44%.

That is exactly what should happen. The old revenue base was high-margin Canadian SaaS. U.S. patient visits carry real delivery cost. Margin compression is the fingerprint of the U.S. business actually mixing into the numbers, it’s the first hard evidence in the filed record that the thing I’ve been arguing would happen is happening. A company that stays at 89% is a company whose U.S. visits aren’t scaling.

The honest implication: any model holding gross margin at 88% through 2035 while projecting U.S.-dominated revenue growth is internally inconsistent, and Maxim’s does. The filed numbers are already thirteen points below that assumption. I’d rather own that now and be right about the direction than defend a number the company itself is moving away from. Lower margin on a much larger revenue base is the trade you want here.

4. Two agreements that open genuinely new service lines

The California IPA agreement (announced June 30, 2026, effective June 1). Rocket Doctor’s first value-based primary care provider agreement, alongside a specialty services agreement, with a California independent physician association. Physicians get assigned patient panels under a population-based reimbursement structure ongoing primary care and population health management rather than episodic urgent care. The specialty services piece carries in-network access to over 5 million patients across 9 payers and 65 insurance products.

Cherniak has said this work commands multiples of standard urgent care reimbursement. That’s management commentary rather than filed disclosure, and I’d note that population-based reimbursement pays for a panel rather than an encounter so it changes the unit economics in ways a simple per-visit multiple doesn’t capture. But the strategic point holds: this moves Rocket Doctor up the value chain from transactional care to continuity of care, where durable payer relationships live.

The California network agreement (announced August 18, 2026, effective July 1). One relationship, five healthcare and insurance channels, including workers’ compensation and auto medical, which Rocket Doctor had no exposure to before. The counterparty connects into a nationwide network spanning over 700 health plans, more than 100,000 employers, and close to 60 million consumers.

To be precise, because precision here is the whole basis of anyone taking my numbers seriously: that 60 million describes the network’s national footprint, and Rocket Doctor’s access under this agreement is California-scoped. Cherniak’s own framing “as we continue to scale in California” says so. The counterparty is a network lessor, not a direct payer like Aetna or Anthem. What’s genuinely additive is the channel access: workers’ comp and auto medical run on different reimbursement dynamics and different claims cycles than Medicaid managed care, and getting into both through a single relationship is efficient TAM expansion.

Filed disclosure supports approximately 21 million covered lives, California ~8.1M, New York ~9.9M, Maryland ~3.1M. Press releases since August 11 cite approximately 24 million following the New York expansion. I use the filed number until the next MD&A confirms the higher one.

The bottom line: the walls are going up

Cherniak’s analogy for where the company sits is right. For two to three years, Rocket Doctor was digging the hole and pouring the foundation. From the outside and to the market it looked like nothing was happening. That’s the phase the share price is still pricing.

The walls are now going above ground. The infrastructure is built, the payer contracts are signed, the physician pipeline is filling and paying to be there, and the marketplace is matching patients to doctors at a rate that was zero nine months ago. As credentialed physicians go live through Q3 and Q4 and the cash-basis recognition lag catches up, the revenue scaling should become visible in the filed numbers rather than only in the operating commentary.

What I’m watching to confirm it:

1. U.S. revenue as a percentage of total revenue. This is the thesis validation metric; everything else leads into it.

2. Credentialing conversion: how many of the 80+ reach billable status, and how fast.

3. Visits per credentialed physician per month. Signing doctors is the easy part; utilization is the business.

4. Gross margin finding a floor. Compression is expected. Where it stabilizes determines the modelis value.

5. Cash. C$3.50 million at March 31, 2026 against C$3.09 million used in operations that quarter, with financing covered by a C$4.96 million placement and C$1.23 million in warrant exercises. Capital markets dependency is a live risk regardless of how the operating metrics trend, and the escrow release calendar runs through August 2027.

None of that changes the thesis. It’s the scoreboard I’ll grade it against.

I’m long, and Q2 made me more comfortable being long, not less.

Citation: Cherniak, Dr. Bill (Co-Founder & CEO, Rocket Doctor Inc.). Q2 2026 corporate update interview. [HOST / CHANNEL], [DATE].

Additional sources: Rocket Doctor AI Inc. Condensed Interim MD&A and Interim Financial Statements for the three months ended March 31, 2026 (SEDAR+, filed June 1, 2026); GlobeNewswire releases dated June 30, August 11, and August 18, 2026; Maxim Group initiation of coverage (Klee, Greenberg); Fundamental Research Corp update (Rajeev).

Disclosure: I hold over 1.5 million shares of Rocket Doctor AI Inc. (CSE: AIDR | OTC: AIRDF | FSE: 939), representing a meaningful portion of my equity portfolio. I have not been compensated by the company or any affiliate for this piece. NFA. DYOR.

Stocks covered:

In clean technology, the walk from a working pilot to a commercial facility is where capital goes to die. For a decade, equity investors have watched plastic pyrolysis companies present immaculate bench data, then hit reality at industrial scale: yields collapse, capex doubles, reactors coke up, and the equity gets recapitalized at a fraction of the last round.

That history is real, and the scar tissue is earned. It has produced a default assumption about Aduro Clean Technologies: scale-up is the existential risk, and the company deserves the same discount as everyone else who has stood at this particular threshold.

I think the market has the risk shape wrong. Not its existence. Its shape.

The specific failure mode that killed legacy pyrolysis at scale is structurally absent from Aduro’s process. What remains is conventional project risk: permitting, schedule, capital deployment, integration risk, but fundamentally different to underwrite. And on August 19, Aduro took the single largest step toward retiring the portion of that residual risk that actually sits on its own balance sheet.

Let me walk through why, and then through what would prove me wrong.

1. The physics: why thermal uniformity is the thing that breaks

Start with the mechanism of failure rather than the fact of it.

Pyrolysis reactors are oxygen-free; that’s the whole point of the process. Heat is applied externally, through the reactor wall or via a circulating heat carrier, into a molten polymer mass that happens to be an excellent thermal insulator. That is the problem. As you scale the vessel, the surface-area-to-volume ratio falls, and the distance heat must travel through a viscous, poorly conducting melt grows. You get thermal gradients. Material adjacent to the hot wall over-cracks and forms char and coke; material in the interior is under-converted. Product slate drifts, run length degrades, and the oil you get out is a variable soup requiring expensive downstream cleanup.

Mariusz Skonieczny, an independent investor and content creator who, like me, holds a position in the company, and whose work I’d treat as independent research rather than company disclosure, framed this well in a recent video, and I’m going to borrow his analogy because it’s the clearest version I’ve seen.

“Aduro’s plastic recycling technology is like cooking pasta in water... It will cook evenly in a small pot, a medium pot, a big pot, or many pots. Modularity means it will cook evenly.”

— Mariusz Skonieczny, MicroCap Explosions [3]

Baking a pizza depends on gradients. Boiling pasta doesn’t. A liquid medium at temperature transfers heat by convection and carries essentially no gradient; the reaction environment at the wall is the same as the environment at the centre.

Aduro’s HCT runs in liquid water, sub-supercritical, in the roughly 240–390°C band, with a proprietary catalyst and an in-situ hydrogen donor. The water is not incidental; it is the thermal medium, and it is the reason the reaction environment is uniform in a way a pyrolysis melt cannot be.

Here is the narrow claim I’ll defend: the dominant historical failure mode of scaled chemical recycling thermal non-uniformity and its downstream consequences does not apply to this process. The chemistry that works at 10 kg/hr is the chemistry that works at 1,000 kg/hr, because the variable that degraded with scale in the old architecture doesn’t degrade here.

Here is the claim I will not make: that scale-invariant chemistry gives you a scale-invariant plant. It does not. Mixing regime, residence time distribution, catalyst recovery and makeup, solids handling, corrosion behaviour under pressure at temperature, heat integration economics- none of that is settled by thermal uniformity. Those are engineering problems, and they remain.

Which is the entire point. Chemical risk and engineering risk are not the same asset class, and the market is pricing this name as though they were.

2. The Lego-block argument, and its one honest gap

A recurring error in capital markets is equating novel chemistry with novel engineering. Aduro’s chemistry is proprietary. Its plant is deliberately not.

Eric Appelman, Aduro’s CRO, has made this point directly:

“We have been able to assemble a process essentially from Lego blocks that have been around for so many years... We do not really need to build a 10,000-ton pilot plant because every builder knows perfectly well how you build those things bigger without running into trouble.”

— Eric Appelman, CRO of Aduro [1]

Sourcing note: this and the pilot-sizing rationale in Section 3 are management commentary from podcast interviews, not filings. I’m carrying them as such. They’re consistent with what the company has disclosed publicly, but they are not independently verifiable, and you should weight them accordingly.

The substance holds up. Extruders, pumps, heat exchangers, distillation columns, CHP units these are catalogue items with published scaling relationships, operating at ten to a hundred times Aduro’s FOAK capacity across the global petrochemical industry today. Nobody has to guess how a distillation column behaves at 10,000 tonnes a year. That collapses a large part of the FOAK premium: no single-source custom fabrication, no bespoke supply chain, no engineering guesswork on the balance of plant.

But the reactor is not a Lego block. It is the one genuinely novel piece of hardware in an otherwise conventional plant, and it is exactly where the residual risk from Section 1 lives. Anyone telling you the standardization argument covers the whole facility is overselling it.

So the real question isn’t whether Aduro can buy standard equipment. It’s whether anyone credible will take the novel unit and integrate it into a standard plant to a bankable cost estimate.

That’s the question that got answered this month.

3. The pilot size: pragmatism, not evasion

Before I get to Saipem, the bear objection that comes up most often.

The NGP pilot in London, Ontario runs at 10 kg/hr, roughly 240 kg/day. The FOAK at Chemelot is designed for about 1,000 kg/hr, or ~10,000 tonnes a year. Critics reasonably ask why the company didn’t build an intermediate 100 kg/hr unit to bridge the gap, and read its absence as avoidance.

Management’s answer, per interviews, is unglamorous: 240 kg/day was the largest reactor that would safely fit in the existing lab. Building bigger meant a new site, new permits, and a 12-to-18-month delay to generate data that standard scaling relationships already provide. They chose speed.

I find that credible precisely because of Section 2. If the balance of plant is catalogue equipment and the reactor chemistry is scale-invariant, an intermediate unit burns capital and calendar to produce very little novel engineering information. If either premise is wrong, the intermediate unit was the right call and skipping it was a mistake. The argument is coherent, but it is load-bearing; it depends entirely on the two claims above holding.

4. Saipem: the residual risk gets a counterparty

On August 19, 2026, Aduro announced the selection of Saipem S.p.A. for Early Works and Services supporting the FOAK facility at Chemelot. Early works are funded from existing cash, and the company expects no additional financing. The scope covers Process Design Package review, optimization of critical equipment packages, preliminary utility integration and capital cost refinement, under a structured stage-gated approach that opens into FEED, detailed engineering, procurement, construction and start-up.

Be precise about the instrument. Aduro’s own forward-looking language describes this as a Conditional Letter of Award. It is an entry point into a staged process, not an unconditional lump-sum EPC contract, and it does not commit Saipem to build anything. Treat it as what it is.

With that said, I think this is the most under-read item of Aduro’s year, and it lands directly on the gap I identified in Section 2.

Saipem is a tier-one industrial contractor. Its core competency is integrating process technology into executable, cost-bounded industrial facilities — including novel units, which is a substantial part of what they get paid for. When you hand a firm like that your Process Design Package and ask them to tear into your capital cost estimate, you are doing three specific things:

You are importing execution capability you do not possess. Aduro is a chemistry company with a pilot plant and roughly a C$10 million annual burn. It does not have in-house industrial project delivery. Saipem does. The engineering competence required for the build now sits with an organization that has done this repeatedly at far larger scale.

You are subjecting the capex number to adversarial review. This is the part I’d emphasize most. FOAK cost overruns are the single most common way clean-tech equity gets destroyed, and they usually originate in an internally generated estimate that never met a contractor’s pricing discipline. A capital cost refinement performed by the firm that would execute the work is a fundamentally different number from one built by a management team that wants the project approved.

What this does not do: it does not eliminate engineering risk. It transfers the residual to a counterparty with a track record of managing it, and it puts that residual under external scrutiny before committing capital. That’s the honest framing, and it’s still a material re-rating of the risk profile, but “de-risked” and “solved” are different words, and only one of them is defensible.

One open question worth flagging: Aduro has not disclosed whether Saipem is the same “leading global EPC firm” behind the March 2026 non-binding licensing-package MOU. The company has kept that counterparty unnamed. I’m not going to assume they’re the same entity, and neither should you.

5. What the NGP actually proved and what it didn’t

Two things get conflated constantly here, including by people who are long. They are separate claims from separate disclosures, and they deserve separate treatment.

The yield claim. On June 9, 2026, Aduro reported that the NGP pilot ran 47 continuous hours on recovered polypropylene at 86% liquid hydrocarbon yield, with roughly 35 hours at steady state. That’s the first public empirical validation of the yield claim under sustained continuous operation rather than batch conditions. It matters less as a number than as a category change: the machine turns on, stays on, and does what the deck said. Note the feed: this campaign was pure polypropylene. Mixed and contaminated feedstock campaigns are ongoing, and their results are not yet public.

The cracker-compatibility claim. This is a different disclosure entirely: the November 20, 2025 press release covering pilot-scale steam cracking trials at a European facility, conducted with a global organization that designs, licenses and services large-scale steam cracking operations. HCT oil was processed as produced, without dilution or pre-treatment, on a mixed waste plastic feed. Ethylene and propylene yields were reported as comparable to fossil naphtha, with a substantially lower boiling range than comparable chemical recycling oils.

That claim drives the economics, because pyrolysis oil is olefinic and cannot enter a cracker without hydrotreatment—a step that peer-reviewed work puts at roughly $400 to $1,200 per tonne, depending on geography. Removing it drops the minimum economic plant scale from the 100,000–200,000 tonne threshold pyrolysis requires to roughly 25,000 tonnes, which is what makes modular deployment work at all.

Carry Aduro’s own hedging with it. The company’s language is “potential” and “may be used,” and it refers to this particular Aduro product. One batch, one feed, one unnamed counterparty. It is a genuine validation milestone. It is not a finished commercial specification, and anyone citing it as one is going beyond what the company said.

CEO Ofer Vicus has been direct that the pilot-to-industrial jump is the largest physical step in the company’s life, and that the FOAK does not need to be the optimum on day one; the goal is steady-state continuous commercial operation, with refinement coming in units two and three. I think that’s the right posture, and it’s more honest than most of what this sector puts in front of retail.

6. What would prove me wrong

I’ve argued that the market is mispricing the shape of Aduro’s execution risk. That is not a claim that the risk is small. Here is what I’m actually watching, and what would break this thesis.

Permitting at Chemelot. Management has itself identified permitting as the most significant gating risk to the FOAK schedule. Nothing in this article addresses it. Standardized equipment and uniform thermodynamics do not move a Dutch environmental permit. The Ebert HERA engagement is the vehicle; visible permitting progress is the first concrete de-risking event, and its absence through year-end would be a genuine problem.

Schedule divergence. The company targets mechanical completion in 2H27 with commissioning in 2028. Sell-side models carry that later; Ladenburg’s runs to late fiscal 2028 with first revenue in fiscal 2029. Analysts are embedding a six-to-twelve-month slip multiplier, and they are usually right to. If you’re underwriting the company timeline without that haircut, you’re underwriting the optimistic case.

The instruments are still conditional. The Saipem award is a Conditional Letter of Award. The offtake is a non-binding LOI. The licensing arrangement is a non-binding MOU. The AstroTurf and Ortessa arrangements are MOUs at the evaluation stage. The thesis is conversion of that funnel into executed contracts, and none of it has converted yet. FEED authorization is a clean, falsifiable marker. If FEED hasn’t been authorized by mid-2027, the schedule has slipped materially regardless of what anyone says on a podcast.

The reactor could still surprise. I’ve argued thermal uniformity travels with scale. Catalyst performance over extended runs, fouling behaviour on contaminated feed, and materials durability at pressure and temperature remain open questions that only sustained operation can answer. The mixed-feedstock NGP campaigns are where this gets tested. Poor results there would directly damage the Section 1 argument.

And the standing bear case hasn’t changed: scale-up failure at FOAK, no commercial license signed by end of 2027, or repeated dilutive raises at falling prices. Any of those and I’m wrong. This remains a pre-revenue microcap whose valuation depends on belief about 2030. Position size accordingly.

The verdict

Scaling custom machinery running on chaotic thermodynamics is a monumental, often insurmountable risk. That risk killed the pyrolysis cohort, and the market is still pricing it into this name.

Scaling standard chemical engineering components around a controlled, water-based reaction with a tier-one contractor inside the data room reviewing the process package and the capex estimate is a different problem. It is a project execution and capital deployment problem. It can absolutely still go wrong. Permits get delayed, schedules slip, and contractors find the number was too low. But it is the kind of problem thousands of industrial facilities solve every year, and it shouldn't carry the same discount as an unsolved chemistry problem.

That distinction is the whole thesis. In March, you were underwriting chemistry. Today you are underwriting a project schedule. The Saipem mandate is what marks the transition, and I don’t think the market has repriced for it.

Sources

Tier 1 — Company disclosure and filings

• Aduro Clean Technologies, “Aduro Selects Saipem for Engineering and Procurement Support for First-of-a-Kind HCT Facility at Chemelot,” GlobeNewswire, August 19, 2026

• Aduro Clean Technologies, pilot-scale steam cracking of plastic-derived Hydrochemolytic oil, GlobeNewswire, November 20, 2025

• Aduro Clean Technologies, NGP pilot continuous operating campaign results, June 9, 2026

• Aduro Clean Technologies Corporate Presentation, March 2026

Tier 2 — Independent research and sell-side

• Roth Capital Partners, Initiating Coverage, Buy, July 27, 2026

• Ladenburg Thalmann (Jon Hickman), coverage updates 2026

Tier 3 — Interviews and community commentary (management commentary; hedge accordingly)

• [1] How to Commercialize Technology with Eric Appelman of Aduro, The Chemical Show:

• [2] Aduro CEO Ofer Vicus: The 86% Yield Run, Stock Therapy with Penny Queen:

• [3] Is Scaling Aduro’s Biggest Risk?, MicroCap Explosions (Mariusz Skonieczny independent investor, holds a position):

• [4] Why is Aduro’s pilot plant so small? Will their tech actually scale?, Amateur Investing

This is not financial advice. I own shares of $ADUR and it represents ~30% of my portfolio, so I am biased. Long $ADUR | Not financial advice | DYOR.

There’s a version of the Rocket Doctor story told in headline numbers: covered lives, payer logos, total addressable market. I’ve written that version. I’ve also spent enough time in the filings to know that the headline numbers are the least interesting part of this company right now.

The interesting part is an accounting mechanic that will make the financial statements look worse every quarter for the next several years, in a way that is actually the clearest available evidence the thesis is working.

Amateur Investing put out a new video this week that walks through it. I want to credit it properly, walk through what it gets right, correct three figures it gets wrong, and add one thing I think it misses entirely, something I only found by going back to the Q1 2026 statements to check its arithmetic.

Source video: Amateur Investing — Rocket Doctor AI financial deep dive

A note on how I label things, because it matters for a company at this stage: I separate filed disclosures (SEDAR+, MD&A, financial statements) from management commentary, analyst coverage, and third-party analysis. Those are four different tiers of reliability, and I’ll tell you which one I’m standing on throughout.

The setup: Shopify for doctors, pointed at the US

The video frames Rocket Doctor as a one-stop digital practice platform; doctors set up their own virtual clinics on it, the way merchants set up stores on Shopify. That’s the company’s own positioning, and I think it’s a fair shorthand.

