YTD return is around a very volatile +100%, which is great, but it’s mainly been driven by sizing STLN and TVA Group correctly. Also, dislocations intra-quarter are getting bigger in small caps, so I think there’s a lot of alpha in buying and holding through earnings, even when the results are fairly predictable.

Not mentioned here, but still bullish and holding 4441 (Tobila Systems).

Subscribe now

Evolent (EVH), not holding. Is on the value-based basket I’m paying attention to. EVH is a company that health plans hire to manage the cost of expensive specialty care, mostly oncology, plus cardiology and musculoskeletal.

In its main model, the Performance Suite, it takes full financial risk on the total cost of care for that specialty. The structural catch is that it’s on the hook for cancer drug spend but doesn’t own the pharmacy that dispenses the drugs, which is exactly what nearly killed it when oncology costs spiked in late 2024 (and why STLN model is working so well). So it’s a bet on the medical expense ratio (MER), the share of premium eaten by claims. Right now that’s the main thing I’m watching. It printed 95.3% in Q2, up from 93% in Q1, because the big new Highmark contract launched May 1 and fresh books always run at a higher cost before they season.

Management guides to roughly 93% for the full year with Q4 improving about 250 basis points. That walk down is the entire thesis. Where the numbers need to go: MER to 93% and then below (which is not an easy thing); adjusted EBITDA margin off its current 4.3% (EBITDA actually fell 25% year-on-year even as revenue rose 47%) back toward the ~8%+ it earned in 2023; and the balance sheet fixed, because ~$808M of net debt against ~$127M of EBITDA is ~6.3x, and they’re only generating $10–20M of operating cash flow this year after ~$60M of interest. If MER normalizes and they term out the 2029 maturities, it’s probably a double as it re-rates. If oncology costs spike again, it breaks again.

P3 (PIII), tracker position. Is one I’m looking at closely. Highly distressed situation. Takes capitated risk on Medicare Advantage members through physician networks: it gets a fixed amount per member per month (pmpm) and keeps the spread if it manages care for less.

It nearly went under, deliberately shrank membership to cut loose bad contracts, and is now inflecting. The trap in the Q2 print is that the headline medical margin, $97.8M or $311 per member per month (PMPM), was roughly half real. Strip out favorable payer settlements and prior-year reserve releases and it’s $52.9M, or $168 PMPM. They raised the full-year adjusted EBITDA guide to $80–110M, but that leans partly on those one-timers, so what I’m watching is clean quarters where the margin holds without accounting help.

The deeper point is that this is a capital structure story more than an operating one. There’s about $282M of preferred stock at a ~16% blended rate accruing ~$45M a year, which is more than the midpoint of the EBITDA guide, so even in a good year the senior claim compounds faster than the business earns.

Where the numbers need to go: clean medical margin up toward the $210–240 PMPM guide; membership turning from shrinking (~116K) back to growth toward ~140K lives; cash building off a thin ~$26M base into real FCF. And the preferred: until that ~$282M is refinanced, restructured, converted, or cleared by an acquirer, the common sits behind ~$520–570M of claims and is basically a levered option on the turnaround working.

Starling (STLN), my biggest position and driving a significant portion of my YTD performance (someone said volatility?!)

Formerly The Oncology Institute (TOI), is community cancer care, and the thing that makes it interesting is that it owns its own pharmacy. Where Evolent takes drug-cost risk without owning the pharmacy, Starling owns the dispensing and keeps the drug margin. The core insight is that a capitated cancer patient isn’t really profitable on the capitation rate from the start, but it is on the prescriptions that patient fills through Starling’s own pharmacy. So each capitated life is essentially an acquisition of future pharmacy volume, and as scripts per patient cohort ramp over time (roughly 1.0 to 3.2), gross profit per thousand lives climbs about 68% with no rate increase at all.

What I’m watching: it just posted its first-ever positive adjusted EBITDA, tiny at ~$0.2M but a real milestone, on revenue up 35% and specialty pharmacy up 58%, and the July refinancing with OrbiMed took the B/S risk off the table.