What matters for the numbers is the geographic pivot. Canada was historically the business. The US is where it’s going. The video cites a virtual care market projected at roughly $50 billion by 2034.

Take that market-size number the way you should take all market-size numbers, including the ones in the company’s own deck; the deck itself carries a disclaimer that third-party market data hasn’t been independently verified. TAM figures are directional, not analytical. The useful question is never how big the pool is; it’s what share you can actually capture and at what unit economics.

Which brings us to the actual point.

The accounting mechanic (this is the whole article)

Here is the distinction that will cause people to misread these financials.

In Canada: the provincial insurer pays the doctor. The doctor pays Rocket Doctor a fee. Per the company’s July 2026 investor deck, that’s an average of 17% per Canadian appointment. Rocket Doctor books only the fee as revenue. Almost no cost of revenue sits against it. Result: gross margins in the high 80s.

In the US: the insurer pays Rocket Doctor the full visit amount. Rocket Doctor keeps its cut and passes the rest to the physician. But the full amount runs through the top line as revenue, and the physician payout sits in direct costs. Result: the same economic transaction produces a dramatically lower gross margin percentage, while producing far more gross profit dollars.

The video uses illustrative numbers: a $100 visit, roughly $25 retained, roughly $75 to the doctor. The company’s deck states a US$25 flat fee per US appointment, which is consistent with the retained figure.

So as the US mix grows, blended gross margin percentage mathematically falls. Every quarter. And that decline signals the thesis is working, not deterioration.

That’s the video’s central claim. I wanted to know whether it was a nice story or something visible in filed numbers.

I checked it. It’s already happening.

From the filed Q1 2026 interim statements (three months ended March 31, 2026, all CAD):

Now compare that to the trailing quarters. Q2 2025: 89%. Q3 2025: 88%. Q4 2025: 87%.

Q1 2026 came in at 75%. That’s a thirteen-point compression in a single quarter.

And here’s the part I find genuinely compelling: both analysts covering this name modelled 88% for that quarter, and both were wrong. FRC’s report noted 87% actual versus 88% forecast for Q4 and carried the assumption forward. Maxim’s initiation model has 1Q26E at 88% gross margin. The actual print was 75%.

Two independent analysts missed the same line by the same mechanism, in the same direction, in the same quarter that US revenue first became material at roughly 24% of the total. That is not noise.

Let me show my work on what the US segment margin actually implies and flag clearly that this next step is my estimate, not a filed figure. The company reports a single consolidated segment and does not break out direct costs by geography.

Canadian direct costs have historically run around 11–12% of Canadian revenue. Applying ~11.6% to Q1 2026 Canadian revenue of $557,753 gives roughly $64,700 of Canadian direct cost. That leaves roughly $119,800 of direct cost against $179,350 of US revenue, implying a US gross margin near 33%.

I stress-tested the assumption, because a single input driving a conclusion makes me nervous. Even if you assume Canadian direct costs are as high as 15% of Canadian revenue, the implied US gross margin only rises to around 44%. Either way, US gross margin is nowhere near the 88% blended rate the analysts are modelling. The conclusion is robust to the assumption.

Honest caveat: I can’t attribute the full 13-point drop to mix shift alone. The Alea Health acquisition closed in January 2026, and other cost movements may be driving it. I’m confident about the direction and the mechanism. I’m not going to pretend I can decompose it precisely from outside the company.

Unit economics: useful, but understand whose numbers these are

The video estimates the ultimate US customer acquisition cost at $15–20, gross profit around $22 per visit, and therefore payback on the first visit. It estimates lifetime value to CAC at roughly 5–6x, driven by targeting chronic care patients who return.

It also argues a US visit is roughly 6x more profitable than a Canadian one. Sanity-checking that: 17% of a Canadian visit versus a US$25 flat fee lands somewhere in the 5–6x range, depending on the Canadian per-visit billing amount and the exchange rate.

But be clear about the evidence tier here. CAC, LTV, payback period, and the 6x multiple are not disclosed in any filing I’ve read. They are the creator’s estimates, built on reasonable logic and stated assumptions. They’re worth engaging with. They are not company-reported metrics and shouldn’t be repeated as if they were.

I’ve said before that the absence of disclosed conversion and retention metrics is a legitimate weakness in this story. It still is. Estimating them carefully is a reasonable response to that gap; it isn’t a substitute for the company closing it.

The cash-basis lag

One thing worth layering on that the video doesn’t dwell on, from the filed MD&A: for US operations, Rocket Doctor recognizes revenue on a cash basis, only upon actual cash receipt, because collectability from Medicare, Medicaid and private payers can’t be established with enough certainty to satisfy IFRS 15 at the time of service.

Practically: US visits show up in the financials 45–90 days after they happen. The MD&A also confirms the company permits out-of-network patients onto the platform to drive adoption, submits those claims anyway, and recognizes anything recovered only on receipt.

This is accounting conservatism, not a red flag. But it means every quarter you read, it understates the operating reality at the time you’re reading it. Combine that with margin compression, and you get financials that structurally lag and structurally look worse right as the business is inflecting.

Path to profitability

The video estimates cash-flow neutrality around 2028 or 2029, with the company deliberately sacrificing near-term profitability to buy market share.

For comparison, the published analyst models show Maxim projecting adjusted-EBITDA-positive by the end of 2027. FRC values the company assuming it captures 1.5% of its Canadian target market and 3% of its US target market by 2032.

The video’s estimate is more conservative than Maxim’s. I’d rather underwrite the conservative one.

What I’m actually watching

Not the margin percentage. That’s the point of this entire piece; it’s going down, and I want it to.

What I’m watching:

1. US revenue as a percentage of total revenue. It was 24% in Q1 2026. This is the single cleanest filed metric for the thesis.

2. Blended gross margin as a confirming indicator. If US revenue climbs and gross margin doesn’t compress, something in my understanding of the recognition treatment is wrong, and I want to know that immediately.

3. Physician credentialing throughput. I’ve said this repeatedly, and it hasn’t changed: the binding constraint here is credentialed physician supply, not patient demand. Everything else is downstream of it.

4. Whether the company starts disclosing conversion and retention metrics. Until it does, the unit economics discussion- mine, the videos’, anyone’s is estimation.

The risks I’m not going to pretend away

Dilution is real and ongoing; both FRC’s fair value cuts were driven primarily by it. The company is unprofitable with a limited operating history. Two of the three published research sources on this name have disclosed financial relationships with it. Segment reporting is consolidated, which means the US margin figures in this article are my inference rather than disclosed fact. And the cash-basis lag that protects against overstatement also means you’re always reading stale operating data.

I own a lot of this stock. That is a reason to read me critically, not a reason to read me.

But I’d rather hold a company whose margin percentage falls because it’s capturing six times the profit per transaction than one whose margin percentage stays beautiful because it never left home.

Disclosure: I hold over 1.5 million shares of Rocket Doctor AI Inc. (CSE: AIDR | OTC: AIRDF | FSE: 939), accumulated over time at an average cost basis I’ve disclosed previously. This position represents a significant weight in my portfolio. I have no compensated relationship with the company, with Amateur Investing, or with either research firm cited. All filed figures are drawn from the Q1 2026 interim financial statements and MD&A available on SEDAR+; all estimates are labelled as such.

$AIDR $AIRDF — NFA. DYOR.

Camp one: a sub-$50M micro-cap is buying NASCAR sponsorship instead of hiring doctors. Camp two: national TV exposure, millions of eyeballs, this is the catalyst.

Neither camp read the agreement. I did. The interesting part isn’t the racing; it’s the consideration structure, and it deserves more scrutiny than either the bulls or the bears have given it.

What was actually signed

Per the June 8, 2026 press release, the definitive Marketing Agreement is dated effective June 5, 2026, between Rocket Doctor AI, Rick Ware Racing, LLC, and FINTEKK AP, LLC.

The terms, stated plainly:

Phase one (June 6 – September 30, 2026): 7,000,000 common shares at a deemed price of USD $0.50, a stated value of USD $3,500,000. Issued in equal monthly installments beginning July 6, 2026, as services are rendered.

Phase two (auto-continuing): Unless the Company delivers written notice at least ten days before September 30, 2026, electing to pause, services continue through December 31, 2026 for an additional 3,000,000 shares at the same deemed $0.50, USD $1,500,000. Split 1.5M on November 6 and 1.5M on December 6.

The 2027 option: The Marketing Partners hold an option, exercisable on mutual written consent, to receive an additional USD $5,000,000 in Company Shares at the greater of $0.50 or a 10% premium to market, for renewal of substantially similar services in calendar 2027.

Lock-up: Every tranche carries a twelve-month contractual lock-up from its issuance date.

Relationship: Arms’ length. No finders’ fees.

The correction I want to make loudly

Several aggregators, including at least one AI-generated summary that has been circulating, have described the US$5 million as a potential future investment into Rocket Doctor AI to support U.S. expansion.

That is backwards. Read the clause. The Marketing Partners have an option to receive $5,000,000 in Company Shares as consideration for renewing their services in 2027. That is money flowing out, not in. It is a renewal price tag, not a funding commitment.

If you have seen that framing repeated and believed it, correct it now. It changes the deal's entire arithmetic.

So what does this actually cost

Ten million shares across 2026 against roughly 96.9 million outstanding as of the Q1 MD&A date. Call it just under 10% dilution on a post-issuance basis for a seven-month brand campaign. If the 2027 option is exercised at the $0.50 floor, that’s another ~10 million shares.

Two honest observations about that:

It preserves cash, which matters. For a company where the binding constraint is physician credentialing throughput and where the balance sheet just took on a c. C$3.1M convertible debenture at 12% (closed August 6, 2026), not writing a $3.5M cheque is a real advantage. Stock-settled marketing is dilution, but it is dilution you choose rather than debt you service.

The deemed price is doing work. A $0.50 USD floor sets a ceiling on share count. If the market price sits below that deemed price when tranches issue, existing shareholders are paying fewer shares than a market-priced deal would require. If it sits well above, the partners are getting a discount. Which way that cuts depends entirely on where the stock trades through the issuance calendar, and the twelve-month lock-up means those shares don’t hit the float until mid-2027 at the earliest.

That last point matters more than the sponsorship itself. Ten million locked shares stacked on top of the December 2026 escrow tranche and the January 22, 2027 warrant expiry is a supply calendar worth mapping.

Now, the Biohaven precedent

Here is where it gets genuinely interesting, and where I want to be very careful not to sell you something I can’t support.

The No. 51 Rick Ware Racing car is not new to healthcare branding.

On December 2, 2020, Biohaven Pharmaceuticals (then NYSE: BHVN) announced that Nurtec ODT (rimegepant), its migraine treatment, would be the primary partner on the No. 51 RWR entry for the entire 2021 NASCAR Cup Series, debuting at the Daytona 500. The program expanded to the No. 51 IndyCar entry in May 2021 and renewed for a second full season across both series in January 2022.

The driver was Cody Ware, the same driver on the Rocket Doctor AI car, and Cody lives with migraine himself. Rick Ware initially connected with Biohaven’s CEO because his own family members suffer from migraines.

Alongside the racing, Biohaven and RWR ran a Military Salutes Program: thirty-five military base visits across 2021, over 1,200 service members honoured, and more than $150,000 donated to service member and first responder charities.

In May 2022, Pfizer announced it would acquire Biohaven for approximately $11.6 billion. The deal closed October 3, 2022, at $148.50 per share in cash, with Biohaven simultaneously spinning off its non-CGRP pipeline into a new listed entity.

Roughly twenty-two months from first green flag to closed acquisition.

What that precedent legitimately establishes

The platform can credibly carry a healthcare brand. This is nothing. Regulated healthcare marketing is hard; a pharmaceutical company with an FDA-approved drug and an army of compliance lawyers signed off on putting a prescription migraine therapy on a Cup Series car and kept it there for two full seasons. That is a real reference point, and it is the strongest argument for the format.

The audience overlap is real. Biohaven’s thesis was that NASCAR reaches rural, working-class, military-adjacent populations with chronic conditions and inconsistent access to care. That's nearly identical to Rocket Doctor’s stated core patient population: rural, underserved, working-class, Medicaid and Medicare. Cody Ware said as much in the Rocket Doctor release, and Biohaven’s team said the same thing five years earlier.

Rick Ware Racing has done this before. They are not learning healthcare activation on Rocket Doctor’s dime.

What that precedent absolutely does not establish

One. Pfizer did not buy a brand. It bought a molecule. Pfizer had already signed a strategic collaboration with Biohaven in November 2021 to commercialize rimegepant and zavegepant outside the U.S., $500 million paid at closing in January 2022, including a $350 million equity investment. The acquirer was already inside the company as a partner and shareholder before the takeout. That is the actual causal chain: a commercial partner exercising an option it had effectively already priced. NASCAR was a line item.

Two. The scale gap is enormous. Biohaven was a commercial-stage NYSE company with an FDA-approved drug generating hundreds of millions in revenue, running celebrity DTC campaigns in parallel. Nurtec crossed $928 million in 2023 sales under Pfizer. Rocket Doctor AI is a micro-cap paying its marketing partners in stock. The reused template doesn't make the outcomes comparable, and pretending otherwise is a category error.

So while I am not saying a deal with Cody or FINTEKK means Rocket Doctor AI will be acquired or re-rate to 30X+ over the next couple of years. I am saying that this partnership is extremely accretive to Rocket Doctor AI and all of its stakeholders.

The actual investable question

Strip out the acquisition fantasy and the honest version of this thesis is narrower and more testable.

Rocket Doctor’s binding constraint is not awareness. It is credentialed physician supply. The company went from roughly 12 clinically active U.S. MDs in April 2026 to 22 active with 33 in credentialing. Patient visits moved from under 100 per month in December 2025 to over 1,000 per month by April 2026. That growth is a supply-side story, not a demand-side one.

So the question is: does a national brand campaign relieve a supply constraint?

It plausibly can, through two channels neither the bulls nor bears talk about much:

Physician recruitment. A doctor deciding whether to build a practice on an unfamiliar platform weighs institutional credibility. A brand visible on NBC and FOX is not a clinical credential, but it is a legitimacy signal, and physician acquisition is the actual bottleneck.

B2B and payer conversations. Health plan business development is a relationship business where being a known name shortens the sales cycle. The campaign’s stated geographic emphasis, California, New York and the Tri-State area, Texas, Florida, maps onto both existing payer geographies and expansion targets.

The weaker channel is the one the press release leads with: direct patient acquisition. National broadcast reaches all fifty states. Rocket Doctor can only see patients where it has credentialed physicians and in-network payer contracts, currently California, New York, and Maryland. A viewer in Ohio who sees the car and searches the platform is a wasted impression. That geographic mismatch is the strongest bear point on this deal, and management has not addressed it directly.

Named risks

• Dilution is real and front-loaded. Roughly 10% in 2026, potentially another 10% in 2027 if the option is exercised. This is not a free deal because no cash moved.

• Auto-renewal is opt-out, not opt-in. Phase two continues by default unless the Company affirmatively elects to pause ten days before September 30, 2026. Watch for that decision; silence is a decision.

• The 2027 option converts a marketing partner into a large holder. Twenty million cumulative shares under lock-up create a future supply overhang with a known unlock schedule.

• Attribution will be difficult. Brand campaigns are hard to measure. If Q3 and Q4 physician credentialing and visit volumes do not inflect, the company will have spent 10% of the equity on something it cannot demonstrate worked.

Five takeaways

1. The $5 million is an outflow, not an inflow. Correct this wherever you see it. The Marketing Partners have an option to receive $5M in shares for 2027 renewal services.

2. The real cost is ~10% dilution in 2026 at a deemed $0.50 USD, all under a twelve-month lock-up, landing in the same window as the December escrow tranche and the January 2027 warrant expiry.

3. The Biohaven precedent is real and useful, as proof the platform works for healthcare brands. Same team, same car number, same driver, two full seasons, a compliance-heavy pharmaceutical sponsor.

4. The Biohaven acquisition is not a roadmap. Pfizer bought a CGRP franchise it was already partnered on. The sponsorship was incidental. Anyone selling you “sponsored, then acquired” is selling correlation.

5. The thesis validation metric has not changed. It is credentialing conversion and per-physician visit volume, not impressions, not followers, not broadcast reach. If Q3 numbers show credentialing accelerating, the campaign has a case. If they don’t, this was an expensive paint job.

I hold over 1.5 million shares of Rocket Doctor AI ($AIDR / $AIRDF / FSE: 939), accumulated over time, representing a meaningful portion of my portfolio. I have no relationship with Rocket Doctor AI, Rick Ware Racing, or FinTekk AP, and I am not compensated for this content. Figures cited are drawn from the Company’s June 8, 2026 press release, the Q1 2026 MD&A and interim financial statements filed on SEDAR+, and publicly available sources on Biohaven Pharmaceuticals and its acquisition by Pfizer. Analyst coverage referenced elsewhere in my work carries disclosed conflicts: Maxim Group makes a market in the security; Fundamental Research Corp is paid for coverage.

NFA. DYOR.

On August 19, Aduro announced that it has selected Saipem S.p.A. for Early Works and Services supporting development of its First-of-a-Kind Hydrochemolytic™ Technology facility at Chemelot in the Netherlands.

I want to be careful here, because this is exactly the kind of announcement that retail investors either dismiss (”it’s just engineering”) or wildly oversell (”Saipem is building the plant!”). Neither is right. What actually happened is narrower than the bulls will say and more significant than the skeptics will admit.

Let me walk through what I think this is.

1. What “Early Works” actually means

Industrial projects don’t go from idea to steel in one step. The conventional sequence in process engineering looks roughly like this:

1. Process Design Package (PDP) — the technology owner’s basis of design

2. Early Works / pre-FEED — an EPC contractor reviews the PDP, tests assumptions, scopes long-lead equipment, refines cost

3. FEED (Front-End Engineering Design) — the detailed engineering package that supports a capital decision

4. Detailed engineering, procurement, construction

5. Commissioning and start-up

Aduro is entering step 2. That’s it. The press release is explicit that progression through FEED, detailed engineering, procurement, construction, commissioning and start-up is conditional on completing each development stage, continued evaluation by both parties, and executing definitive agreements.

Read that again, because it’s the whole ballgame: Saipem has not committed to build this plant. The forward-looking statements section even lists, as a named risk, the possibility that the parties never enter a full FEED services contract.

2. But the counterparty is the point

Here’s why I still think this is one of the year's more meaningful disclosures.

Saipem is not a boutique. FY2025 revenue was roughly €15.5 billion, with a backlog above €31 billion and something north of 30,000 employees across offshore and onshore engineering and construction. This is a firm that builds trunklines for Aramco and process facilities in the energy and chemicals sector at industrial scale.

Firms of that size have gatekeeping functions. They have technology assessment teams whose job is to tell business development “no.” They have reputational exposure; a Tier-1 EPC that puts its name on a first-of-a-kind unit that fails has a problem that a €15 billion revenue base does not paper over.

The press release notes the selection followed a comprehensive evaluation process. I read that as bidirectional. Aduro evaluated contractors; contractors evaluated Aduro. Saipem looked at the Hydrochemolytic chemistry, the NGP pilot data, and the FOAK basis of design, and decided the project was worth putting engineers on.

That is not the same as technical validation; Saipem is being paid, and paid engineers work on projects that later fail. But it is a counterparty with a reputation, a technical assessment function, and no obvious incentive to attach itself to a small-cap process that doesn’t hold up. In a sector where every developer claims a breakthrough, who is willing to sit across the table from you is real information.