Where the numbers need to go: adjusted EBITDA from ~0 up to a normalized ~5–6% of revenue (~$35M) as operating leverage plays through, with SG&A expense continuing down from 19% (already off ~30%) toward the mid-teens; the pharmacy to capitation flywheel extending beyond Florida into the new Nevada, Oregon, and California contracts. I believe the runway here is huge, as I see a path to $1bn of revenue with 6%-8% margins in the next couple of years, on 15x-20x EBITDA, so still a double from here.

HireQuest (HQI), holding around 5%. Is a staffing franchisor and an old friend. I lost some money with them but I’ve kept an eye on the company over the years. It doesn’t run staffing offices itself; it franchises the brand and collects a royalty of about 6% of everything those offices bill (system-wide sales), and it finances the franchisees’ payroll and workers’ compensation through its own receivables.

Capital-light, no debt, ~20% net margins, and it stayed profitable through eleven straight quarters of a staffing downturn while the big direct operators lost money. The reported Q2 is muddy because it just divested a breakeven perm-placement business (MRINetwork).

Headline revenue was only up 6%, but strip the divestiture and the continuing royalty base grew 13.8%.

Where the numbers need to go: system-wide sales, the royalty base, needs to re-accelerate as the labor market stabilizes, and management flagged the first sustained favorable weekly comparisons, so I’m watching whether that clean 13.8% organic royalty growth holds, which according to management, they are seeing those numbers already during early Q3. Margins are already high, so it’s about protection rather than expansion, which mostly means holding the ~6% take rate and keeping an eye on workers’-comp cost swings, the one thing that can genuinely go wrong and what made me lost money last time (workers comp going the wrong direction + macro headwinds). The B/S is very strong, so not a worry there. The inflection is a cheap acquisition converting/continued buybacks (20% of the market cap original authorization last year), plus the cycle turning.

Subscribe now

Not investment advice. Do your own work.

The thesis

Trifast is a genuinely good niche business that spent a few years being run to achieve volume rather than profitability. They got a new management team and an activist on the board since late 2023, and has since rebuilt its margins and balance sheet while the market carried on treating it as the broken cyclical it used to be.

On a lease-adjusted basis (to make it comparable with peers) it trades at roughly 7.6x EV/EBIT and 5.9x EV/EBITDA, while its peers (Bofard 12x, Bufab 22x), trade significantly above.

Some of that discount is fair: Trifast still earns a lower group margin, its revenue is shrinking while both peers grow, and only about a third of the shares actually trade. But I believe the discount is too wide. The margin gains are structural self-help delivered against an 8 point volume headwind; the revenue fall is roughly 93% market-driven and only 7% deliberate pruning; every region except Asia grew operating profit; and the largest shareholder, a well-known activist, bought more stock the day the results came out.

It is a two-part setup: as the margin gap to peers keeps closing and the top line stabilises, you get a higher normalised profit and a higher multiple at the same time, worth roughly 75% or more together, with a strong balance sheet capping the downside.

Subscribe now

What Trifast does

Trifast supplies industrial fasteners and category-C components: the screws, bolts, nuts and washers that hold an engineered product together. On any given assembly these parts are only about 2% to 5% of the total value, but they can be 60% of the part count. And the price of the bolt itself is only around 15% of what the true cost for a customer is. The other 85% is sourcing it, qualifying it, holding it, kitting it, inspecting it, and the risk that a single missing 20p component halts a production line worth far more than the entire fastener contract.

That gap is the whole business. Trifast’s job is to make sure the 85% part works smoothly, which is why around 60% of its revenue is now design-led or engineered into the customer’s product rather than sold off a shelf. Its vendor managed inventory systems sit physically on the customer’s production line and reorder automatically, which raises switching costs and supports a margin structurally above commodity fastener distribution. This is a sticky, engineered, embedded supplier.

The business before, and after

For a while, TRI was not run to capture any of that. Under the previous management the priority was volume and acquisitions. Revenue peaked at £244m in FY23. But the quality underneath deteriorated. Gross margin slid to around 25%, underlying EBIT margin collapsed below 5%, ROCE fell to the mid-single digits, and net debt ballooned to £38m as inventory and working capital bloated. It was a good business being run for the top line, carrying too much low margin work and too much stock, at exactly the moment the cycle turned.

Iain Percival became CEO in late 2023, with Kate Ferguson as CFO, and Nicholas Mills, from the activist investor Harwood, joined the board at the same time. Their strategy, Recover, Rebuild, Resilience, inverted the old priority. Margin and cash first, volume second. That meant repricing or walking away from low margin business such as white goods and low margin automotive, squeezing procurement, closing the Malaysian manufacturing operation, moving back office work to a shared service centre in Hungary, consolidating the UK distribution centre for about £3m of annual savings, and putting in a single ERP system that finally gave them the data to see, contract by contract, where they were underpriced. Headcount came down about 10% and payroll 6% along the way.

The result shows up when looking at the last 5 years numbers:

Looking at the bottom rows, the turnaround is visible. Since the FY23 peak revenue is down about 15%, but gross margin has gone from 25.3% to 30.0%, underlying EBIT margin from 4.9% to 7.9%, and net debt from £38m to £16m, with leverage now 0.75x and cash conversion near 98%. Margins are now higher than they were before the downturn, on lower revenue. This is a better business than the one three years ago. The mix has been deliberately rebuilt as well. Automotive is down from about half of revenue to 37%, while Smart Infrastructure has roughly doubled to 17% of revenue and is targeted at 30% by FY30, alongside a fast-growing Medical business, a signed five-year Saudi contract, and an India operation that grew about 90% YoY in Q1 27.

The numbers, and the gap to peers

The clearest single picture of the turnaround is the margin line.

Two years of improvements has taken gross margin up nearly 5 points and closed more than half of the historic gap to peers. Management’s own bridge shows this is structural. Over two years they added 10.8 points of margin from optimization; that is pricing, mix, efficiency and a leaner organisation, against an 8 point headwind from falling volumes, netting the 2.8 points up to 7.8%. The margin expanded while the cycle worked against them, which is the opposite of most recovery stories and why this one is not fully visible for the market.

Set against the two clean comparables, both running the same VMI and engineered-fastening model, the gap that is left is smaller than the share price implies.

Trifast FY26 underlying; Bossard TTM to Jun 2026; Bufab adjusted EBITA Q2 2026. Peer margins from company reports.

On gross margin, Trifast at 30.0% is now only three to four points behind Bossard and Bufab, less than half the gap of two years ago. On operating margin the distance is larger, about 350 basis points to Bossard. A gap that should close with operating leverage.

But even that gap is really not structural. Every region except Asia grew operating profit in FY26; Europe already earns around 11% underlying EBIT margin and North America close to 9%, at or near peer level. The entire group drag is one region, Asia, hit by Chinese EV competition and currency, plus central cost. Strip that out and Trifast already looks like its peers.

Why the peers grow and Trifast not

The obvious objection in the valuation right now is growth. Bossard grew organic revenue about 10% and Bufab over 5% on their latest numbers, while Trifast fell 7%.

That divergence is the best single reason for a discount. The drivers can be divided in 4 things, only one being a competitiveness problem.

1. Mix: Trifast is over indexed to European automotive, while the peers lean more to the Americas and to electrification end markets that have been recovering.

2. Choice: some of Trifast’s decline is deliberate pruning of low margin work, part of the quality over volume reset. The revenue bridge attributes only £2.0m of the fall to deliberate exits against £26.7m of market contraction.

3. Sequencing: Trifast fixed margins and the B/S first and is only now turning to growth, whereas the peers never had to repair and have compounded throughout.

4. The tell: Trifast’s own growth engines, North America, India, Medical and Smart Infrastructure, are already growing at peer like or better rates. They are just outweighed for now by the shrinking legacy base. Management has named the return to top-line growth as the FY27 priority.