I’ve written before that the most useful signal in a pre-revenue story is not what management says about the technology; it’s who else is willing to spend their own time on it. Zeton built the pilot. Siemens did the automation. An unnamed steam-cracker licensor ran the oil through a furnace. Ebert HERA took the permitting. Now Saipem takes the engineering. The list of counterparties has gotten steadily less exotic and more industrial.

3. The five words I care about most

“funded from the Company’s existing cash resources”

Early Works activities are expected to be funded from existing cash and are not expected to require additional financing.

For a development-stage company, this line does a lot of work. It tells you the near-term engineering spend does not come with a raise attached. Aduro closed a LIFE offering in June 2026; depending on the source, pro forma cash sits somewhere in the C$65–72 million range against roughly C$10 million of annual burn. (Roth modelled approximately US$50.6 million post-offering; I’d treat any specific figure as needing verification against the FY2026 year-end filing before you build on it.)

The point isn’t the exact number. The point is that Aduro can fund the engineering phase without going back to the market, and management chose to say so in the second paragraph of the release. That sequencing is deliberate. It preempts the reflexive small-cap assumption that every partnership announcement is a financing announcement in disguise.

4. What Saipem is actually doing and the number it will produce

The initial scope covers four things: reviewing the Process Design Package, optimizing critical equipment packages, preliminary utility integration, and refining capital costs.

That last one deserves its own paragraph.

We do not currently have a company-disclosed capex figure for the Chemelot FOAK plant. What we have are sell-side estimates. Roth’s July initiation modelled roughly US$50 million of total installed cost to reach 25,000 tonnes per year, with the facility built initially at 10,000 t/yr. That’s an analyst’s number, not Aduro’s.

A Tier-1 EPC refines capital cost estimates into a real number. And a real number is the input every valuation model in this story is currently guessing at. If you’re running a licensor DCF as I am, the capex per tonne of installed capacity drives licensee project economics, which drives royalty capacity, which drives the entire licensing thesis. A credible FOAK capex disclosure would be one of the highest-information events available to us in the next twelve months, and it could cut either way.

I’d rather have the honest number early than a flattering one late.

5. The pilot-engineering feedback loop

The release says engineering activities are deliberately aligned with ongoing NGP pilot campaigns, so operating lessons feed directly into industrial design.

Two ways to read this, and I think both are true.

The bull reading: this is exactly what a disciplined scale-up looks like. The June 2026 campaign produced roughly 86% liquid hydrocarbon yield with about 85% at C20-and-below over a 47-hour continuous run, including a two-hour recovery from deliberate disturbance. Those are design inputs. Feeding them into engineering as they’re generated, rather than freezing a design in 2025 and hoping, materially lowers the risk of building the wrong plant.

The bear reading: a design that keeps absorbing new data is a design that hasn’t been frozen. Design freeze is what makes schedules real. Every process developer that has slipped has slipped partly because “one more optimization” kept arriving. The integrated approach is genuinely lower-risk on technology and genuinely higher-risk on schedule.

I hold both. I don’t think you get to claim the first benefit without accepting the second cost.

6. The open question nobody has answered

In March 2026, Aduro signed a non-binding MOU with an unnamed “leading Global Engineering, Procurement, and Construction organization” to jointly develop a commercial licensing package and a pre-engineered plant concept, the foundation of the licensing business model that carries most of the long-term value in my framework.

Is Saipem that GEPC?

Aduro has not said so, and I am not going to assert it. The August 19 release does not connect the two, and the March release described a counterparty with a long track record of working alongside technology licensors, a description that fits several firms. It would be tidy if they were the same party. Tidy is not evidence.

This matters more than it looks. If Saipem is the GEPC, then the licensing pathway and the FOAK execution pathway run through one relationship that is now deepening a big deal. If they’re separate, then Aduro has two independent Tier-1 engineering relationships, which is a different and arguably more robust structure. Either is fine. Guessing which one is true is how people end up wrong in public.

If someone asks you on X whether Saipem is the GEPC, the correct answer is “the company hasn’t disclosed that.”

7. What this announcement does not tell you

Let me be explicit, because I’d rather own the limitations than have them pointed out to me:

• It does not validate the chemistry. Saipem is an engineering contractor, not a technology certifier.

• It does not commit Saipem to FEED, construction, or anything beyond the Early Works scope.

• It does not disclose contract value, and Early Works scopes at this stage are typically small relative to project capex.

• It does not solve permitting, which remains the stated gating risk at Chemelot.

• It does not convert the offtake LOI into a binding agreement.

A Conditional Letter of Award is referenced in the forward-looking statements. “Conditional” is doing real work in that phrase.

8. What I’m watching from here

In rough order of information value:

1. Execution of a full FEED services contract with Saipem. This is the first genuine confirmation that Early Works cleared its gate. Absence of this by mid-2027 would be a meaningful negative signal.

2. A company-disclosed FOAK capital cost. The single most valuable number Aduro could give us.

3. Chemelot permitting milestones under the Ebert HERA engagement.

4. FID and groundbreaking. The transition from planning to capital deployment.

5. Offtake LOI → binding agreement, which sets the first real benchmark on circular naphtha realization.

6. Whether the GEPC is ever named, and whether the licensing package progresses to definitive terms with performance guarantees.

7. Continued NGP campaign data, particularly on mixed and contaminated feedstocks and energy consumption.

9. What would change my mind

Unchanged from what I’ve written before, and this announcement doesn’t move any of them:

1. Scale-up failure at FOAK. The step from 10 kg/hr at NGP to roughly 1,000 kg/hr at Chemelot is the central technology risk. Engineering support reduces execution risk; it does not eliminate scale-up risk.

2. No commercial license signed by end of 2028. The licensing model drives the valuation. An EPC relationship is the delivery mechanism, not the revenue.

3. Repeated dilutive raises at falling prices. Today’s release addresses the near-term version of this. It doesn’t address the FOAK construction financing that follows.

Druckenmiller’s line, put your eggs in one basket and watch the basket very carefully, is the whole discipline here. Watching carefully means noticing when a stage gate opens, and equally noticing when one doesn’t.

The short version

Aduro moved from “we have selected a site and a permitting consultant” to “a top-tier global EPC is now doing engineering on our plant, and we’re paying for it out of cash on hand.” That’s a real step up the ladder. It is also, precisely, one rung, with FEED, FID, financing, permitting, and construction all still above it.

The stage-gate structure cuts both ways, and I think that’s the honest frame. It means Aduro is proceeding with discipline rather than committing capital ahead of data. It also means every gate is a place the project can stop.

I own this. I’m watching the gates.

Sourcing note: All company facts above are drawn from Aduro’s August 19, 2026 press release and prior GlobeNewswire disclosures. Saipem financials are from Saipem’s own FY2025 reporting. Roth Capital and Water Tower Research estimates are labelled as such. Verify cash figures against Aduro’s most recent filing before use. Where I’ve speculated, I’ve said so.

This is not financial advice. I own shares of $ADUR, and it represents ~30% of my equity portfolio, so I am biased. Long $ADUR | Not financial advice | DYOR.

Two videos dropped in the last stretch that, taken together, tell you almost everything you need to know about this company, and they tell it from opposite ends of the table.

The first is the Rocket Doctor AI Town Hall, June 3, 2026 management’s own account, delivered by Dr. Essam Hamza (CEO of the parent company) and Dr. William “Bill” Cherniak (Founder & CEO of Rocket Doctor Inc.). That’s the inside view. Treat it as management commentary, not filed disclosure.

The second is “This AI Company Gives Doctors a Money Machine” from Amateur Investing, an outside retail analyst with no relationship to the company who builds the payer-economics case from first principles. That’s the outside view, and it’s the more interesting of the two, because it independently arrives at the same place management does.

I want to walk through what both said, separate what’s filed from what’s asserted, and then tell you where I think the market is wrong.

1. The problem isn’t money. It’s throughput.

The US spends roughly $5 trillion a year on healthcare and still leaves about 100 million people without a dedicated primary care physician. Canada leaves an estimated six million people without a family doctor (Town Hall, ~03:42). When people have nowhere else to go, they go to the ER for sore throats, UTIs, rashes, and medication refills.

Meanwhile, the supply side is drowning. Both videos land on the same number independently: physicians spend 40–50% of their working hours on paperwork, billing, and compliance rather than practicing medicine (Town Hall, 09:12; Amateur Investing, 30:32).

So you have unmet demand on one side and misallocated supply on the other, with an expensive middle layer absorbing the difference. That’s not a funding problem. That’s a routing problem, and routing problems are exactly what software solves.

2. What’s actually under the hood

I’m allergic to “AI healthcare” companies that are a GPT wrapper with a stethoscope logo. This one isn’t, and the distinction matters.

The Global Library of Medicine (GLM). Management describes it as built over seven to eight years by more than 200 practicing physicians contributing over 25,000 curated hours, mapping 17,000+ symptoms across 1,000+ diseases (Town Hall, 11:06).

It reasons rather than retrieves. The engine dynamically shifts its questioning based on patient answers, updates a differential diagnosis in real time, and can suggest labs or imaging before the physician ever joins the call (Town Hall, 11:49). Amateur Investing describes the same mechanism as Bayesian probability applied to prior clinical data (Amateur Investing, 04:11). That is a fundamentally different architecture from a model that scrapes and summarizes.

It’s been externally graded. Medical schools including the University of Minnesota have used the system to evaluate students during OSCE clinical-skills examinations (Town Hall, 14:27). Academic institutions do not put a hallucination machine in front of their licensure pipeline.

Around the GLM sits the actual product, what management calls a “Shopify-like” ecosystem for independent practice (Town Hall, 19:00), and these new features are being launched between now and the rest of the year, each of which is a new revenue engine if launched successfully:

• AI Nurse/intake: 24/7 triage, severity assessment, ER routing for genuine emergencies, pre-visit lab ordering, and booking (Town Hall, 16:48)

• AI Scribe: ambient transcription and charting, so nobody types during the visit (Town Hall, 17:34)

• Remote care hardware: Bluetooth stethoscopes and otoscopes for examination at distance (Town Hall, 19:28)

• Patient marketplace: demand delivered to the physician, so they aren’t buying ads or signing leases (Town Hall, 20:24)

• Automated billing: claims submitted and reconciled by the platform (Amateur Investing, 05:51)

The doctor logs in, sees patients, and gets paid. That’s the whole pitch, and it’s a good one.

3. The part that actually convinced me: the payer

Most digital health companies die because someone in the chain loses. Amateur Investing’s contribution, and I think this is the sharpest thing in either video, is the payer arithmetic (Amateur Investing, 20:14).

Health insurance runs on brutally thin net margins, roughly 0.4% to 1.2%. An average treat-and-release ER visit costs an insurer somewhere around $1,000. A virtual visit costs them something closer to $100–$150. His conclusion: if a payer diverts even 1% of low-acuity ER visits into virtual care, corporate net profit could move 8–10%.

Be clear about what that is. That is an independent creator’s model, not company guidance and not a figure in any filing. I have not verified his inputs. But directionally it explains something that would otherwise be strange: why national payers are signing in-network agreements with a company this small. They aren’t doing it out of charity. Virtual diversion is one of the few levers that moves a 1%-margin business.

And the patient side isn’t a compromise. Same-day, in-network, covered, from home. Amateur Investing notes Rocket Doctor’s customer reviews average around 4/5 on Trustpilot against roughly 1.2/5 for UnitedHealthcare (Amateur Investing, 25:01). Whatever you think about review-site methodology, the gap is not subtle.

4. Now the filed numbers, because this is where the thesis lives or dies

Everything above is narrative. Here is what’s actually in the Q1 2026 MD&A and financial statements filed on SEDAR, which is the only source I’ll build a position on.

Monthly completed US visits rose 283% during Q1 2026, then another ~69% month-over-month in April (Q1 2026 MD&A). Canada, the mature business, did 47,790 completed visits in Q1 2026 against 37,176 a year earlier — still growing, but this is not where the story is.

The covered lives: approximately 21 million in-network lives across California (~8.1M), New York (~9.9M), and Maryland (~3.1M, raised to ~3.2M after an April payer addition). The MD&A states plainly that less than 3% of projected revenues associated with these payer agreements are reflected in current financial statements.

The revenue: C$737,103 in Q1 2026, up ~5% sequentially from C$697,340. That’s it. Against 21 million covered lives.

That gap — 21 million contracted lives, three quarters of a million dollars of quarterly revenue — is the entire investment thesis. Everything else is commentary.

Why the gap exists, and why I think it closes. Two reasons, and neither one is “patients don’t want this.”

1. Credentialing is the binding constraint, not demand. Commercial and Medicaid credentialing runs three to six months per physician. Management reported roughly 80 active providers by end of May 2026 — double the prior count — with nearly three times that number sitting in the credentialing pipeline, acquired with effectively zero physician-marketing spend (Town Hall, 34:50). Cherniak added that not a single physician has left the US platform since launch (Town Hall, 41:36). The supply side isn’t leaking; it’s queued.

2. Cash-basis government billing lags visits by 45–90 days. A visit completed in April shows up in revenue in Q2 or Q3. The reported line always trails the operating reality by roughly a quarter. If you mark the company on trailing revenue, you are marking it on a stale picture by construction.

Strategic note worth appreciating: management deliberately opened in California, New York, and Maryland first — three of the most regulatory-hostile states in the country (Town Hall, 26:30). That’s counterintuitive until you realize the playbook you build for California generalizes downward to everywhere else. They took the hardest exam first.

5. Where I’d push back on my own thesis

I own this. That obligates me to be harder on it than the bears are.

The per-visit economics don’t reconcile. Management and Amateur Investing both cite roughly US$25 per US visit (Town Hall, 27:48; Amateur Investing, 17:02). Maxim Group’s initiation note says US$18. I’ve also seen much higher per-visit figures used in penetration math floating around retail channels. Three different numbers, three different sources, and the spread is material; a 28% difference in take rate compounds hard across any volume forecast. Until this reconciles against filed disclosure, I’d describe the US opportunity as a nine-figure USD revenue opportunity at scale and refuse to be more precise than that. Anyone giving you a decimal point here is guessing.

“Minimal debt” is no longer a thing you can say. On August 6, 2026, the company closed a first tranche of unsecured convertible debentures for gross proceeds of C$2,269,000 at 12% annual interest, maturing in twelve months. Twelve percent is not cheap money. It tells you the company needed capital and chose this over an equity raise at the prevailing price. You can read that charitably (unwilling to dilute at these levels) or uncharitably (limited options). Both readings are live.

The cash burn is real. C$3.5M cash at March 31, 2026, against C$3.09M used in operating activities during Q1 alone, and a C$4.65M comprehensive loss for the quarter against C$737K of revenue. Marketing spend nearly tripled year over year to C$963K. This company is buying growth, and the runway is measured in quarters, not years.

There’s a warrant stack. Roughly 22.1 million warrants outstanding as of March 31, 2026, at a weighted average of $0.79, including ~5.2M at $0.75 expiring August 21, 2026 (that’s days from now), ~3.7M at $0.75 expiring October 2, 2026, and ~7.4M at $0.85 expiring January 22, 2027. Above strike, those are non-dilutive proceeds and a genuine reason the company is spending on investor visibility right now. Below strike, they expire worthless, and the funding question comes back. Layer the new debenture conversion on top, and you have a multi-tranche overhang that any honest bull has to hold in view.

Quality control sits outside the company’s walls. Physicians operate independently. A rude doctor or a missed appointment is Rocket Doctor’s brand damage and not Rocket Doctor’s employee (Amateur Investing, 39:42).

6. The valuation, and why I’m still here

Management pegged the market cap around C$50–60M during the town hall (Town Hall, 38:50); Amateur Investing put it near US$40M (Amateur Investing, 41:34). Meanwhile, private health-AI comparables are printing at $1B to $12B+ on far less clinical infrastructure.

I don’t put much weight on comp-table arguments; private marks are marketing documents. What I put weight on is this:

The expensive, slow, un-fakeable work is done. Payer contracts across three of the hardest states: signed. Credentialing infrastructure: built over three and a half years. Revenue operations, billing, collections: running. Physician supply: queued three-deep with zero acquisition cost. The clinical engine: eight years and 25,000 physician-hours in.

What remains is throughput. Credentialing pace, and the 45-to-90-day billing lag catching up to a visit curve that has gone from 86 to over a thousand a month in five months.

That’s my thematic filter and always has been: setups where the structural work is complete but the market is still pricing the “before” picture. This is the cleanest example of it I own.

For balance: analyst coverage exists, and both notes carry conflicts you should know about. Maxim Group rates it Buy with a C$3.00 target. Fundamental Research Corp has it Buy at C$1.86 fair value, and is paid for coverage.

7. What would prove me wrong

I’d rather name these now than rationalize later.

1. Q3/Q4 revenue doesn’t inflect. If the billing lag catches up and revenue is still under C$1M a quarter with visits above 1,000/month, the per-visit economics are worse than any published figure and the thesis breaks.

2. Credentialed physician count stalls. If active providers sit near 50 into Q4 while the pipeline stays “three times” that number, credentialing isn’t a queue; it’s a wall.

3. Visit growth flattens before it hits meaningful scale. The ramp is off a tiny base. Compounding off 86 is easy. Compounding off 1,000 is the actual test.

4. The next raise comes at a materially lower price. Another 12% instrument or a discounted equity round would tell you the debenture wasn’t opportunistic; it was necessary.

5. A payer churns. Losing one in-network contract would say the ER-diversion economics don’t work the way Amateur Investing modelled them.

Q2 2026 results are the next real data point, and I’m watching them.

Sources

• Rocket Doctor AI Town Hall, June 3, 2026 — Dr. Essam Hamza & Dr. William Cherniak:

• “This AI Company Gives Doctors a Money Machine” — Amateur Investing:

• Rocket Doctor AI Inc. Q1 2026 Interim MD&A and Condensed Interim Consolidated Financial Statements (SEDAR+)

• Rocket Doctor AI Inc. news releases, July 28 and August 6, 2026 (GlobeNewswire)

• Maxim Group initiation of coverage; Fundamental Research Corp. update, both carry disclosed conflicts

Video timestamps are cited as given. Statements attributed to the Town Hall are management commentary. Statements attributed to Amateur Investing are the independent analysis of a third-party creator with no disclosed relationship to the company. Only figures attributed to the MD&A, financial statements, or news releases are filed disclosure.

Disclosure & Disclaimer

I am long Rocket Doctor AI Inc. (CSE: AIDR | OTC: AIRDF | FSE: 939), representing approximately 6% of my equity portfolio as of publication. I have not been compensated by the company or by any third party for this article. I may buy or sell shares at any time without notice or update.

This is a thinly traded micro-cap. Liquidity is poor, spreads are wide, dilution risk is real, and the company is not profitable. Nothing here is a recommendation to buy or sell anything. Everything in this piece is my own opinion and analysis, offered for informational and educational purposes only. Verify every figure against the company’s filings on SEDAR+ before you act on anything.

NFA. DYOR.

On August 11, Rocket Doctor announced it had gone in-network with another major national insurer in New York — adding more than 100,000 eligible covered lives across Medicare Advantage PPO and commercial plans in the New York City metro region and select upstate counties.

If you only read the headline number, you will underreact. If you read what’s underneath it, you’ll see why I’m adding.

Let me be honest about both.

1. What was actually announced

Straight from the release, so we’re working from the same facts:

• More than 100,000 additional eligible covered lives across individual Medicare Advantage PPO and commercial plans.

• Geography: NYC metro region plus select upstate counties.

• The payer participates in a national reciprocal network program, meaning eligible out-of-state plan members can access Rocket Doctor’s in-network physicians while travelling in New York.

• The estimate excludes members covered under administrative services-only (ASO) or self-insured plans.

• The agreement was signed on July 15, 2026, and announced on August 11. Initial one-year term, automatic one-year renewals unless terminated.

• Running total: ~10 million covered lives in New York, ~24 million across the U.S.

That last number matters. The Q1 2026 MD&A put the total at approximately 21 million covered lives across California, New York and Maryland. We’re now at roughly 24 million. That’s meaningful movement in a single quarter, and it happened without a single press release that moved the stock in any durable way.

2. 100,000 is the wrong number to anchor on

Let’s kill the obvious criticism first, because it’s a fair one.

100,000 lives on a 24 million base is a 0.4% increase. On its own, that is not a stock-moving event, and anyone telling you it is doesn’t understand the math. I’m not going to pretend otherwise.

What I actually care about is the composition.

This is Medicare Advantage PPO and commercial. Those are not Medicaid managed care lives. In U.S. virtual care, reimbursement per encounter is not uniform — Medicare Advantage and commercial lines generally reimburse at higher rates than Medicaid, and Medicare Advantage members are, by definition, an older and higher-utilization population. Rocket Doctor’s early U.S. footprint was heavily weighted toward Medicaid and Medicare/Dual-Eligible. Adding commercial and MA PPO shifts the quality of the covered-life base, not just the count.

The reciprocal network detail is the other piece most people will skim past. A national reciprocal program means eligible members of that plan from outside New York can use Rocket Doctor’s in-network physicians while they’re in the state. It’s a small structural widening of the funnel that doesn’t show up in the headline count at all, and neither do the ASO and self-insured members, the company explicitly excluded from its estimate.

So: small number, better-quality number, and a conservatively stated number. That combination is what I’m paying for.

3. The gap is the entire thesis

Here is the setup in one comparison, and it is the reason I own this at all.

Contracted, in-network access: ~24 million covered lives. Q1 2026 reported revenue: C$737,103.

That’s the whole trade. Rocket Doctor has spent three-plus years building payer contracts, credentialing infrastructure and revenue operations specific to Medicaid and Medicare, and the market is pricing the company off a P&L that reflects almost none of it.

Management’s own language in the Q1 MD&A is that less than 3% of the projected revenues associated with the newly finalized payer agreements are reflected in current financial statements. That’s a management estimate, not an audited figure, and you should treat it as such. But the direction is not in dispute: the contracts exist, the revenue does not yet.

I’ve written before that this is a before picture still being priced by the market. Nothing in this release changes that. It extends it.

4. The bottleneck is credentialing supply, not patient demand

This is the part I’d want any skeptic to sit with.

Rocket Doctor does not have a demand problem. Per the Q1 MD&A, monthly completed U.S. patient visits rose 283% during Q1 2026, and April 2026 came in around 1,200 completed visits, a further 69% month-over-month increase. (The July investor deck cites 1,144 for April; I’d use the deck’s more precise figure and treat ~1,200 as rounding in the MD&A.)

What it has is a supply problem. The number of active providers and specialists on the platform roughly doubled by the end of May 2026.

Very few physicians. Twenty-four million covered lives.

That ratio is the thesis and the risk in a single line. Every new payer agreement expands the addressable pool. None of them converts to revenue faster than the company can credential physicians into those networks. Credentialing is measured in months; it is payer-by-payer and state-by-state, and it does not accelerate because a press release went out.

So when I track this name, I am no longer tracking covered lives. I’m tracking active credentialed physician count and pipeline. That is the forward-looking number. Everything else is downstream of it.

5. Why the reported revenue understates what’s happening

One structural point that I think the market genuinely misreads.

For its U.S. operations, Rocket Doctor recognizes revenue on a cash basis, not when the visit occurs, but when payment is actually received. The MD&A is explicit that collectability cannot be established with sufficient certainty at the time of service to satisfy IFRS 15, given varying collection timelines across payers.

The practical consequence: a visit completed in July may not appear as revenue until Q4. There is a structural lag of roughly 45–90 days between visit volume and reported revenue.

I’d argue that is conservative accounting, not a weakness. It means reported revenue is a lagging, cash-confirmed indicator of an operating business that has already grown past what the income statement shows. It also means Q2 and Q3 prints are the real test, that’s where the Q1 visit surge should begin landing.

If it doesn’t land there, my thesis has a problem. I’ll say so when the numbers come out.

6. What else happened this week

On August 5, Rocket Doctor entered a digital advisory agreement with B2i Digital, joining its Featured Company Program and its self-described network of more than 1.7 million retail and institutional market participants.

Two things worth stating plainly. First, the 1.7 million figure is self-reported by B2i and its methodology is not publicly defined; treat it accordingly. Second, IR spend is immaterial relative to this company’s burn, so “is it accretive” is the wrong question. The right question is sequencing: the company is spending on investor awareness roughly six months ahead of a January 2027 warrant expiry at $0.85. If the stock is above $0.85 into that window, roughly C$6.3 million in non-dilutive proceeds unlock. If it isn’t, those warrants expire worthless, and the company raises again on worse terms.

That’s the actual logic. Whether it works is a separate question.

7. What would prove me wrong

I don’t publish a thesis without naming its failure conditions:

1. Credentialed physician count stalls. If active providers sit still through Q3, the number of covered lives is decorative.

2. Q3 and Q4 revenue doesn’t inflect. Cash-basis lag explains a quarter. It does not explain two.

3. Additional debenture tranches close on terms similar to or worse than those.

4. Covered lives keep growing while revenue per covered life stays flat. That would mean access is being signed but not converted, and the press-release cadence would be running well ahead of the business.

5. Payer non-renewal. This agreement has an initial term of 1 year. So do others. Access can be withdrawn.

Why am I adding anyway?

Because I think the market is pricing a company with C$737K of quarterly revenue, and I think it is buying an infrastructure position, payer contracts, credentialing workflows, revenue operations, that took three and a half years to build and that a competitor cannot replicate with capital alone.

Because the constraint is credentialing throughput, which is a solvable operational problem, not a demand problem. Demand problems kill companies. Throughput problems get fixed.

Because the covered-life base is growing and improving in mix, Medicare Advantage PPO and commercial are better lives than the Medicaid base the company started with.

And because the financing terms that make this uncomfortable are exactly why the price is where it is. If the balance sheet were clean and the credentialing pipeline were full, it would not be trading at the current valuation. I’d be reading about it in a research note.

The risk is real. The dilution is real. I’m sizing accordingly; over 6% is a conviction position, not a bet-the-portfolio position, and it stays that way until the Q2 and Q3 numbers tell me something new.

The next thing I’m watching is not a press release. It’s the credentialed physician count.

Sources: Company press releases via GlobeNewswire (August 5, 6 and 11, 2026); Rocket Doctor AI Inc. Q1 2026 interim MD&A and condensed interim consolidated financial statements (three months ended March 31, 2026, filed on SEDAR+); Rocket Doctor AI July 2026 investor presentation; Maxim Group initiation of coverage (January 2026); Fundamental Research Corp. update (May 2026).

Note on analyst coverage: Maxim Group rates AIDR Buy with a C$3.00 target and makes the security a market. Fundamental Research Corp. rates it Buy with a C$1.86 fair value estimate and is a paid-coverage provider. Both conflicts are disclosed in their own reports. Weight them accordingly.

Note on per-visit economics: Maxim’s initiation models a flat fee of US$18 per U.S. appointment. The company’s July 2026 investor deck cites US$25 per U.S. appointment. I have not been able to reconcile these from filed disclosures, and I’m flagging the discrepancy rather than picking a side.

Disclosure: I am long Rocket Doctor AI Inc. (CSE: AIDR | OTC: AIRDF | FSE: 939). The position now accounts for over 6% of my equity portfolio, and I have recently added to it. I receive no compensation for this or any other piece I write about AIDR. I may buy or sell at any time without notice. This is a micro-cap company with negative cash flow, going-concern language in its filings, and an active dilutive financing underway. It is not suitable for every investor. NFA. DYOR.

PureCycle Technologies reported Q2 2026 this morning. The market values that business at roughly $1.6 billion in enterprise value.

Aduro Clean Technologies has roughly $410 million in assets.

The reflexive response to that gap is “Well, PureCycle has revenue.” I want to take that response seriously and then show you why it doesn’t stand up to the numbers.

This is not a hit piece. PureCycle just put commercial recycled resin into a Procter & Gamble consumer product, which is technically hard, and the team deserves credit. But there is a difference between proving a product exists and proving a business works, and the market is currently pricing PureCycle as though it has done both.

1. What PureCycle Reported

Operational:

• First P&G commercial resin deliveries began; select Downy detergent caps are now in commercial production

• Revenue of $4.5 million, up ~173% year over year, a sixth consecutive quarter of sequential growth

• New Jersey DEP approved PureFive resin as post-consumer recycled content

• On-site compounding is now operating

• Initial PureFive shipments in Q3 to all three major converters for QSR cold cup trials

Financial:

• Net loss of $(142.2)M vs. $(144.2)M in Q2 2025

• Operating loss improved to $(41.3)M from $(45.6)M

• Adjusted EBITDA of $(31.7)M vs. $(27.8)M

• Total liquidity of $236.9M

• Core operations spending of $8.3M/month, down ~8% from $9.0M

• Materials spending of $2.1M/month, up from $0.7M/month in Q1

• A non-recurring legal settlement of $20.4M in cash

Give real credit for two items. The New Jersey approval converts PureFive from a product into a compliance instrument under a state mandate beginning in 2027, creating a genuine moat. And core monthly spend came down while production grew ~32% year over year through a planned turnaround. That is competent operating management.

Now let’s look at what it adds up to.

2. Skip the $142 Million

A lot of commentary will lead with the net loss headline. Don’t — it’s the weakest number in the release, and using it will get you correctly dismantled.

The $142.2M is dominated by non-operating, largely non-cash items; the same pattern produced a $144.2M net loss in Q2 2025, compared with a $45.6M operating loss. PureCycle says so directly: the year-over-year adjusted EBITDA comparison reflects roughly $7.8 million of lower non-cash add-backs rather than any deterioration in operating performance.

The number that actually matters:

• ~$31.7M/quarter negative adjusted EBITDA

• ~$8.3M/month core operations plus ~$2.1M/month materials

• $39–45M guided FY2026 project spend

Call it $150M+ of annual cash consumption against ~$18M of annualized revenue.

3. The Utilization Problem

This is where “but they have revenue” breaks.

Ironton’s nameplate capacity is 107 million pounds per year (~48,500 tonnes). Management has publicly estimated cash production costs at roughly $0.50–0.60 per pound.

Run the coverage math. Core operations spending is $8.3M/month (~$100M per year) and excludes materials. To cover corporate operating cash spend alone, the contribution margin has to clear $100M:

Required contribution: ~$100M At 107M lbs (100% of nameplate): need ~$0.93/lb of margin At $0.55/lb cash cost: need a realized price of ~$1.48/lb

That is the breakeven case. At full nameplate, at a premium realized price, PureCycle roughly covers operating spend — and has contributed nothing toward $438M of debt, nothing toward $39–45M of annual project spend, and nothing toward a return on the $360–365 million already sunk into the facility.

Drop to 80% utilization, and contribution falls to roughly $77M against $100M of spend. Still negative. The intuition that they need ~80% is conservative: the real bar is closer to 100% at premium pricing, and even then, it only reaches operating breakeven.

Where is utilization today? Q1 2026 was a record quarter, with approximately 10 million pounds of feedstock throughput, roughly 37% of nameplate throughput on an annualized basis. On resin actually produced, disclosed volumes have run lower. And on the metric that pays the bills, resin sold, $4.5M of quarterly revenue implies something in the neighbourhood of 15% of nameplate, depending on your assumption for realized pricing.

(Realized price per pound is my estimate; PureCycle does not disclose it directly. Substitute your own; the shape of the conclusion holds across any plausible range.)

And here is the detail that should give you pause: materials spending tripled from $0.7M/month to $2.1M/month as production restarted. Cost is scaling faster than revenue.

Revenue growing 173% sounds like de-risking. Revenue growing while the cost of producing it grows faster, at ~15% utilization, against a $438M debt stack, is something else. It proves the product ships. It does not prove the unit economics work. Those are different claims, and only the first has been demonstrated.

4. What a Pound of Capacity Costs

Ironton costs approximately $360–365 million for 107 million pounds of annual capacity, roughly $7,400 per tonne of installed capacity.

PureCycle’s own CEO has publicly acknowledged the inefficiency, contrasting the Ironton greenfield build with an estimate that a 130-million-pound line in Thailand could be built for “250-ish” by siting it at an integrated petrochemical complex rather than a greenfield. That is management telling you the flagship asset was capital-inefficient.

Aduro’s Chemelot FOAK is estimated by Roth at $50 million to reach 25,000 tonnes, about $2,000 per tonne. Water Tower Research models it higher (~US$39M for the initial FOAK plus shared site infrastructure, plus ~US$35M for the +15kt Phase 2), landing closer to $3,000 per tonne.

Somewhere between 2.5x and 3.7x less capital per tonne of capacity.

Two caveats, stated plainly rather than buried:

1. Ironton’s number is the money spent. Aduro’s is money projected. First-of-a-kind projects overrun. If Chemelot comes in at 1.5x budget, the advantage narrows materially.

2. The products differ. Ironton makes purified resin sold into the PP market. Chemelot will produce circular naphtha for sale into the petrochemical feedstock chain. Revenue per tonne differs, so capital-per-tonne is incomplete on its own.

The structural point survives both. Aduro is sitting at an existing integrated chemical park, precisely the strategy PureCycle’s CEO now says he’d prefer, and Aduro’s model transitions to licensing, where someone else’s balance sheet funds the next plant. PureCycle must build every incremental pound itself.

5. Where $30 Million Went

To push $216M of debt from a 2030 maturity to a 2032 maturity at a lower coupon, PureCycle paid approximately $246.3 million in cash in June 2026. Roughly a $30 million premium, real money, permanently gone, buying two years of runway on the maturity wall and nothing else.

The concurrent equity is trading at $8.21, down from a 52-week high of $17.37. The new 4.75% notes convert to around $11.08. The stock closed yesterday near $6.87, some 38% below the conversion price. Short interest was above 40% as of March 31.

Total consolidated indebtedness was $403.8M at March 31, 2026. Pro forma for the $250M issuance, it would have been $653.8M; net of the ~$216M repurchase, the stack lands around ~$438M, plus Series A preferred sitting ahead of common.

(Reconcile against the Q2 10-Q when it posts; I’m working from the Q1 filing plus the June offering documents.)

Contrast Aduro’s June 2026 raises: US$15.54M at US$15.20 and C$9.15M at C$21.20, with insider participation and no warrants attached. No convertibles, no ATM, no floorless instruments, no debt. Total debt of C$0.1 million, functionally zero.

Roth models an additional $50–60M of equity issuance through 2030. Aduro shareholders will be diluted; I’d rather say that than pretend otherwise. But there is a real difference between equity raised at rising, milestone-linked prices and cash paid at a premium to refinance a maturity wall. One fund’s progress. The other fund’s time.

6. The Scope Asymmetry

This is what I think the market isn’t pricing at all.

What PureCycle is

One polymer. Polypropylene. Not PE, not PS, not mixed streams, not multilayer flexibles. PP is a large market, but it's a one-lane road.

One mechanism. Solvent dissolution, a purification process. It removes colour, odour, and contaminants. It does not break carbon–carbon bonds. It cannot make naphtha, cannot feed a steam cracker, and cannot convert a polymer into anything other than a cleaner version of itself. Functionally, a very sophisticated washing machine for plastic.

Licensed IP. The process was developed by Procter & Gamble; PureCycle operates under license, and its own risk factors flag reliance on P&G-licensed technology. Whatever you think of the process, the invention is not the company’s.

One product, one market. Purified PP resin competes against virgin PP on price. When virgin PP is cheap, the spread compresses. Realized pricing is structurally tethered to a commodity that PureCycle does not control.

Feedstock-constrained. Dissolution needs to be sorted, reasonably clean PP. The hardest, dirtiest, most abundant fraction of the waste stream, the material that actually carries negative value, is not addressable.

One plant. Every incremental pound requires another balance sheet event.

What Aduro is

Three polymers, plus contamination tolerance. PE, PP, and PS together account for more than 70% of municipal plastic waste, precisely the streams that mechanical recycling cannot reach. Aduro’s November 2025 disclosure describes steam cracker trials in which HCT oil derived from mixed waste plastic, including PE, PP, PS, PET, and polyamide, was processed as-produced, without dilution or pre-treatment, at an established European pilot facility, with a stated boiling range substantially lower than that of comparable chemical recycling oils.

Read that against the section above. PureCycle needs the clean sorted fraction. Aduro’s stated target is the fraction nobody else wants.

Three verticals off one chemistry. Plastic upcycling is one application. Paraffinic and heavy crude upgrading is a second CIP patent filing, a dedicated program director hired in May, Utah Petroleum Association membership, and Uinta Basin feedstock secured in June. Renewables upgrading to aromatics and SAF is a third. Roth already carries paraffinic crude royalties from 2030 at ~95% gross margin. Management has indicated the paraffinic crude application requires its own separate pilot, treat commercial timelines, which are unconfirmed.

Every vertical shares the same R&D base. The marginal cost of the second and third applications is a fraction of the first. A single-polymer purification company structurally cannot replicate that, because dissolution does one thing to one material.

Owned IP. Eleven patents, seven granted, four pending, covering core chemistry and process integration, in a field historically dominated by ConocoPhillips, Shell and Total.

A feedstock, not a resin. Circular naphtha sells into the petrochemical value chain, not the PP resin market. Different buyers, different pricing dynamics, and a product that steam crackers can take directly, rather than a resin competing head-to-head with virgin pellets.

A licensing endgame. Build-own-operate the FOAK to prove the numbers, then license. Royalty revenue at ~95% gross margin, per Roth. The capital intensity that defines PureCycle’s entire story is exactly what Aduro’s model is designed to hand to somebody else.

No debt. 34% insider ownership. The founder has never sold.

Four independent sell-side firms cover the name with Buy ratings, D. Boral ($46), Roth ($30, initiated July 27), Ladenburg Thalmann (C$24.75), H.C. Wainwright ($22). Water Tower Research also publishes on Aduro; that is paid IR coverage and should be weighted separately from independent sell-side.

7. The Enterprise Value Bridge

Market cap is the wrong lens when capital structures diverge this hard.

PureCycle: ~$1.38B equity (≈200M shares near $6.87) + ~$438M debt − $237M liquidity ≈ ~$1.58B, before considering preferred.

Aduro: ~$460M equity (~33.7M shares near $13) − ~$50.6M cash (Roth’s post-June estimate) ≈ ~$410M.

A ratio of roughly 3.9x, and it holds better on EV than on market cap because EV captures the debt.

Per dollar of enterprise value, you’re choosing between:

8. What I Still Have to Answer

I’m long and biased, so let me make the case against myself rather than wait for someone else to.

Ironton’s struggle is a warning about Chemelot, not just about PureCycle. This is the one that matters, and I won’t have it both ways. If I use PureCycle’s utilization problem as evidence against them, I have to concede Aduro has never operated anything larger than a 10 kg/hr pilot. The ~86% liquid yield, the 47-hour continuous run, the ~85% at C20-and-below, that is pilot data from a single campaign. Ironton was going to hit the nameplate, too. Everything above about capital efficiency and licensing economics is contingent on a plant that doesn’t exist yet performing as well as a pilot did. If Chemelot has an Ironton-style ramp, this thesis is wrong, and the discount was correct.

“X is overvalued” does not prove “Y is undervalued.” Both can be mispriced in the same direction. If the market sours on advanced recycling as a category, PureCycle compresses, and Aduro doesn’t automatically expand.