The multiple gap

Quick note that Bossard reports under Swiss GAAP, which does not capitalise leases, while Trifast and Bufab use IFRS 16, which does. So below is a table with the appropiate adjustments.

Bossard TTM = FY25 + H1’26 - H1’25. Bufab Q2’26 adjusted EBITDA.

On the comparable basis Trifast trades around 7.6x EV/EBIT and 5.9x EV/EBITDA against Bossard’s 11.7x and 9x, a discount of roughly 35%, and it does so with a balance sheet a third as levered. Bufab, on a near-record 38x P/E, is a premium outlier and the wrong anchor for a base case.

The discount to Bossard is wider than the margin gap justifies. Trifast earns about 70% of Bossard’s operating margin but trades at around 65% of its EV/EBIT, and it carries by far the strongest balance sheet. The one axis where the discount is fully earned is growth, and even there the decline is mostly the cycle, not the business. And the cyclicality exposure of the business is going down…

The europen automotive market

The single macro variable that would flip Trifast’s top line from drag to tailwind is the european automotive market, and after three bad years it is finally recovering. EU vehicle production fell around 9.6% in 2024 and a further 4.3% in 2025, and the forecasters now cluster on 2026 landing somewhere between flat and modestly negative before a return to growth in 2027. EUROFER pencils in +3%, others sit in the +1% to +2.5% range. This lines up with the FY27 return to growth management has named as its next priority.

Trifast rebuilt its margin structure and repriced its worst automotive work during the downturn, so whatever volume returns lands on a leaner, better-priced base than the one that existed at the FY23 peak. The incremental margin on recovering auto volume is therefore higher than it was before the reset, which makes an end market inflection a bigger earnings lever today than the 37% auto weighting alone implies.

M&A as a de-cyclicalisation tool

The balance sheet could buy the option to accelerate the shift out of cyclical end markets. At 0.75x leverage, near-full cash conversion and roughly £70m of undrawn facilities, Trifast has enough cash to buy something small. The instinctive objection is that M&A is precisely how the previous management fucked up. But the new team has inverted the old logic everywhere else, and there is no reason capital allocation should be the exception. They can pull the main source of revenue away from European automotive faster than organic growth can manage. Remember that the governance is aligned for it, with Nicholas Mills in the board I expect him to favour either value realisation or disciplined allocation.

Valuation

As usual, using rough numbers. Holding revenue flat at about £210m and net debt flat at £16m, which is conservative since it is still falling, on an ex-lease basis at around 80p:

The 10x EV/EBIT anchor is deliberately conservative, since Bossard sits near 11.7x on the same basis, so it already discounts Trifast for lower margin, negative growth and thinner liquidity. Neither lever needs heroics.

Re-rating from 7.6x to 10x with no margin improvement is already about a third; delivering the double-digit margin management guides to takes the total toward 75%; and the growth pockets in India, Saudi and Smart Infrastructure are unpriced optionality on top.

Management is explicit that the margin target does not depend on any industrial recovery, with PMIs stuck below 50 throughout, and 3 out of 4 self-help levers have already delivered. The one still to come, profitable growth, is what carries the margin the last leg.

It helps that I’m not alone in the trade. About 2/3 of the shares sit with seven holders, and the largest, at 20.5%, belongs to vehicles connected to the activist investor Christopher Mills, whose son Nicholas Mills sits on the board. Mills has a long record of pushing smallcap boards toward value realisation, and connected vehicles bought a further £2.7m of stock at 68p on the day the FY26 results landed, taking the market’s shrug as their entry.

Risks

The same concentration is also the clearest risk. With only about a third of the shares in free float this is an illiquid micro-cap, and the discount can persist for years regardless of fundamentals.

Also, the top line is now a market call, so it inflects only when automotive and industrial end markets do. And Asia can swing group margin by a point in either direction. Additionally, until the ERP investment finishes rolling off, statutory profit stays near zero and screens will keep flagging Trifast as loss-making, which keeps generalist buyers away.

The offset is the balance sheet. At 0.75x leverage, near-full cash conversion, a yield around 2.4% and about £70m of undrawn facilities, the realistic bear case is dead money at 7x to 8x for a couple of years, not a permanent loss.

Bottom line

The re-rate needs the margin trajectory to continue, which management attributes to structural self-help rather than the cycle, and the top line to stabilise. If both happen, a move from 7.6x toward a Bossard-like 10x to 12x on a modestly higher profit base is worth roughly 75% or more, with an activist on the register already buying and the best balance sheet of the three underneath. Not as expensive as Bufab, still a fair discount to Bossard, but too cheap for what it has become.

Subscribe now

Not investment advice. Do your own work.

TOI reports next week. I’m long and have slightly changed how I view the oncology value-based care model.

Long term, I still think it’s going to be a very big trend. At scale, oncology VBC players get enough operating leverage to absorb drug-pricing volatility, which matters most in oncology, where that exposure is huge. A single bad drug year sinks a small book but barely dents a large one. I think TOI is well positioned to get there, as it’s starting from attractive markets (mainly thinking about Florida) and I’d expect further expansion, maybe to be announced in the Q2 call next Thursday.

The LT case is about operating scale, enough volume to spread volatility. The ST problem is actuarial credibility, individual oncology books are just too small to be statistically reliable. Small denominators create noise, disease heterogeneity is high, and drugs dominate the spend. That’s why demand flows toward episode based and pathway/drug-management models rather than pure per-member risk, which favors $TOI’s pathways and dispensing over raw risk-taking. We’ve seen this with ridiculous growth in TOI’s dispensary segment.

But the ST limitation doesn’t worry me because it’s self-correcting. The demand that outruns full capitation today (payers desperate to delegate oncology as drug trend inflates) is what delivers the volume that makes capitation work tomorrow. Delegation pulls volumes forward. It’s a flywheel, and while it plays out, the dispensary business just prints FCF and grows fast.

Subscribe now

On FY26 guidance TOI screens as barely profitable, adjusted EBITDA of $0 to $9M on ~$640M revenue, at $560M EV. The business runs three lines: low-margin FFS, capitation that keeps a medical-margin spread at a blended MLR in the low 80s, and an owned specialty pharmacy carrying an 18% GM because TOI retains the Part D drug economics rather than passing scripts through. Blend those company margins on a mature mix of roughly half pharmacy, thirty percent capitation, twenty percent FFS and gross margin lands near 17%, a figure that reconciles to the reported ~16% today. The key is that a capitated life is the acquisition event, not the profit, since the capitation rate barely moves ($50 to $54 PMPM over three years) while scripts per cohort climb from 1.0 to 3.2, tripling attached pharmacy revenue and lifting gross profit per 1,000 capitated lives ~68%, from $9,949 to $16,685, with no rate increase. Pair that with the operating leverage already visible, SG&A down from 30% of revenue in early 2024 to 19% by Q1 2026, and a normalized EBITDA margin of ~5 to 6% (about $35M on the current base) is sensible. Of course, with further scale, operating leverage will keep improving. That gives us a multiple of ~16x normalized, for a business still guided to grow revenue near 20% a year. A further squeeze by the IRA on the pharmacy margin will only increase the demand for capitation, improving the operating leverage of the business.

I’m pretty sure researching while on a cabin in the woods boosts returns

TVA Group

I recently wrote it, so feel obliged to say that I closed my TVA position yesterday (+35% in 25 days).

The business income statement was being dragged down by a money losing national French language NHL sublicense. The market was anchoring on the top line and missing that the loss-making deal was rolling off with the 25/26 season. Once it was gone, revenue would fall and that would be good news, because the bleeding stopped with it.

I’m not sure why but I like to play a lot of these revenue shrinks-EBITDA inflects plays. The market re-rates once it stops reading the headline number and starts looking at profitability and the new economics of the business. I did it with BELFB on a huge size and going to share another idea shortly where I’m seeing the same situation.