Regulatory definitions could cut against conversion. If PCR-content rules continue to favour purification pathways over molecular conversion, that’s a structural headwind Aduro needs ISCC PLUS mass-balance certification to address. Watch the Delphi LCA.

I hold my position knowing all of this.

9. So What Is the Claim?

Not “Aduro should be worth $1.6 billion tomorrow.” That’s lazy, and it isn’t what I think.

The claim is this:

The market has demonstrated it will underwrite advanced recycling at $1.6 billion of enterprise value for a business that is one polymer, one mechanism, one plant, licensed IP, $438M of debt, and roughly 15% utilization, where full nameplate at premium pricing gets you to approximately operating breakeven, before debt service.

That is the revealed price of the category. It isn’t hypothetical; it was printed today.

Against that benchmark, a debt-free platform addressing three polymers across three verticals on owned IP, with a capital intensity per tonne of a quarter to a third of PureCycle's, and a licensing model designed to avoid PureCycle’s exact problem, trades at one-quarter of the enterprise value.

The gap I’m underwriting isn’t “PureCycle is wrong.” It’s that the execution discount applied to Aduro is wider than the category premium it earns if Chemelot works. That’s the trade, and it lives or dies on one plant in Sittard-Geleen.

10. What Would Change My Mind

1. Scale-up failure at the FOAK. If Chemelot can’t reproduce pilot yields at 10,000 t/yr, there is nothing to license.

2. No commercial license signed by the end of 2028. The whole thesis is that partners will pay for this.

3. Repeated dilutive raises at falling prices. Capital discipline is a core reason I own this.

4. FOAK capex overrun beyond ~2x. The capital-efficiency argument is the load-bearing wall of this piece. If it goes, so does the piece.

11. What to Watch

• Aduro’s consolidated NGP validation data package (Q4 2026 window)

• Chemelot permitting via Ebert HERA

• Offtake LOI Phase 1 sample qualification converting to a Phase 2 FOAK parcel commitment

• Naming of the EPC MOU counterparty

• Formal HBU / paraffinic crude pilot announcement

• Delphi LCA and ISCC PLUS certification progress

• PureCycle Q3: whether the P&G ramp moves utilization meaningfully off ~15%, and what materials cost per pound does as it goes

Figures are drawn from PureCycle’s Q2 2026 press release, its June 2026 offering documents and Q1 10-Q, public management commentary on Ironton capacity and cash costs, Aduro’s corporate disclosure, and Roth Capital Partners’ July 27, 2026, initiation. Realized pricing and utilization percentages are my estimates, where PureCycle does not disclose directly, and are labelled as such. Market data as of August 5–6, 2026. Verify against primary filings on SEDAR+ and EDGAR before acting on any of it.

This is not financial advice. I own shares of $ADUR, which represent ~30% of my equity portfolio, so I am biased. Long $ADUR | Not financial advice | DYOR.

Roth Capital initiated coverage of Aduro yesterday with a Buy rating and a $30 target. That makes four independent sell-side firms covering the name.

Where the coverage stands now

Four firms. Four Buys. No holds, no sells. The spread is wide, $22 to $46, and the stock is at $13.40.

Even the lowest target on the street implies roughly 64% upside. The highest implies about 243%. The average of the four sits near $30.70, or roughly 129% above where the shares closed.

That’s the headline. But the headline isn’t the interesting part.

The interesting part is what’s underneath

Wide target ranges usually mean analysts disagree about the business. Here, they don’t. They’re modelling the same two revenue lines, plastics upcycling and paraffinic crude upgrading, and using broadly the same architecture: build-own-operate first, royalty licensing second.

What separates $22 from $46 is mostly timing, multiple, and discount rate. Not a fight over whether the chemistry works.

And when you actually open the models, the common denominator is that every one of them is built on inputs at the low end of the plausible range.

Roth’s numbers

Roth puts circular naphtha at $1,250–$2,050 per tonne against virgin naphtha at roughly $685–$800, a 30–40% premium supported by EU recycled-content mandates phasing in around 2028–2030.

Then they model $1,250. The bottom of their own range. They say so explicitly: conservative.

Same pattern on the margin. They estimate a gross margin of roughly 50% at industrial scale and 50–60%+ at commercial scale, and note that smaller 10K MT facilities likely come in below that, while larger ones exceed it. The model runs 50%.

Same pattern on the oil side. They see a paraffinic royalty model worth $3–$5 per barrel. They model $3, and write plainly that the rate could be higher, but they prefer to stay conservative.

Same pattern on royalty rate for plastics licensing: 5% of revenue.

Now, the penetration assumption

This is the one that should stop you.

Roth’s 2035 estimate has Aduro at 625,000 tonnes per year of total plastics capacity owned plus royalty. Global plastic production is roughly 400 million tonnes a year.

That is about 0.16% of the stream.

On paraffinic crude, they model 30,000 barrels a day by 2035, which they peg at roughly 15–20% of the Uinta Basin’s daily production. One basin. In one state.

From those inputs, Roth gets to CAD$403.1mm of adjusted EBITDA in 2035, discounts it back to 2027 at 15%, applies a 13.3x multiple, and lands on $30.

Ladenburg gets there differently and lands in the same place

Hickman models fiscal 2030 with exactly three assets: one company-owned FOAK plant, one licensed plant, and 50,000 bbl/day of paraffinic crude. He uses $1,500 per tonne, 60% gross margin, and $2.00 per barrel on the oil side. That produces roughly $63.4mm in revenue and $1.42 in non-GAAP cash EPS, which, at 30x, discounted back three years at 20%, gives $24.75.

His prior DCF was even more restrained: 214 licensed plants by 2050, which he notes represents just over 1% of today’s global plastic-waste stream. At a 15% discount rate that produced $19.

Two firms. Different methods, different years, different multiples. Both assume Aduro captures a rounding error of the addressable market.

So what does that actually mean

It means the targets on the street are not blue-sky numbers. They are floor-ish numbers dressed in analyst formatting.

The models assume:

• Naphtha sells at or near the bottom of the observed premium range

• Margins land at the low end of what the unit economics suggest

• Royalty rates come in at the conservative end of the negotiated range

• Aduro captures well under 1% of the plastics market and a slice of a single basin on the oil side

• Meaningful further dilution happens along the way

And they still produce 64% to 243% upside from $13.40.

The bull case for this stock has never required Aduro to win the market. It requires Aduro to work, and then take a sliver. Four independent firms just published models that say exactly that, using inputs they themselves flag as deliberately understated.

What I’m not going to pretend

A few things belong in the same post, not a footnote:

Roth disclosed that it received investment banking compensation from Aduro within the last twelve months. That doesn’t make the research wrong; it’s still independent sell-side research subject to Reg AC, and it’s a different category entirely from paid IR coverage. But you should know it, and I’d rather you hear it from me.

Roth’s own model assumes another $50–$60mm equity raise to fund the FOAK build-out and cover burn through 2030. That’s dilution baked into the $30. Anyone quoting the target without quoting that is selling you something.

These are 2030 and 2035 numbers. Roth has plastics commercialization in 2028 and paraffinic crude in 2030. Ladenburg’s revenue model is for fiscal 2030. Nothing in these targets happens next quarter. The FOAK has to get financed, permitted, built, and commissioned first, and each of those is a real place for this to go wrong.

Price targets are opinions, not floors. All four of these can be cut. That is what analysts do when milestones slip.

What would change my mind hasn’t changed: FOAK scale-up failure, no commercial license signed by the end of 2027, or repeated dilutive raises at falling prices.

But if you’re looking at $13.40 and wondering what the professionals see, they see a company that only has to be modestly successful for the math to work. And they’re not even giving it credit for that.

This is not financial advice. I own shares of $ADUR, which represent ~30% of my portfolio, so I am biased. Long $ADUR | Not financial advice | DYOR.

Amateur Investing published a cost model this week that does what almost nobody in chemical recycling bothers to do: it follows a tonne of mixed plastic waste through every step, pre-processing, conversion, upgrading, steam cracking, polymerization, and asks what the finished polymer costs at the end.

His baseline answer for pyrolysis: €2,968 per tonne, about $3,375. For solvolysis, roughly €2,200.

Watch it here:

It’s his model and his work, independent investor analysis, not company disclosure. Everything below is my reading of it, and any errors in interpretation are mine.

What I want to do is put pyrolysis and Aduro’s Hydrochemolytic Technology next to each other, line by line, at boundaries that actually match, and source every single Aduro number to something published. No hand-waving, no blank cells where I can avoid them.

Step 1: the pyrolysis cost cascade

Each stage carries forward everything upstream, adds its opex and capex, subtracts co-product credits, and divides by surviving mass. I rebuilt it from his inputs. It reconciles.

Cost added at each stage is opex plus capex charge.

€1,698 is the row to remember. That’s the cost of a ton of cracker-ready oil, the same product Aduro sells. Everything past that row is the steam cracker’s business, not the recycler’s.

Step 2: where the plastic actually goes

Before the money, look at the mass. This is the part of the model I can’t stop thinking about.

You feed in 3.67 tonnes of plastic waste. You get back one ton of plastic.

Nearly three-quarters of what goes in doesn’t come out the other end as polymer. To be precise about it: not all of that is destroyed, some leaves as co-products that get sold, and those credits are already netted into the cost figures above. But it doesn’t become plastic again, which is the entire point of the exercise.

And note where the losses sit. 58% of the original waste is gone before it ever reaches the steam cracker. The cracking and polymerization losses below are ordinary petrochemical processing; virgin naphtha loses mass the same way. The 58% is specific to the recycling chain itself. That’s the part that a better technology can attack.

This is why the cost per tonne climbs so violently across the cascade. It isn’t mainly that each stage is expensive. It’s that you keep dividing by a smaller number. Yield isn’t one variable among several in this model; it is the multiplier on every other cost in the chain.

Which is exactly where Aduro’s claim lives.

Run the same chain with Aduro’s published conversion yield of 86%, and with the upgrading step deleted, and the mass balance changes shape: the same 2.39 tonnes of prepped feedstock produces 2.06 tonnes of cracker-ready oil instead of 1.53, 34% more. Carried through cracking and polymerisation at the model’s own ratios, the waste needed per tonne of finished plastic falls from 3.67 tonnes to about 2.73.

That’s my calculation, applying Aduro’s numbers to Amateur Investing’s chain, not his figure and not Aduro’s. But it’s arithmetic, not opinion, and it’s where the cost advantage comes from before a single euro of opex is compared.

Step 3: the side-by-side

This is the comparison. Every Aduro figure is sourced.

Sources. ⁽¹⁾ Amateur Investing baseline model. ⁽²⁾ Ladenburg Thalmann, June 9, 2026 note, FY2030 FOAK model: 40,000 tonnes processed, 85% yield, $1,500/tonne, 60% gross margin. ⁽³⁾ Aduro NGP pilot campaign, June 2026: ~86% liquid hydrocarbon yield, ~85% at C20 and below, 47-hour continuous run. ⁽⁴⁾ Aduro’s own TAM calculation, reproduced in D. Boral Capital’s March 2025 initiation: 90% conversion yield, output priced at €1,200/tonne. ⁽⁵⁾ Aduro press release, November 20 2025, HCT oil run as-produced in a pilot steam cracker, no dilution or pre-treatment. ⁽⁶⁾ My recalculation of Amateur Investing’s cascade stripping capex charges, holding co-product credits constant. ⁽⁷⁾ Implied: $1,500 price × (1 − 60% margin) = $600/tonne of product. ⁽⁸⁾ No published Aduro plant capital model. ⁽⁹⁾ Ofer Vicus, Gabelli Waste Symposium. ⁽¹⁰⁾ D. Boral Capital model, commercial production gross margin. ⁽¹¹⁾ Aduro sells at the plant gate; it doesn’t own the cracker or the polymerisation unit. ⁽¹²⁾ My calculation applying Aduro’s 86% conversion yield and no upgrading step to Amateur Investing’s chain. Converted at EUR/USD 1.137, July 24 2026.

Reading it

The yield row is the foundation. Three independent sources put Aduro at 85% to 90% at the conversion step, compared with 71% for pyrolysis. The pilot measured 86%. Ladenburg models 85%, notably, the analyst assumes no improvement from scale-up, which is the conservative posture. Aduro’s own TAM math uses 90%.

That 14-to-19 point gap is not a rounding difference. It means more product from the same tonne of waste, which lowers the cost per tonne of output. Yield is the multiplier on everything else in the table.

The hydrotreatment row is the structural one. Pyrolysis oil cannot be fed into a steam cracker. It’s an olefin-and-wax mix that fouls furnaces, so it has to be hydrotreated with hydrogen under pressure, using a catalyst, and a second plant with its own capital cost. In the model that box costs €225 per tonne of feed and destroys another 10% of the mass.

Aduro claims to produce cracker-ready oil directly. That received its most meaningful external test last November, when a global steam-cracking licensor ran HCT oil as-produced, no dilution, no pretreatment, with stable furnace operation and olefin yields comparable to those of fossil feed. Delete that box from Amateur Investing’s model, and the finished polymer falls from €2,968 to roughly €2,375, about 20%. That’s my sensitivity on his model, clearly labelled.

The two cost rows need to be read together. On operating costs alone, pyrolysis produces cracker-ready oil at about $997 a tonne. Aduro’s implied cost is about $600, inferred from Ladenburg’s 60% gross margin on $1,500 output, so treat it as an analyst assumption, not a disclosure. That’s roughly a 40% advantage, and it’s the honest version of the number.

Add capital charges and the picture changes character entirely. Pyrolysis full-loaded is $1,931 per tonne. And that is where the table stops being a comparison and starts being a verdict.

The line that matters

Cost to produce a tonne of cracker-ready oil: $1,931.

Price of a tonne of cracker-ready oil: $1,200–1,500.

Three separate sources bracket that price: Aduro’s own TAM math at €1,200, Ofer Vicus publicly at $1,200–1,500, Ladenburg modelling $1,500. They agree.

So, in the baseline case, the pyrolysis chain costs more to run than its product is worth. Not thin margins. Underwater at the gate, before anyone argues about royalties or multiples.

This is also, I think, the real explanation for something that has puzzled people for years: why so much pyrolysis output ends up as fuel rather than as circular feedstock. It isn’t a strategic choice. Fuel has a lower spec bar and doesn’t require hydrotreating. The economics push operators there, and once the output becomes fuel, the circularity claim evaporates, and under EU rules, it stops counting toward recycled content altogether.

Where the table is weak

One cell is genuinely blank. Aduro has published output price, processing cost, yield, and margin. It has not published a plant capital model. Until FOAK is built, nobody, including me, knows Aduro’s full-loaded number. The $600 operating cost is real; the capital charge that sits on top of it is unknown.

86% is a pilot number. Ten kilograms per hour, defined feed, 47-hour run. Pilot yields compress at scale. Anyone treating it as commercial is selling you something.

Pyrolysis isn’t useless. It works on clean single-polymer industrial streams and where policy support is strong. What the model shows isn’t that pyrolysis fails; it’s that pyrolysis aimed at messy post-consumer plastic and genuine polymer-to-polymer circularity has a cost structure that doesn’t close on its own. That’s the narrower claim, and it’s the one that matters, because messy post-consumer waste is where the actual problem lives. PE, PP and PS are more than 70% of municipal plastic waste, and they’re exactly what mechanical recycling can’t touch.

What would change my mind?

1. FOAK scale-up fails. If yield or product spec degrade between 10 kg/hr and 10,000 tonnes/year, the entire table above collapses.

2. No commercial license signed by the end of 2028. A licensing model is validated by licensees, not by pilot data.

3. Repeated dilutive raises at falling prices. Roughly 35M shares, ~32% insider ownership, no founder sales, no debt. Watch whether that holds.

Amateur Investing’s video matters because it puts numbers on something this industry has talked around for a decade. Once you can see the cost stack, you can see exactly which boxes a technology has to delete to be competitive — and you can hold any company, including this one, to account for whether it actually deletes them.

Go watch it:

This is not financial advice. I own shares of $ADUR and it represents ~30% of my portfolio, so I am biased. Long $ADUR | Not financial advice | DYOR.

Amateur Investing’s cost model is its own independent investor analysis, with permission granted. It is not an Aduro disclosure and has not been reviewed or endorsed by the company. Analyst figures cited are the published estimates of Ladenburg Thalmann and D. Boral Capital, respectively, not company guidance; both firms disclose investment banking relationships with Aduro. Aduro figures are from company press releases and recorded public remarks. The opex-only recalculation, the upgrading-step sensitivity, and the implied $600 cost figure are my own calculations, labelled as such. Do your own diligence.

Disclosure: I hold a long position in Rocket Doctor AI, representing approximately 5% of my portfolio. I have not been compensated by the company. Nothing here is financial advice; do your own research. Full disclosure repeated at the bottom.

The company just published its July 2026 investor deck (https://www.rocketdoctor.ai/investors/), and buried in it are the clearest numbers yet on the question that drives this entire thesis: can Rocket Doctor actually convert its US payer footprint into visit volume?

I’ve been writing about this company for a while, and my thesis has never been “the financials look great today.” They don’t. Trailing revenue is small, the company is burning cash, and the stock has drifted from its November highs. My thesis is that the market is pricing $AIRDF off its rearview mirror while ignoring what the operational data is now showing month by month. At a market cap of roughly C$59.5M (about US$43M), call it the low-$40s, I think the mispricing is significant. Here’s why.

1. The US ramp happened with almost nobody on the platform

This is the single most important slide in the deck, and I want you to actually sit with the numbers.

US completed patient visits, per the July 2026 deck:

From 86 visits to over 1,100 monthly visits in four months. The Q1 2026 MD&A (SEDAR-filed) puts hard percentages on it: monthly US completed visits grew 283% during Q1, and April was up a further 69% month-over-month to approximately 1,200 visits, with management stating May is projected to continue the trend.

Now here’s the part the market seems to be missing entirely: that ramp was produced by 22 clinically active US physicians. Per the deck, another 33 MDs are in credentialing right now, with 3x that MD capacity pending in the pipeline.

The US has over one million licensed physicians. I am not arguing that Rocket Doctor needs some meaningful fraction of them; that would be silly. I’m arguing something much narrower: when 22 doctors generate 1,100+ monthly visits within four months of the patient-growth phase starting (December 2025, per the company’s own timeline), the binding constraint on growth is provider capacity, not patient demand. That is the single best problem a marketplace business can have. Demand-constrained platforms have to buy growth. Supply-constrained platforms just have to onboard.

And credentialing is a known, mechanical process: 3–6 months per physician per the deck. The 33 MDs credentialing today are the visit volume of Q4 2026 and beyond.

2. Less than 3% of the payer opportunity is in the numbers you’re looking at

This is stated plainly in the SEDAR-filed Q1 2026 MD&A, not in a promotional slide: the in-network US footprint now spans approximately 21 million covered lives across California, New York, and Maryland, and less than 3% of the projected revenues associated with these payer agreements are reflected in current financial statements.

The deck breaks the footprint down by state: ~9.88M covered lives in New York, ~8.13M in California, ~3.18M in Maryland. Recent additions include a January 2026 contract adding 175,000 members in California and an April 2026 agreement adding ~2 million members in Maryland through a first in-network agreement with the state entity of a major national insurer.

My standard caveat applies, and I will repeat it every single time: covered lives are not revenue. They are an access footprint. Conversion depends on credentialing pace, patient awareness, and visit ramp. But that’s precisely why the monthly visit data above matters so much; it’s the first hard evidence that conversion is actually happening and accelerating.