I mentioned as well that there was optionality via a plausible take private by Quebecor, and an unresolved NHL rights question where non renewal of the bad contract was the bullish outcome.

The NHL rights overhang substantially resolved. The old national sublicense was confirmed on its way out; for 26/27, TVA Sports keeps the Canadiens games while the older national deal expires.

Second, the Q2 print on Thursday announced revenue falling sharply YoY on the sublicense roll-off, which read as ugly, but adjusted EBITDA moved the right way, extending the favorable swing already visible in Q1.

Quebecor is still a strong take private, and I expect further margin recovery, but that is more speculative and longer in duration, while holding a not so good business. So of to the next one.

Subscribe now

Not investment advice. Do your own work.

I’m taking the chance to review the business since my last post in 2024 and explain what has happened in the last two years, as I still hold it and has generated (and I expect it to continue to generate) good returns for me.

If there is one lesson from this review is that when buying good businesses at cheap prices that keep growing, you don’t really need a mulitple re-rating, or the market discovering the idea.

Moury Construct

In May 2024 I wrote about Moury. A fourth-generation family business in Liège, running a negative working capital model because their public clients pay in advance, with half of its market cap sitting in cash. I paid around €600 avg for my shares and argued that with a balance sheet like that and 15%-20% FCF/EV at the time, very few things could go wrong. Growth, hinted at by new contracts that were to be announced in September 2024, was the possible catalyst.

What happened

The Sep 2024 contracts came later than expected. 2024 revenue actually fell 4%, to €186 million, because many of the new projects were in their start-up phase during H2. Percentage of completion accounting punishes you in these situations. You could only see the improvement of the business that year looking at the backlog.

And finally 2025 arrive with revenue growing 34% to €249.6 million. Operating profit grew 31% to €38.1 million, which means the operating margin held above 15% for the fourth year in a row (in construction!). Net profit grew 41% to €34.5 million, or €87 per share. And a key number, net cash, ended the year at €152.5 million, up from €119.6 million. The company grew revenue by a third, paid dividends, and still added €33 million to the pile. That is what negative working capital looks like when it works, because growth does not consume cash here, it actually increases it.

Visibility got better too. The order book at the end of February 2026 stood at €371.5 million, roughly 18 months of activity already signed, and management guides to keeping the record 2025 level of activity through 2026.

Subscribe now

A key change

The most important development of these two years is that in late 2024, the CEO, Gilles-Olivier Moury, bought out his siblings through his holding company and now controls just over 50% of the shares.

For a decade, the company said it looked at the future avec sérénité, paid a modest dividend, published only in French, and let the cash accumulate. If you are planning to buy out your family, a depressed share price is a good thing. Then the buyout closed, and within months the company introduced its first ever special dividend of €5, raised the ordinary from €11 to €12.50, and committed to repeating the special in 2026 and 2027. A year later the ordinary goes to €14.20 and the special to €7.50, both paid this June, €21.70 gross in total.

When you have a family-controlled business, incentives are the main thing to watch. Gilles-Olivier controls the company and only gets paid when the rest of us get paid. That alignment did not exist when I wrote the original piece, and it’s a step forward towards getting the accumulated cash distributed.

What the stock did

The share price went from the €635 where I finished buying to around €750 today. Add roughly €50 per share in gross dividends collected along the way and the total return sits somewhere near 30% over two years.

This is not the 20% IRR path I expected in my original thesis, but the whole point around holding these kinds of businesses is generating predictable returns, which Moury certainly does (just look at the backlog).

And the business actually ran much faster than the stock. EPS are up over 40% in the last year, the cash pile grew, and the ex-cash multiple is still around 4x earnings. Strip out €380 of net cash per share and the market is paying about €370 for a business earning €87/share, growing, with 18 months of revenue signed. The market is basically paying (cheaply) for earnings and refusing to pay for the balance sheet.