3. The reported financials lag the operations, by design, of how healthcare billing works

Q1 2026 key financials (deck, unaudited; consistent with SEDAR-filed interims):

• Revenue: C$737,103, up from C$697,340 in Q4 2025, the fourth consecutive quarter of growth (C$512.8K → C$529.1K → C$697.3K → C$737.1K)

• Gross margin: 75%

• ARR: ~US$2.9M (management’s non-GAAP measure, note it includes contracted-but-unrecognized revenue and is stated in USD while IFRS revenue is CAD, so treat the comparison carefully)

Remember the mechanics I’ve written about before: much of the US revenue is Medicare/Medicaid billed on a cash basis, with a 45–90-day lag between the visit and the revenue. The April visit surge doesn’t show up in Q1 numbers. It barely shows up in Q2. The income statement is a delayed echo of the operational data, which is exactly why I watch visits, not just revenue.

The Canadian base, meanwhile, keeps compounding quietly: 47,790 completed visits in Q1 2026, up from 37,176 a year earlier, with 750,000+ lifetime patients seen and 350+ physicians.

4. The moat is boring, and that’s the point

The deck lays out what a competitor would need to replicate: ~1 year of infrastructure setup, 8–15+ months per payer to secure contracts, 3–6 months per physician to credential, 1–2 months for billing setup. Management’s estimate: 2.5+ years to replicate, with no guarantee of securing payers at all.

This is not a sexy AI moat. It’s a regulatory-and-operations moat specific to Medicaid, Medicaid Managed Care, Medicare, and Medicare Advantage, the populations with the most severe access gaps (79M+ Americans on Medicaid and CHIP, 100M+ without a primary physician, per the deck’s cited sources). Every well-funded telehealth startup can build a slick app. Very few have spent 3.5 years grinding through state-by-state payer contracting and credentialing infrastructure. That infrastructure is what the ~US$43M market cap is buying, and I don’t think the market is ascribing it much value at all.

The Global Library of Medicine sits on top of this, 25,000+ clinician-hours of development since 2016, ~10,000+ expert medical reviews, real-time coverage of 1,000+ diseases and 17,000+ symptoms, plus the NIH-funded work (US$2M grant, non-dilutive, running through 2027) continuing via the Rush River Research partnership per the MD&A. I treat the GLM as optionality, not the core thesis. The core thesis is that payer infrastructure is converting to visits.

5. What I’m watching, and what would make me wrong

I want to be explicit about the risks, because that’s the deal we have here:

Cash burn is real. Q1 2026 operating expenses were ~C$4.9M against C$737K revenue. Dilution risk at the micro-cap scale never fully goes away. Fully diluted share count is 135.4M vs. 99.2M basic.

Insider ownership is 4.29%. I’d like that number higher. It’s disclosed right in the deck, and I’m not going to pretend it isn’t a fair criticism.

Small-number effects. Going from 86 to 1,144 visits is a 13x, but it’s 13x off a tiny base. The thesis test is whether the trajectory holds as the base grows. May/June/Q3 data will tell us far more than April did.

Credentialing pace is the whole ballgame. If the 33 in-process MDs and the pipeline behind them stall, the growth story stalls with them. Management says it’s evaluating proprietary pathways to fast-track credentialing; I’ll believe the acceleration when I see it in active-MD counts.

Covered lives ≠ revenue. One more time, for the record.

What would make me wrong: US visit growth flattening over Q2/Q3 despite more credentialed MDs coming online. That would suggest the demand assumption is weaker than the early data implies, and I’d reassess the position.

The bottom line

At roughly US$43M, the market is pricing Rocket Doctor as a struggling Canadian telehealth micro-cap with C$3M of trailing revenue. What the filed MD&A and the operational data actually describe is a company that spent 3.5 years building a US payer and credentialing infrastructure across 21 million covered lives, flipped the switch on patient growth in December 2025, and 13x’d monthly US visits in four months, with 22 doctors.

Less than 3% of the projected payer revenue is in the financials. The credentialing pipeline is 1.5x the current active base, with 3x capacity behind it. Either the visit trajectory breaks down, or the income statement catches up to operations. I’m positioned for the second outcome. I’ve sized the position (5% of my portfolio) so I can be wrong without it ruining me, and the monthly visit data is the scoreboard I’ll be watching.

Disclosure: I am long Rocket Doctor AI (CSE: AIDR / OTCQB: AIRDF), representing approximately 5% of my portfolio. I may buy or sell shares at any time without notice. All figures are sourced from the company’s July 2026 investor deck, SEDAR-filed Q1 2026 interim financials and MD&A, and third-party analyst coverage (FRC, which is paid coverage; Maxim Group, which makes a market in the stock, both conflicts are disclosed in their reports). Management-stated figures are labelled as such. Micro-cap stocks are illiquid and high-risk. This is not financial advice. Do your own research.

Most micro-cap catalysts are events a company is subjected to. This one is different: The Precision Peptide Company spent the last eight months building for a specific July date, and on July 23, that date arrived with the outcome it needed.

That’s the setup. Whether it’s an opportunity depends on something narrower — whether the business underneath the catalyst is structurally hard to copy, or just early. Here’s my read on both.

The thesis: vertical integration in a category that has almost none

The default peptide competitor imports API powder, frequently from China, filters and vials it domestically, and calls it American-made. The label reads USA; the molecule doesn’t.

The Precision Peptide Company's claim is control of the entire chain: from synthesis, purification, verification, to delivery. Four links, as laid out by CSO Justin Kirkland on Dave Asprey’s The Human Upgrade (June 16):

Synthesis. Chemically assembling the amino-acid chain in the US rather than inheriting whatever a supplier shipped.

Purification. A TFA-to-acetate salt exchange strips the trifluoroacetic acid residue that solid-phase synthesis leaves behind. Purely chemical synthesis also sidesteps the LPS endotoxin that E. coli– and CHO-derived peptides can carry. Both are real problems in the gray market, and both are checkable independent of anything the company says.

Verification. HPLC and mass spec on every batch. Miraxis Labs independently returned 95.5% purity on the BPC-157 patch material, a primary artifact, not a marketing number.

Delivery and distribution. Auto-injector pens and dissolving microneedle patches that address the two adoption barriers keeping peptides in the niche: needle phobia and reconstitution math. Sold through a physician-consult telehealth loop and 503A compounding, not gray-market forums.

Where the buyer can’t verify what’s in the vial, “we control the chain and can document it” is the asset. Add LegitScript certification (June 23), which unlocks the Meta and Google ad platforms that gray-market sellers are banned from, plus US cGMP manufacturing and a 50-state prescribing network, and the shape of this becomes clear.

This is a commercialization bet. The rails are built. It was waiting on the regulatory door.

I wrote about The Precision Peptide Moat: What Kirkland Told Asprey and What the Deck Still Says ‘Under Development” here:

Why now

On July 23, the FDA’s Pharmacy Compounding Advisory Committee reviewed BPC-157, PNGAF’s lead molecule and the basis of its flagship product, for the 503A Bulks List, and the read is favourable.

That matters because it moves the regulatory direction from gray-market tolerance toward a legal pathway. It is precisely the environment the company pre-built: the 503A agreement, the telehealth network, the LegitScript clearance, and US manufacturing. None of that infrastructure made sense in a world where these peptides stayed permanently outside the compounding framework. In a world where they’re moving inside it, all of it does.

The larger peptide-exposed names, , are already reacting. Illiquid micro-caps typically lag the initial move by days, not minutes. That lag is where the opportunity sits.

The risks that actually matter

PCAC is advisory, not law. A favourable vote is a recommendation. FDA staff briefing documents recommended against all seven peptides; the panel drew AP scrutiny over members with industry ties; and formal 503A rulemaking typically runs 12–24 months. “Recommended” is not “legal to compound tomorrow.” Treat this as direction de-risked, not a finish line crossed.

Compounding is not synthesis. The public disclosures describe 503A compounding and US fulfillment more clearly than they describe in-house chemical synthesis. Those are different claims, and the synthesis link is the most load-bearing and least documented of the four. I’d want it evidenced before pricing it.

Pre-revenue and dilutive. FY2025 revenue was ~CA$50k. Cash is ~$3M following the $2M placement, which closed on July 15. Fully diluted share count is 54.7M+ and climbing. This company needs more capital, and scaled execution is unproven.

Micro-cap mechanics. Thin float, OTC listing, sentiment-driven. Moves both ways violently.

Where this leaves me

A tiny valuation, genuine, if replicable, infrastructure, first-mover positioning, into a regulatory tailwind that just turned favourable.

That’s an asymmetric setup. But I want to be precise about where the asymmetry actually lies: in the small-cap and execution-window markets, not in an IP moat, and not in any claim that peptides are now legal.

One more piece of precision, because I’ve written about this company before, and a reader deserves consistency. When I covered the Kirkland interview, I said the two things that would close the gap between the story and the price were the July PCAC meeting and the first real patch revenue. One of those has now happened. The other hasn’t. On July 23, the regulatory variable was moved. It did not move the execution variable. The patch still has to ship, the synthesis claim still needs documentation, and the next twelve months still need funding without crushing the share count.

That’s the follow-through I’m watching. I’m long the setup with my eyes open on all three.

Sources

• Mariusz Skonieczny, In 1 Week — Everything Changes for the Precision Peptide Company: mariuszskonieczny.substack.com

• The Human Upgrade with Dave Asprey featuring Justin Kirkland, CSO: youtube.com/watch?v=1JD8FizxJEA

• FDA Pharmacy Compounding Advisory Committee, July 23–24, 2026 meeting materials

Disclosure: I hold a position in $PNGAF. This is my own opinion, not investment advice. Statements attributed to Justin Kirkland reflect his characterizations in the linked interview and are the company’s claims, not independently verified by me. Forward-looking product timelines, including the BPC-157 transdermal patch, are described by the company as under development and are subject to change. Share structure and market cap figures should be re-verified against a current quote. Do your own diligence; this is a speculative, pre-revenue micro-cap, and you can lose your entire position.

Something interesting is happening beneath the surface of Aduro’s tape, and it’s worth spelling out plainly: a meaningful short position has been built into a name that is stacking catalysts — and that combination has a way of resolving violently in one direction.

The Numbers

Per the latest NASDAQ short interest report:

• Short interest: 2,119,673 shares

• Short interest as % of float: 8.82% (NASDAQ short interest / Capital IQ float)

• Days to cover: 2.35

Add to that the daily tape: on July 20 alone, 414,734 shares, 61.35% of off-exchange volume, were marked short per FINRA (including dark pool venues), following 194,994 on July 17 and a heavy 643,980 on July 16. A caveat for the rigorous reader: daily short volume includes market-maker hedging flow and isn’t a one-to-one proxy for positioning. But taken alongside a hard 8.82%-of-float short interest figure, the picture is coherent, sellers keep leaning on this stock, and someone is on the other side of every one of those trades, obligated to buy back eventually.

Why 8.82% of Float Matters Here Specifically

On a mega-cap, 8.82% short interest is background noise. On a small-cap specialty chemicals company with a tight float, a concentrated long-term shareholder base, and a stock that just printed +6.87% in a single session — it’s ammunition.

Here’s the mechanical reality: 2.1 million shares must be repurchased at some point. Days-to-cover at 2.35 sounds manageable in calm conditions. But days-to-cover is calculated on average volume. When a binary positive catalyst hits a small-cap, average volume goes out the window, spreads widen, and shorts who wanted an orderly exit discover there isn’t one. The exit door on small-caps is narrow in both directions.

And this is not a company short-sellers should want to be leaning on into the second half of 2026.

The Milestone Stack

The short thesis on Aduro, to the extent it can be reconstructed, is that the commercialization narrative stalls. The problem is the calendar is working against that bet:

1. The GEPC EPC framework. The MOU establishing Aduro’s “master franchiser” licensing architecture is the structural unlock, it converts Aduro from a single-plant science story into a licensing and royalty model with a Tier-1 engineering partner attached. Every incremental step in that relationship maturing (FEED work, first-licensed deployment scoping) is a headline shorts have to absorb.

2. Chemelot / Ebert HERA permitting progress. The permitting contract at Chemelot puts Aduro’s technology on a path to a flagship European deployment inside one of the continent’s premier chemical clusters. Permitting milestones are exactly the kind of “boring but binary” de-risking events that re-rate small-caps quietly — until suddenly it isn’t quiet.

3. The paraffinic crude CIP patent expansion. The bench-scale results on Uinta Basin waxy crudes open an entirely separate addressable market beyond plastics — upgrading problem crudes. This is optionality the market is arguably paying nothing for at current levels.

4. Analyst targets far above the tape. H.C. Wainwright carries a $22 target; D. Boral Capital has published work in the $46–50 range. Whatever you think of sell-side math, the gap between $14.32 and the lowest published target is the runway shorts are standing on.

The Setup

Squeezes don’t happen because short interest is high. They happen because short interest is high, and a catalyst forces a repricing that turns short covering from a choice into a necessity. Aduro today has:

• A hard, verifiable 8.82% of float sold short

• Persistent daily shorting pressure that has failed to keep the stock down (+6.87% today, against 61% short volume, read that again)

• A multi-catalyst calendar in H2 2026, where any single positive milestone forces the question: what exactly is the short thesis now?

When a stock rises nearly 7% on a day when the majority of off-exchange volume is marked short, it tells you demand is overwhelming supply, even amid selling pressure. Remove the pressure, or worse, force it to reverse, and the supply-and-demand math gets asymmetric fast.

The shorts have loaded the gun. The next positive milestone pulls the trigger.

Not investment advice. Do your own due diligence. Small-cap securities are volatile and illiquid; short interest data is reported with a lag, and daily short volume figures include market-making activity.

Disclosure: standard §17(b) / NI 51-102 disclosure block author holds a long position in Aduro Clean Technologies

I want to walk through his framework here, with full credit to him for the structure and the math, because I think the exercise itself is valuable even if you disagree with every input. All of the vertical-level assumptions below are his; the commentary on where I’d push back or add caution is mine. His full article is worth reading in the original, and I’d encourage you to do that before forming a view.

One framing point before the numbers: this is a long-horizon, everything-works scenario. The author is upfront about that, and so am I. Aduro is pre-commercial. The plastics vertical, the furthest along, doesn’t see meaningful commercialization until 2028 at the earliest, and most of the other verticals trail it by a year or more.

The framework: one platform, six businesses

The core idea will be familiar to anyone who’s followed my writing: Aduro isn’t a single-product cleantech company. Hydrochemolytic™ Technology (HCT) is a platform chemistry, a water-based catalytic process that breaks down hydrocarbon-based materials, and any material with a hydrocarbon backbone is, in principle, addressable. Plastics, heavy oil, waxy crude, renewable oils, cross-linked polymers, and synthetic turf.

Where the 10klist’s piece goes further than most analysis (including my own three-vertical licensor model) is in splitting the opportunity into six distinct segments and valuing each one independently:

Rounded to $2,000 “because it sounds cooler,” in his words. Fair enough.

Vertical 1: Plastics — the anchor

His plastics model assumes a licensed plant processes 25,000 tonnes per year at a 95% conversion rate, with naphtha selling at $1,500/tonne, feedstock at $150/tonne, and ~$312/tonne in operating cost, roughly $1,038 of gross profit per tonne, or about $24.6M of gross profit per plant. Scale that across 75 deployed plants and apply a 12x multiple, and you get $22.2B. His low case ($900/tonne naphtha) still produces an $11B segment value.

This is the vertical with the most real-world support. The FOAK plant at Chemelot is progressing, the NGP Pilot Plant’s June campaign delivered ~86% liquid yield in continuous operation, the offtake LOI with a global commodities trader includes a binding carve-out on initial FOAK production, and the third-party steam cracker trials confirmed HCT oil ran as-produced, no dilution, no hydrotreatment, with olefin yields comparable to fossil naphtha. That last point is the one I keep coming back to, because it’s the structural difference between HCT and the pyrolysis field: pyrolysis oils typically need hydrotreatment and low blend rates before a cracker will take them.

My caution: 75 plants is a full build-out scenario, not a forecast. His own sensitivity table runs from ~$21/share (25 units, low margins, 2x multiple) to ~$759/share, which shows how wide the range of outcomes is.

Vertical 2: Bitumen and heavy oil — the sleeper

This is the largest single segment in his model, and it borrows Amateur Investing’s framework (credit where due, this is Tier 2 community analysis, not company guidance). The value-per-barrel logic: partial upgrading lifts the selling price by an estimated $11–15/bbl, eliminating diluent and transport costs saves another $10–13/bbl, and HCT processing costs run $5–8/bbl, netting producers roughly $13–23 of incremental value per barrel. His model then assumes Aduro eventually captures 15% of an 8.3M bbl/d global heavy crude market at a 20% licensing take, yielding ~$1.3B in annual licensing profit at a 20x multiple: $26.4B.

The commercialization signals here are real: Scott Smith (17 years at Cenovus) joining as Program Director of Petroleum Technology Solutions, the Prospera pilot collaboration, and the January 2022 independent validation lifting Alberta bitumen from 14.6 to 19.5 API in continuous flow. But I’d flag two things. First, management has indicated HBU requires its own pilot program; this vertical is not riding the Chemelot plant’s coattails. Second, 15% global penetration is an enormous assumption for a technology that hasn’t yet run a dedicated heavy-oil pilot. His own framing calls the scenarios “highly speculative,” which is the right posture.

Vertical 3: Waxy and paraffinic crude, the potential fast mover

The Uinta Basin problem is easy to understand: the crude is so paraffinic that it behaves like peanut butter at ambient temperature, forcing producers to use heated trucks and insulated railcars and capping how much can be economically extracted from the basin. In April, Aduro announced that HCT had been extended to paraffinic crudes, that bench-scale testing on actual Uinta yellow and black wax produced an ambient-stable, lighter product, and that it filed a continuation-in-part patent application. The Utah Petroleum Association membership followed in May.

His valuation: producers gain up to ~$15/bbl, Aduro captures 25% of that value (~$3.75/bbl), applied across current Uinta production, generating ~$250M in licensing revenue and ~$200M in net profit at 80% margins. At 20x: $4.0B.

What I find genuinely interesting is Ofer’s own framing in the July Penny Queen interview, that the paraffinic application is chemically an extension of the plastics work and that the company wants to move quickly on it. From an execution standpoint, there’s a real argument that this vertical de-risks faster than plastics: wellhead crude is a homogeneous feedstock, and the offtake infrastructure (pipelines, refineries) already exists. No circular-economy supply chain needs to be built.

Where I’d hedge harder than the article does: the suggestion that a major Utah oil company prompted the testing comes from Mariusz’s commentary, not from any Aduro disclosure. Treat it as informed speculation until a named producer, paid program, or MOU appears. That said, this is exactly the kind of catalyst that could re-rate the stock well before 2028, and I agree with the author that assuming no re-rating before plastics commercialization is a mistake.

Vertical 4: Renewable oils, SAF, and BTX have the biggest TAM, with the least evidence

Aduro’s own materials put the renewables TAM at ~$121B by 2030, the largest of the three verticals in the corporate deck. The thesis: HCT can use low-cost biogenic hydrogen donors instead of expensive externally supplied molecular hydrogen, potentially undercutting hydrotreating economics for SAF and renewable diesel, and a September 2024 patent filing covers converting renewable oils and certain waste plastics into BTX aromatics.

His high case assumes 5.45% market penetration, a 7% royalty on HCT-enabled product value, and a 21x royalty revenue multiple, totalling $9.7B. And he uses that high case in the headline number.

This is the segment where I’d apply the biggest haircut. Aduro itself classifies renewables as “advanced research,” and no named partner, MOU, or paid test program has been disclosed for SAF or renewable diesel. In my own valuation work, I’ve excluded HRU entirely and treated it as free optionality, and I still think that’s the honest way to hold it. A 21x revenue multiple on a vertical with no disclosed commercial engagement is a bet on the platform thesis, not on evidence.

Verticals 5 and 6: Thermosets and synthetic turf, small numbers, real signals

These are the smallest segments in his model ($2.0B and $440M respectively, on simple EBITDA scenarios at 20–22x), and appropriately so. But they’re worth watching for a different reason: both show the pattern investors should want to see from a platform company, technical result, industrial interest, and paid engagement.

On cross-linked polymers: HCT achieved up to 84% yields of lower-molecular-weight hydrocarbons from XLPE waste in 2024 testing, Georg Fischer signed an MOU that November covering Uponor-branded PEX waste, and by March 2026, that had progressed into a paid test program. On turf: successful testing on real end-of-life turf in July 2025, followed by the June 2026 AstroTurf MOU to evaluate PE and PP recovery. Note that the AstroTurf agreement is non-binding and evaluation-stage, a signal, not a contract.

The repeatable motif matters more than the dollar figures: difficult feedstock, promising chemistry, named counterparty willing to pay. That’s the platform thesis showing up first in the smallest verticals.

Where I land

I won’t pretend $2,000/share is my base case, it isn’t. My own licensor model is based on much lower penetration assumptions across three verticals, and I’d rather be surprised to the upside. The value of the 10klist’s exercise is different: it maps the full option surface. Analysts today model essentially one vertical (plastics) at sub-1% penetration. Everything else, HBU, paraffinic crude, HRU, thermosets, turf, is optionality the market is pricing at roughly zero. His article is the most complete attempt I’ve seen to put numbers on all of it at once, and for that, it deserves a wide readership.

For balance, my standing “what would change my mind” framework applies to every scenario in this post: scale-up failure at the FOAK plant, no commercial license signed by the end of 2028, or repeated dilutive raises at falling prices. None of those has happened. The June financing was done at a premium structure with no warrants; the balance sheet shows no debt, and management (which holds roughly a third of the company) has been buying, not selling. But those are the tripwires I watch.

Go read the original article, watch Mariusz’s chemistry series, work through Amateur Investing’s models, and check every number against Aduro’s filings yourself.

This is not financial advice. I own shares of $ADUR, which represent ~30% of my portfolio, so I am biased. The valuation frameworks discussed above are credited to @the10klist, Amateur Investing, and Mariusz Skonieczny; they are independent investor analyses, not Aduro company disclosures. DYOR.

When a discovery-stage micro-cap puts its Chief Scientific Officer on one of the largest platforms in the longevity space, the interesting question is never “did they get exposure.” It’s whether the conversation reveals a structural reason the business is hard to copy or just another well-produced pitch.

Justin Kirkland’s appearance on Dave Asprey’s The Human Upgrade (released June 16, 2026) is worth parsing on exactly those terms. Strip away the biohacking framing and the interview is really an argument about one thing: that the peptide market is a low-trust, quality-fragmented “wild west,” and that whoever controls the chain from synthesis → purification → verification → delivery owns something the fill-and-vial crowd can’t easily replicate.

That’s the moat thesis. Below is the moat as Kirkland articulated it — and, because this is coverage and not a company deck, the parts that are verified science versus company claims, plus the facts that cut the other way.

Interview link: The Human Upgrade with Dave Asprey — Justin Kirkland (YouTube) (also on Apple Podcasts, Spotify, and daveasprey.com)

The moat, as told

1. US-based chemical synthesis does not fill-and-finish of imported API

Kirkland’s central distinction: most companies marketing “American-made” peptides are buying the active pharmaceutical ingredient as a powder from overseas, frequently China, then filtering it and vialing it domestically. The label reads USA; the molecule doesn’t. (~15:57)

The claimed differentiator is that Precision Peptide performs the actual chemical synthesis, physically assembling the amino-acid chain, inside the US, retaining oversight of the full process rather than inheriting whatever a supplier shipped. (~16:41)

Reader’s note: This is the load-bearing claim of the entire moat, and it’s the one to hold loosely until independently confirmed. It sits in some tension with how the company describes itself elsewhere, a 503A compounding agreement under which peptides are “100% compounded in the United States,” plus a third-party telehealth/fulfillment partner. Compounding and fill-finish are downstream of synthesis. “Manufactured in an approved U.S. facility” and “we chemically synthesize our own peptides in-house” are not the same statement, and the public disclosures lean toward the former. Worth pinning down before leaning on the synthesis claim in any valuation framing.

2. Delivery systems built to remove the two biggest adoption barriers

The two frictions that keep peptides niche are needle phobia and dosing math (reconstituting powder with bacteriostatic water and calculating micrograms). The interview frames two answers:

• Auto-injector pens are spring-loaded, EpiPen-style, pre-loaded at standard concentrations. The user dials a prescribed dose and administers it without seeing a raw needle or doing the math. (~08:00, ~09:36)

• Dissolving microneedle patches, clear cosmetic patches with micron-scale needles, are described as feeling like a cat’s tongue on thin skin. Peptides are embedded in natural polymers (e.g., hyaluronic acid) that dissolve into interstitial fluid and enter systemically over ~30–60 minutes. (~44:21, ~45:12)

Reader’s note: this is the single most important caveat in the piece. The needle-free BPC-157 transdermal patch is the company’s flagship delivery method, and its June 2026 investor deck labels it “Currently Under Development.” It is showcased at conferences and described in interviews as a capability; it is not yet established as a shipping, revenue-generating consumer product. The existing product line purchasable today (Ageless, Blue, BPC, Chill, KPV, plus Hair and Skin) is oral/topical, not the patch that anchors the “we own the delivery layer” thesis. The gap between operational readiness and live revenue on the differentiating product is the core thing to watch, and obscuring it would be the fastest way to make this coverage read as promotional.

3. Purification and toxin removal

This is the strongest part of the interview on the merits, because the chemistry is real and checkable independent of the company:

• TFA → acetate salt exchange. Solid-phase peptide synthesis grows chains on resin beads, which are cleaved with trifluoroacetic acid (TFA), a residue that is genuinely worth removing. Kirkland describes a rigorous salt-exchange step converting residual TFA salt to a benign acetate salt. (~18:32, ~19:43, ~20:34)

• No LPS from biological manufacturing. Peptides made in engineered E. coli or CHO cells can carry lipopolysaccharide (LPS) endotoxin byproducts. Purely chemical synthesis sidesteps that class of contaminant. (~20:52, ~22:37)

The category-level point is accurate: TFA residue and LPS are real problems in the gray market. The company-specific point that Precision Peptide’s process reliably eliminates them batch-to-batch is a quality claim that only third-party COAs can substantiate, which is precisely why pillar 4 matters.

4. Telemedicine plus third-party analytical verification

• Telemedicine loop. Rather than routing customers through unregulated forums, the model pairs a physician consultation with fulfillment through legitimate compounding pharmacies. (~09:13)

• Batch testing. Every batch is described as run through HPLC and mass spectrometry, so a stated purity (e.g., 99.9%) is backed by analytics rather than a marketing number. (~46:10)

Third-party validation is where a trust-based moat becomes real or falls apart. The prior externally validated data point, the Miraxis Labs 95.5% purity result on the BPC-157 patch material, is the kind of primary artifact that carries more weight than any interview claim. More published COAs, not more podcasts, are what would harden this pillar.

The synthesis of the thesis: vertical integration

Put the four pillars together, and the moat Kirkland is describing is vertical integration in a category that mostly doesn’t have any control from domestic synthesis, through documented purification, through analytics, into delivery formats that remove the friction keeping peptides niche. In a market where the default competitor is repackaging someone else’s powder, “we control the whole chain and can prove it” is a genuinely defensible position if the claims hold at each link.

That “if” is the entire investment question.

The facts that cut the other way

Coverage that only relays the bull case isn’t coverage. Four things belong in the same frame as the moat:

The differentiating product isn’t shipping yet. Again: the patch is “under development” by the company’s own admission. Today’s revenue does not come from the delivery layer on which the thesis is built.

Compounding is not synthesis. The public disclosures describe compounding and US fulfillment more clearly than they describe in-house chemical synthesis. Until the synthesis claim is documented, the “we make the molecule ourselves” pillar is the least verified and the most load-bearing.

Dilution is live. On June 11, 2026, the company announced a non-brokered private placement of up to 4,000,000 shares at $0.50 (up to ~$2.0M gross). Against the deck’s fully-diluted count of ~54.7M shares (43.6M basic + 8.4M warrants + 2.7M options, as of June 2026), any raise plus attached warrants moves the denominator. This is a company funding itself through the pre-revenue phase of its flagship product, and the share count should be re-verified against a published quote, not the deck.

Where the signal vs. price gap sits

The bull framing writes itself: an under-followed micro-cap assembling a vertically integrated quality story in a category the GLP-1 wave has made mainstream, validated by a marquee platform appearance. The disciplined framing is narrower, the moat is real in architecture and still unproven in execution. The patch is pre-launch, the synthesis claim is under-documented, the balance sheet is being topped up through dilution, and the delivery layer already has a funded challenger.

Comparables like Hims & Hers, Ro, and AgelessRx define the ceiling of what a consumer-health peptide platform can be worth; they are category context, not a price target for a company whose flagship product hasn’t shipped. The interview raised the ceiling of the story. It didn’t close the distance to it. The July 23 PCAC meeting and the first real patch revenue are the two events that would.

The Human Upgrade with Dave Asprey featuring Justin Kirkland (CSO): youtube.com/watch?v=1JD8FizxJEA

Disclosure & disclaimer

I hold a personal position in The Precision Peptide Company (CSE: BPC | OTCQB: PNGAF). This is not investment advice. Nothing here is a recommendation to buy or sell any security. Statements attributed to Justin Kirkland reflect his characterizations in the linked interview and are the company’s/speaker’s claims, not independently verified by me. Product timelines, including the BPC-157 transdermal patch, are forward-looking and described by the company as under development; forward-looking placement, purity, and commercialization claims are per company releases and the linked interview and are subject to change. Share-structure and market-cap figures are drawn from the company’s June 2026 materials and should be re-verified against a current quote before relying on them. Do your own research and consult a licensed professional before making any investment decision.

They never were. They were scaffolding, and rescheduling just started dismantling it.

Hirsh Jain, one of the sharpest policy minds in cannabis and a consultant to LEEF Brands, published a piece this week that everyone in this sector should read: Interstate Commerce After Rescheduling: Why the Debate Has Changed. Full credit to Hirsh, this is his framework, and I’m going to walk through why I think it matters enormously for how we position from here. Go read the original. Then come back.

The State-Siloed Market Was Never the Law, It Was a Workaround

Here’s the part most investors miss, and Hirsh nails it: Congress never mandated that cannabis markets be sealed inside state lines. The siloed structure emerged as a defensive response to the 2013 Cole Memo, which flagged interstate diversion as a federal enforcement trigger. States built import and export bans into their programs to stay off the DOJ’s radar — and kept operating that way even after Sessions rescinded the memo in 2018. A “shadow Cole Memo,” as Hirsh puts it.

The legal shield protecting those state walls was always the same: cannabis sat in Schedule I, so federal courts held that the Dormant Commerce Clause — the constitutional doctrine that stops states from protectionist discrimination against out-of-state goods, simply didn’t apply. You can’t invoke the Constitution’s protection of commerce for a product that federal law says has no lawful commercial channel.

Schedule III breaks that premise. A substance with federally recognized medical use and a DEA registration pathway is very hard to characterize as sitting wholly outside lawful commerce. Hirsh points to the Ninth Circuit’s June oral arguments in Lamar Central Outdoor v. City of Perris, where the panel showed visible skepticism toward arguments built on cannabis being categorically federally illegal. Judicial thinking is shifting in real time.

Three Pathways And They Compound

Hirsh lays out three credible routes to interstate commerce, and the key insight is that they’re not mutually exclusive:

1. State legislative reform. States wrote these bans; states can repeal them. New medical markets may prefer importing products over waiting years to build cultivation infrastructure. Mature markets may want lower prices and better access to squeeze the illicit market.

2. Dormant Commerce Clause litigation. The challenges that failed for a decade now face a fundamentally different legal environment. Litigation is slow, years, not quarters, but the underlying theory just got materially stronger.

3. Interstate compacts. This is the one I’d watch most closely. California’s SB 1326, passed in 2022, authorizes the state to enter cannabis trade agreements with other jurisdictions and can be triggered without Congress doing anything. DOJ guidance, favourable court rulings, or even a determination by California’s own Attorney General that the legal risk is manageable would suffice. Oregon, Vermont, and Washington have similar statutes on the books. The trigger conditions that once looked theoretical now look live.

The realistic path isn’t one dramatic federal moment. It’s the alcohol playbook post-Prohibition: incremental, state-by-state, compact-by-compact, with each development reinforcing the next. Hirsh is appropriately careful on timing; the FD&C Act remains a genuine open question, regulatory harmonization will take time, and entrenched incumbents in protected markets will fight this. But the direction of travel has changed. That’s the point.

Why This Is the $LEEEF Setup

Now connect the dots to positioning, because this is where I think the market is asleep.

When state walls come down, the industry stops rewarding presence and starts rewarding cost structure and scale. The multi-state operator model, duplicating expensive cultivation and manufacturing footprints in every state to comply with the silo rules, becomes a stranded-cost problem. The winners are the low-cost producers in the states best positioned to export.

And no state is better positioned to export than California. Ideal growing climate, the deepest cultivation talent pool in the world, the largest legal market, and, critically, SB 1326 already sitting on the books waiting for a trigger.

LEEF Brands ($LEEEF) is California’s largest vertical extractor. Salisbury Canyon Ranch provides LEEF with a scaled, low-cost biomass supply to feed its extraction platform. Today, that asset base is priced against a single state market of ~40 million people. In an interstate world, even a partial one, even a two-state compact, that same asset base gets optionality on a continental market.

That’s the asymmetry. The downside case is the status quo: LEEF keeps executing as the dominant extractor in the world’s largest single cannabis market. The upside case is that California becomes the export engine of American cannabis, and the state’s biggest low-cost extractor is standing at the loading dock when the doors open.

I’m not going to pretend the timeline is knowable, Hirsh explicitly doesn’t, and neither should anyone else. But investing in small caps is about identifying where the legal and structural landscape is moving before consensus reprices it. Rescheduling moved the landscape. The market hasn’t moved yet.

Read the Original

Again, full credit to Hirsh Jain for the legal framework here. His piece is the most clear-eyed treatment of the interstate question I’ve read since rescheduling, and it deserves your full attention: Interstate Commerce After Rescheduling: Why the Debate Has Changed.

Disclosure: The author holds a long position in LEEF Brands securities. Hirsh Jain is a consultant to LEEF Brands. This publication is provided pursuant to Section 17(b) of the U.S. Securities Act of 1933 and applicable Canadian securities regulations, including NI 51-102. Nothing in this article constitutes investment advice, a recommendation, or an offer to buy or sell any security. The author may buy or sell securities mentioned at any time without notice. Cannabis-related investments involve significant legal, regulatory, and market risk. Always conduct your own due diligence and consult a licensed financial advisor before making investment decisions.

Watch the full interview here:

The execution recap

The interview opens with a run-through of what Aduro has actually shipped over the past few months: a confirmed FOAK site at Chemelot Industrial Park in the Netherlands, the strongest NGP pilot data to date, MOUs with Ortessa Group and now AstroTurf, a TSX uplisting that went live May 27, and a capital raise completed at a premium to market with zero warrants attached. Vicus’s framing was simple: this isn’t a lucky streak, it’s the plan they told investors about at the start of the year, executed on schedule.

For a microcap, that last sentence is the whole ballgame. Plans are cheap. A checklist of delivered milestones matching a previously stated roadmap is not.

The NGP numbers, and what Vicus says they actually mean

The headline data: roughly 86% liquid hydrocarbon yield from the NGP pilot plant running on recovered polypropylene, with about 73% of the feed ending up as cracker-ready output. Per Aduro’s June 9 press release, the plant ran continuously for 47 hours (Vicus rounds to 48 in the interview), held steady state for roughly 35 of those hours, and recovered from a deliberately introduced disturbance in about two hours. The 73% figure reconciles cleanly with the disclosed data: ~86% liquid yield, with ~85% of that liquid falling at C20 and below, the naphtha-range cut steam crackers actually want.

Two things Vicus said here matter more than the numbers themselves.

First, he was explicit that the NGP is not a go/no-go experiment. The conversion chemistry, Hydrochemolytic Technology, a catalytic, water-based process operating below supercritical conditions, not pyrolysis, was proven years ago at bench and continuous-flow scales. The NGP exists to generate continuous operating data for commercial-scale engineering. That reframing matters: the question being answered isn’t “does it work,” it’s “what are the design parameters for the plant we’re about to build?”

Second, and this is the part I found most interesting: the 86% run happened on a unit that Vicus says is not yet fully optimized. The team is still tweaking valves, pipelines, and heating zones. Whether further optimization moves the yield number materially is something we’ll only know from future disclosures, but hearing a CEO say the best run to date came from an unfinished tuning process is a different message than “we squeezed out our peak result.”

The usual qualifiers apply, and I’ll keep applying them: this was one feedstock (clean recovered PP), at pilot scale. Mixed, contaminated post-consumer results are a separate proof point I’m watching.

The FOAK philosophy: deliberately imperfect

My favourite section of the interview. Vicus openly stated that the first-of-a-kind commercial unit at Chemelot won’t be perfect, by design. Aduro plans to use off-the-shelf components wherever possible to control cost and keep the unit easy to modify, fully accepting that generation-two and generation-three plants will be more efficient than the first.

This sounds unglamorous, and that’s exactly why I like it. The graveyard of chemical recycling is full of companies that bet everything on a bespoke, capital-heavy first plant and then discovered scale-up problems with no budget left to fix them. Designing the FOAK as a learning machine rather than a showpiece is the capital-disciplined version of this journey. It also fits the modular thesis: Aduro’s scale-up path is replication of ~10,000 t/yr units, not a leap to a mega-reactor, which sidesteps the heat-transfer wall that punishes pyrolysis scale-ups.

Why Chemelot, really

The conventional answer is regulatory: Europe’s mass-balance framework and recycled-content mandates (EU PPWR becomes generally applicable August 12, 2026) reward high-yield chemical recycling and make life difficult for lower-yield pyrolysis routes. Vicus acknowledged that, but said the primary driver was the ecosystem, being physically embedded among service providers, civil engineers, and potential customers who specialize in plastics and petrochemicals. When you’re a small company building your first industrial plant, having your supplier and customer network within walking distance is worth more than any single incentive.

AstroTurf and the B2B waste-stream strategy

The AstroTurf MOU (June 30, 2026) got real airtime. The market Vicus cited is striking: over 275 million square feet of end-of-life synthetic turf in North America alone, and it’s a stream that mechanical recycling essentially cannot touch: PE and PP fibres embedded in cured polyurethane backing, mixed with sand, rubber infill, and adhesives.

Two strategic points from the interview are worth underlining:

The first is sequencing. Aduro is deliberately targeting unified, consistent B2B waste streams (turf, agricultural film, XLPE pipe) before attacking the chaos of mixed post-consumer waste. Each of these streams has a known composition, a motivated counterparty, and a disposal cost problem, which makes the commercial conversation far simpler than “bring us your municipal garbage.”

The second is deployment geometry. Because turf is consistent, Vicus floated the possibility of smaller, configurable, dedicated HCT units deployed close to where turf is collected, flipping the logistics problem of hauling heavy, bulky turf across the continent into a distributed processing model. That’s a natural expression of the modular architecture, though I’d stress it’s a concept at this stage, not a commitment.