What I’m looking at now

1. The revenue mix is shifting.

My original thesis leaned on public clients that prepay. The recent growth has come partly from selective private projects, and the subsidiary driving it, BEMAT, grew over 50% last year and is now around 40% of group revenue. Private clients do not systematically pay in advance the way public administrations do. Gross margins already ticked down in 2025 even though operating margins held. The negative working capital is still doing its thing (as the 2025 cash build proves), but the blend is changing and the balance sheet might become weaker, Despite actual earnings growth, if the % of private projects keeps increasing by needing to fund growth.

2. The cash is not all cash.

Around 1/4 of the treasury sits in a securities portfolio of equities and bonds, individual positions capped at 2%, plus a gold position that has appreciated to almost 6% of the book. Part of the €8.5 million financial result, which more than doubled last year, is portfolio gains rather than deposit interest. Please do a haircut to the pile when valuying this kind of situations, it’s never straightforward.

3. The 15% margin assumption can still be challenged.

A 15% operating margin in construction is extraordinary. The bear case for 4 years has been mean reversion to mid single digits. It hasn’t happened, but that doesn’t mean it won’t happen in the future. If margins revert, the current multiple and the cash mean the downside is not as significant, which is the asymmetry I bought. But I hold this position knowing the margin is the key risk.

4. The discount may simply never close.

11 shares traded on a recent session, so this is illiquid! Reports in French, no analysts, a controlling family, no index that will ever include this... This is a considerable ceiling on any significant rerating. I’m ok with the idea that the return here may come entirely from dividends plus retained earnings compounding, with the multiple doing nothing. At these prices, that is still an acceptable outcome for me.

What actually needs to happen for this to go great

1. The first is a step change in distributions.

The specials divs are committed through 2027 and then the framework is open. The company returns about €8.6 million a year while generating 4x-5x that. Buybacks cannot absorb it because the float will not allow it. A large extraordinary distribution or a tender offer would be a surprising outcome given the characteristics of the owners (traditional, conservative), but for the first time there is a controlling shareholder which will significantly benefit from it. I don’t expect any significant change regarding distributions before end of 2027.

2. The second is a fifth and sixth year of 15% margins.

I do believe at some point the market starts recognizing this is a superior construction business if the margin holds.

3. The third door is the wild card.

Someone simply deciding to take private the whole thing. I do not underwrite takeovers, but the cheaper and cleaner an asset stays the longer that option runs without costing anything.

Check the dividend tax

Since a significant part of the return here comes from dividends, the tax matters. Belgium withholds 30% at source, which is why the €21.70 gross became €15.19 net this June. Most tax treaties, including those with Spain, most of Europe and the US, cap the Belgian withholding at 15%, but getting the difference back means filing a reclaim with the Belgian tax authorities, paperwork that many brokers will not do for you on a stock this small. The treaty portion is generally creditable against your home taxation but the excess above it usually is not, so the lazy path costs you real basis points every year. Not tax advice, just a reminder to check what your broker actually does with Belgian withholding before you assume the net yield.

Where I stand

Same place I was two years ago. This is not a place to bet half of your portfolio, but a derisked way to probably earn 15%-20% IRR. Nothing that happened since has changed that framing. The price is actually lower against what the business produces than when I bought it.

The thesis is playing out in the fundamentals and dragging its feet in the price (I should get this tattooed as this is my day to day). But this is a more confortable position than a value trap where there is no growth and you are just waiting for a catalyst to happen.

You do not need the market to agree

If it has been cheap forever why would it stop being cheap? My answer is that it does not need to stop.

Run the math and suppose the multiple never moves. You still collect a dividend that has grown every year and now yields around 3%, and you own a per-share stream of earnings and cash that keeps compounding. If earnings per share grow high single digits and the payout keeps climbing, the stock price has to follow the fundamentals eventually, without the market re-rating it. Cheap forever + growing forever is a perfectly good setup. As long as the fundamentals keep improving we are fine.

If a step change in distributions arrives after 2027, or the market one day decides a construction company with these margins and this balance sheet deserves a more fair? multiple, that is just additional returns.