And the necessary honesty: this is an MOU, not a commercial contract. The press release explicitly states that it establishes no commercial deployment terms, and any licensing, supply, or tolling arrangement would require separate definitive agreements. It’s a pipeline entry, not revenue.

The quiet petroleum playbook

While most of the market has been focused on plastics, Vicus confirmed the parallel track in petroleum: testing Uinta Basin yellow and black wax crudes. His framing was that highly paraffinic crude is actually a simpler application of the same chemistry as waste plastic, the feedstock flows when heated and contains none of the random contamination of post-consumer waste (his phrase: no “cats, dogs, and chicken bones”), so it doesn’t even require extruders.

The velocity claim was notable: because years of hard engineering went into the plastics process, designing the petroleum technology demonstration unit took only 2 to 3 months. The next step is to finish that demonstration unit and scale it into a localized pilot campaign processing 5–10 barrels per day to prove commercial viability to oil producers.

My qualifiers, as always on this vertical: everything published so far on paraffinic crude is bench-scale (the April 23, 2026, continuation-in-part patent filing, the -21°C pour point result). A 5–10 bbl/day pilot campaign is guided by an interview, not a contracted program with a named partner. This is a potential pathway, and the Uinta Basin context (roughly 185–200k bbl/d of waxy production against ~70k bbl/d of pipeline-compatible capacity) explains why producers would care, but I’ll believe the vertical is real when I see a pilot partner and data at pilot scale.

Vicus also hinted that the R&D team is working on the bitumen application (B2N) in the background, with news expected “shortly.” Filed under forward-looking statements; watch the wire.

Insider behaviour and the clean raise

PennyQueen pressed on the financing track record, and this is where Aduro genuinely stands apart from the microcap norm. The recent raises the US$15.54M offering at US$15.20/share and the C$9.15M LIFE offering at C$21.20, both at premiums to market, with no warrants. Vicus confirmed on record that insiders have never sold shares and continue buying when windows allow (verifiable via SEDI). Pro forma, the company has roughly C$65M in cash and no debt.

Fifteen years into developing this technology, the founders’ capital behaviour is the tell I keep coming back to. No warrants means no structural overhang; a premium raise means institutional demand, not desperation. Compare that to the ATMs, ELOCs, and death-spiral converts that litter this sector.

What comes next

Vicus’s closing message: the company is transitioning from planning to heavy physical execution. Real capital deployed on long-lead items, engineering finalization, and breaking ground on FOAK civil works at Chemelot. Management’s commissioning target remains mid-to-end 2027, worth noting that Ladenburg’s model conservatively assumes first revenue in fiscal 2029, a discrepancy I flag in everything I write.

My watch list coming out of this interview: (1) the consolidated NGP data package, including mixed/contaminated feedstock results; (2) Updates on the FOAK progress; (3) a named petroleum pilot partner; (4) the hinted B2N announcement; (5) conversion of any MOU, Ortessa, AstroTurf, or the EPC partner into a definitive agreement. That said, this company has so many irons in the fire that I would not be shocked if we see other de-risking and valuing creating milestones.

Sources

• PennyQueen’s interview with Ofer Vicus — [INSERT VIDEO LINK] | Her Substack: pennyqueen.substack.com | X: @The_PennyQueen

• Aduro/AstroTurf MOU press release, June 30, 2026: GlobeNewswire / Aduro IR

• Aduro press releases (NGP campaign results June 9, 2026; Chemelot FOAK site; Ortessa MOU June 17, 2026; TSX uplisting; financings): Aduro investor relations

• Insider transactions: SEDI | Canadian filings: SEDAR+

Attribution note: the interview summarized above is PennyQueen’s work. Sections describing management’s forward plans (petroleum pilot scale-up, B2N news, FOAK groundbreaking) are forward-looking statements by management and are subject to change. Bench-scale results are bench-scale results.

This is not financial advice. I own shares of $ADUR, which represent ~30% of my portfolio, so I am biased.

If you spend any time discussing Aduro Clean Technologies with a thoughtful investor, you eventually arrive at the same objection. It usually sounds like this:

“This is a binary, pre-revenue, scale-up bet on chemistry that hasn’t been proven at commercial scale, with first real revenue in the back half of the decade, in a sector where the graveyard is full of companies whose bench results didn’t survive the jump to continuous operation.”

That is a fair, well-informed summary. It’s also, in my view, the right thing to worry about. Not the TAM, not the policy tailwind, not the analyst price targets, the scale-up gate. If the chemistry doesn’t hold when you go from kilograms an hour to thousands of tonnes a year, nothing else on the page matters.

So rather than wave that concern away, I want to take it head-on: what the risk actually is, what evidence we have so far, why management and many long-term holders weigh it lower than the sector’s track record would suggest, and, just as importantly, where the risk genuinely still lives.

Let me be clear up front about my bias: I own the stock, it’s a large position, and I’m involved with the company in an advisory capacity. Read everything below with that in mind, and do your own work.

First, name the gate honestly

Here’s what we actually have as of now, and what we don’t.

What we have is the June 2026 data from the Next Generation Process (NGP) Pilot Plant: 86% liquid hydrocarbon recovery over the steady-state window, with 85% of that liquid landing in the C20-and-below carbon range, the band associated with naphtha-cracker feed. The number I care about most isn’t the yield, though. It’s that steady-state conditions were re-established within roughly two hours after deliberate operating changes. That’s the first real evidence that the process is controllable in continuous operation, not just repeatable in a batch flask. Separately, an independent third-party steam cracker processed the oil and reported ethylene and propylene yields comparable to fossil feedstock.

What we do not have is the thing that actually clears the gate: sustained, continuous operation at commercial scale. The pilot runs in multi-day campaigns at roughly 10 kg/hr. The First-of-a-Kind (FOAK) plant planned for Chemelot is designed for about 10,000 tonnes per year and will need to run for weeks at a time. That jump, solid handling, heat transfer across a physically larger reactor, and feedstock consistency when the input is real-world contaminated garbage rather than curated test batches, is precisely where the clean-tech graveyard gets filled.

I don’t think that risk is retired. What I think the pilot data does is move the risk from one category to a smaller one: from “the chemistry is unproven” to “the chemistry is proven, but unproven at scale.” Those are genuinely different things to underwrite. The first is a science bet. The second is an engineering bet. Engineering bets fail too, but they fail less often, and for more diagnosable reasons, than science bets.

Why is the chemistry itself the de-risked part

The strongest reason to weigh the science bet as largely settled is the distance already travelled.

Aduro has been developing this chemistry since 2011. It went from vial-scale batch reactors to a continuous-flow plastic reactor commissioned in 2023 with hundreds of runs on polyolefins to today’s continuous NGP pilot. The single hardest transition in all of chemical process development, moving a reaction out of a batch flask and proving it works in continuous flow, is the one Aduro has already made.

Aduro’s technical team has spoken to this progression of risk directly. On a recent facility tour, Aduro’s Birendra Adhikari identified the biggest scaling risk as the transition from small vials to the next-generation continuous system and characterized the subsequent step, from pilot to commercial unit, as proportionally lower risk.¹ Chief Scientist Anil Jhawar has made the same point: that the chemistry itself is proven, and that he doesn’t see the core process chemistry as the technology risk; the remaining work is chemical engineering, not chemistry.¹

To understand why the chemistry scales so predictably, it helps to contrast it with the sector’s default: pyrolysis. Pyrolysis relies on thermal energy, essentially baking the plastic in a massive oven. Like baking a pizza, heat distribution is naturally uneven. Scaling up to a larger oven doesn’t solve the problem of uneven heat transfer; it just leaves you with burned edges and an undercooked center (or, in chemical terms, unpredictable, messy yields).

Aduro’s hydrochemolytic process, by contrast, is a water-based, catalytic reaction. Think of it less like baking a pizza and more like cooking pasta in boiling water. Water is a superior heat conductor, creating a far more even reaction environment. Scaling it up is closer to using a bigger pot; the chemistry behaves the same way. The technical team has gone so far as to describe the process as getting simpler and cleaner as it moves toward commercialization, rather than more fragile.²

Why the engineering bet is smaller than it looks

Three structural features make the remaining scale-up more tractable than the sector average.

1. The equipment is off-the-shelf, not custom. This is the one most people miss. Aduro isn’t inventing a novel machine and hoping it scales; it’s running its proprietary chemistry inside proven, commercially available hardware. Company representatives have addressed this directly, describing the equipment as standard rather than exotic and framing the differentiator as the core process itself rather than a custom-built machine; they’re not reinventing the wheel in hardware.² The technical team reiterated on a recent tour that they rely on off-the-shelf equipment of the kind regularly used in petrochemical operations.¹ Standard equipment has known scaling curves. Custom equipment doesn’t.

2. The absolute jump ahead is a more understandable kind of jump. Going from roughly a couple of hundred tonnes/year of pilot throughput to 10,000 TPA at FOAK is a big step in magnitude. But it’s a step within continuous-flow operation using standard unit operations, not another format change. The genuinely treacherous transition is behind them. What remains is scale-up engineering: bigger vessels, more robust feed handling, sustained runtime. Hard, capital-intensive, and not guaranteed but a category the process-engineering world knows how to attack.

3. They’re not scaling alone, and the partners are the ones who scale for a living. This is where the counterparties stop being logos on a slide and start being risk-reducers:

• External engineering scale: Company representatives have noted that Aduro is working with established external engineering firms that have built commercial units many times the size of what Aduro is constructing at the pilot scale.²

• Shell GameChanger: Beyond funding, graduation from this program provided third-party technical validation and market direction from a major industry stakeholder engaged with the scale-up path.³

• Siemens: Supplies the industrial automation and control system (SIMATIC PCS neo), chosen to ensure that control strategies developed at the pilot scale translate directly to larger-scale facilities.

• Ortessa Groep: Signed a non-binding MOU (June 17, 2026) to evaluate a dedicated feedstock logistics center at Chemelot sourcing, sorting, cleaning, and delivery of specification-ready feedstock directly targeting the single most underrated scale-up killer: inconsistent input.

• TotalEnergies: Working alongside Aduro through technology evaluation and cooperation arrangements.

None of these relationships guarantees anything. The Ortessa arrangement is a non-binding MOU, not a signed supply contract, and I’d hold you to reading it that way. But the collective signal matters: the companies that would actually do the scaling engineering, automation, feedstock logistics, and petrochemical integration are in the room and engaged, not being pitched cold.

The part that keeps me honest

Here’s what I’m not claiming.

I’m not claiming FOAK will run clean for weeks on real-world contaminated waste. That is still the gate, and it is still unproven at that scale. Feedstock variability, trace-contaminant accumulation on reactor walls over long runs, and solids-handling at higher throughput are real, and a multi-day campaign window doesn’t prove a 30-day run. The whole reason the NGP pilot exists is to surface these problems before FOAK, and the development timeline for genuinely new industrial processes is measured in years, not quarters.

I’m also not claiming an information edge. Everything above is from public filings, press releases, independent research, and public corporate updates. My proximity to the company doesn’t give me a clearer view of the FOAK outcome than you can get from the same disclosures, and if it ever did, I couldn’t act on it or post about it anyway.

What would actually change my mind on this specific question:

1. The pilot fails to sustain steady-state as campaign lengths extend toward continuous operation (e.g., recurring fouling, coking, or control instability on longer runs).

2. FOAK slips materially or is redesigned in a way that signals the current process economics don’t hold at 10,000 TPA.

3. The feedstock logistics and pre-treatment work (Ortessa and others) reveal that real-world contaminated input degrades yields below the economics the licensing model needs.

Any one of those would tell me the engineering bet is harder than I currently think.

So where does that leave the scale-up worry?

Roughly here: the science bet, does this chemistry work in continuous flow? has largely been answered, and answered across increasingly realistic feedstocks. The engineering bet can it run for weeks at 10,000 TPA on real garbage? has not, and won’t be until FOAK actually operates. What’s changed over the past year is that the risk has migrated from the first bucket to the second, and the second is being de-risked with standard equipment and industrial partners rather than improvised alone.

“Unproven chemistry in a founder’s lab” and “proven chemistry being scaled inside standard equipment with Siemens, Ortessa, and Total engaged” are different risk profiles. Both can fail. But if you’re going to underwrite a pre-revenue hard-tech name at all, the second is the version you want to be underwriting, and it’s the version I think is actually on the table here.

That is not the same as saying it’s derisked. It’s saying the nature of the remaining risk is narrower, more diagnosable, and more heavily supported than the “bench-to-continuous graveyard” framing implies. Whether that’s enough for you depends entirely on your tolerance for a binary that won’t fully resolve until the back half of the decade.

For me, it is. For plenty of sensible people, it won’t be and watching this one from the sidelines until the FOAK gate clears is a perfectly rational call. You’d give up meaningful upside for a lot less uncertainty, which, for most portfolios, is the right trade.

This is not financial advice. I own shares of $ADUR, which represent ~30% of my portfolio, so I am biased. This piece is for informational and educational purposes only, is not a recommendation to buy or sell any security, and I am not a registered investment advisor. All figures are drawn from Aduro’s public disclosures, press releases, and independent third-party research, and are point-in-time as of writing. Do your own work.

Sources

Footnotes

1. Aduro Clean Technologies facility tour / technical team discussion (video):

1. Aduro management discussion on equipment and scale-up (video):

1. Aduro corporate update / Shell GameChanger discussion (video):

Every so often, a regulatory change quietly rewrites the scoreboard for an entire industry, and most people don’t notice until the game is already being played by the new rules. That just happened in European plastic recycling, and the person who laid it out most clearly is Aduro’s own Chief Revenue Officer, Eric Appelman, in a Plastics News op-ed published July 1.

I want to walk through what changed, why it matters, and why I think it plays directly into the thesis I’ve been writing about for months. I’ll flag the bear case at the end, because this is exactly the kind of “the rules now favour my stock” argument that deserves scrutiny before you believe it.

What actually changed

The short version: Europe has formalized mass balance accounting under the PPWR (Packaging and Packaging Waste Regulation) and the implementing guidelines of the Single-Use Plastics Directive, including a fuel-exempt clause. In plain terms, the carbon a recycler loses to fuel gas, char, and wax cannot be counted toward recycled-content targets. Only the carbon that actually ends up in a new product counts.

That sounds like accounting minutiae. It isn’t. It quietly promotes yield from an engineering metric that plant operators cared about to the number that determines whether a chemical recycling business is economically viable at all. If a third of your carbon walks out the door as low-value fuel, that third now earns you fuel value, not premium recycled-content value.

Appelman’s framing is that this makes carbon yield the primary measure of success, and that the industry is being steered toward its next phase of maturity as a result. I’d put it more bluntly: the regulation just turned yield into the moat.

The 70% wall

Here’s the problem the incumbents run into. For the largest category of plastic waste, polyolefins (PE, PP, PS, which account for>70% of municipal plastic waste), chemical recycling today essentially means variations of high-temperature pyrolysis. And pyrolysis leaks. It loses good carbon to fuel gas, char, and heavier residues, which typically caps the fraction of carbon that makes it into new polymer at roughly 65–70%.

Under the old “bookkeeping” regime, you could paper over that. Under the fuel-exempt clause, you can’t. The 30–35% you lose is now stranded at fuel value, and the margin math that made many pyrolysis projects pencil out gets a lot uglier.

This is the part worth sitting with: the constraint isn’t a subsidy or a talking point. It’s thermodynamics meeting an accounting rule. Thermal cracking hits a practical yield ceiling, and the regulator just made that ceiling a P&L problem instead of an engineering footnote.

Why this is an Aduro story

Aduro’s Hydrochemolytic Technology (HCT) is designed for exactly the metric the rules now reward. Appelman’s own language in the op-ed is stage-appropriate, and I’m going to keep it that way: early-stage data suggest the potential to push yields into the 85–90% range under optimized conditions. He also notes that HCT generates a substantially paraffinic product without the extra hydrogenation step that other routes typically require.

Do the arithmetic that the regulation now forces you to do. Moving carbon retention from ~70% to ~90% isn’t a 20-point improvement; it’s the difference between a third of your output being penalized and almost none of it being penalized. That’s a step-change in chemical recycling margins, not a rounding error.

And this isn’t only a slide. Aduro’s own disclosed pilot data (Tier 1, from the NGP campaigns) show roughly 86% liquid hydrocarbon yield, with about 85% of that liquid at C20-and-below, i.e., in the naphtha range that steam crackers actually want. In his Sustainable Plastics “5 questions” interview, Appelman cited yields of up to ~95% on pure PP feed, a baseline figure, on a clean single-polymer stream, that I flag as such rather than treat as a mixed-waste number.

The through-line: the mechanism was built to keep carbon in the product. The regulation just made keeping carbon in the product the thing that pays.

The firewall (because this always gets muddled)

I keep this distinction in every piece, and I’ll keep it here: HCT is not pyrolysis. Pyrolysis snaps chains thermally and more or less at random, leaving reactive carbons that recombine into an inconsistent gas/wax/liquid soup, which is why it needs costly hydrotreatment before a cracker will take it, and why the carbon losses stack up. HCT operates at lower, sub-supercritical temperatures in a water-based, catalytic environment that selectively cleaves C–C bonds while sparing C–H bonds, with a hydrogen donor capping the freed carbons. The result is a consistent, lighter, paraffinic output that has been fed to a steam cracker as-produced. Different mechanism, different yield ceiling, different economics. The “70% wall” is a pyrolysis wall, not a chemistry-wide one.

What would change my mind

I’d be doing you a disservice if I only gave you the bull case, so here’s what I’m watching that would break this:

1. The 85–90% number stays a promise. Right now, it’s early-stage, optimized-condition data, not a commercial guarantee on messy, mixed, contaminated feedstock at scale. The credibility threshold is contaminated/mixed-feedstock campaigns at NGP, and the full validated data package. Until then, treat the yield as directional.

2. FOAK scale-up disappoints. The Chemelot first-of-a-kind plant has to actually run at ~10,000 t/yr and hit these numbers outside a pilot. Scale-up is where a lot of good chemistry goes to die.

3. No commercial license signed by the end of 2029. The whole thesis is a licensing model. Regulatory tailwind or not, if there’s no signed commercial license by then, the “yield is the moat” argument stays theoretical.

4. The regulation is EU-specific, and enforcement can slip. The fuel-exempt clause is a real tailwind in Europe. Timelines get pushed back, carve-outs appear, and North America becomes a different, more fragmented regime. Don’t extrapolate one jurisdiction’s rulebook to the whole TAM.

Sourcing

• Tier 1 (company / regulatory): Aduro NGP pilot data and disclosures; the PPWR/SUPD mass balance framework as described in Appelman’s Plastics News op-ed.

• Tier 1-adjacent (Aduro executive, published): Eric Appelman’s Plastics News op-ed and his interviews (linked below). These are Aduro’s CRO speaking, informative and directional, but they are the company’s framing, not independent third-party validation.

• What this is not: Independent sell-side coverage, and not a claim that 85–90% is a proven commercial spec. It’s stage-appropriate potential plus formalized regulation that rewards it.

Sources & further reading

• Eric Appelman’s op-ed (the piece that prompted this): Why mass balance is a lifeline for plastic recycling economics — Plastics News

• Eric Appelman interview (most recent, full-length): How To Scale Sustainable Technology — The Chemical Show

• Eric Appelman on mass balance specifically: 5 questions for Eric Appelman — Sustainable Plastics

Long $ADUR | Not financial advice | DYOR

Disclosure, again, because it matters: I hold $ADUR at ~30% of my portfolio. I’ve tried to keep every claim above stage-appropriate and sourced, but I am not a neutral party. Read the primary sources and decide for yourself.