Subscribe now

This is not investment advice, do your own work.

TASE, BELFB, EVC, MIPS, TOI

I’ve been looking at what has and hasn’t worked for me over the last years. Most of the write-ups I’ve published here (17 ideas so far) have worked (15 out of 17), but only three have really worked. By that I mean being in the 4x-6x territory. But guess what? My portfolio is not up by that much despite all of these successes being the top position in my portfolio at some point in time.

Tel Aviv Stock Exchange (TASE) is up around 6x since December 2023, Entravision (EVC) has around 5x since November 2025, and Bel Fuse (BELFB) is a 4.5x since February 2024. Numbers are approximate, measured from the publication date of each article.

The pattern for these three businesses is one you’ve probably heard numerous times. A cheap entry point and a high quality management on a decent business will do most of the work for you (TASE is an exceptional one to be fair). The relevant part is that there wasn’t either one factor or the other, but both happening at the same time. I still managed to leave most of the money on the table.

Subscribe now

Tel Aviv Stock Exchange: right thesis, unpredicted magnitude

I published the TASE write-up in December 2023, two months after the war started in October 7th. That timing felt uncomfortable, which in hindsight was the point. The thesis was a clearly laid roadmap of value creation: a monopoly exchange that had only converted from a not-for-profit into a for-profit business in 2019, with operating margins still far below global peers, management incentivized with options struck far above the share price, and a company buying back its own stock aggressively while the war headlines kept everyone away.

Margins kept expanding, volumes recovered and then broke records. Foreign capital flooded into Israeli assets once the conflict de-escalated, and TASE became the tollbooth for all of it. The company moved to a Monday-Friday trading week in January 2026 to align with global markets. Q1 2026 revenue grew 40% year over year with EBITDA up 87%. A business that originally guided to 10-12% annual revenue growth is compounding at multiples of that.

The truth is that I underwrote an operational improvement story at a low multiple. The stock re-rated from the mid-teens to more than 50x earnings. Earnings growth explains maybe a third of the return. The rest is the market changing its mind about what this business is worth. I deliberately don’t try to predict multiple expansion because I want my theses to work without a change of mood in the market. But, of course, when you own high quality businesses, a repricing is always possible. It always helps if you buy at a price where being wrong about it can’t hurt much.

Bel Fuse: the thesis paid, then the surprises paid more

I wrote Bel Fuse in February 2024, right after a 20% post-earnings drawdown, although I have held it for a year or so already. The thesis was simple: a historically mediocre electronic components company transformed since 2021 by a new CFO, Farouq Tuweiq, who had already delivered 3 years of cost optimization and margin expansion, was buying stock personally, and the market was still pricing the company below peers as if the transformation was unproven. I wrote at the time that even assuming a disastrous 2024 with no growth afterward, the valuation looked cheap.

2024 was indeed soft, and it did not matter. The company acquired Enercon Technologies later that year, adding more aerospace and defense exposure exactly as that cycle was going vertical. Then AI data center demand turned power components from a boring category into a strategic one. Revenue grew 26% in 2025, earnings grew 50%, and Q1 2026 showed 17% growth with strength in defense, space and AI-related demand. The stock went from roughly $62 to above $270, briefly above $300.

If you look at the components separately the picture is humbling. The part I underwrote, margin expansion and a modest re-rate toward peers, would have justified just a double. The part I did not underwrite, an acquisition I could not know about, a defense supercycle, and an AI capex wave, delivered the rest. The multiple went from around 12x to above 60x earnings. Even if my work identified a company with potential for optimization and a great new CFO that was doing an excellent job, the tailwinds of the industry did the rest.

Again, where do I really contribute? Increasing my position after a drawdown in the market post a negative quarter, judging there was no damage to the transformation itself. The market was over punishing but the business still had so much potential.

Entravision: six months of nothing, then everything at once

I published in November 2025, after Q3 results, with the stock around $2.50. It was already a popular fintwit stock. EVC was a left-for-dead Hispanic broadcaster that had lost its Meta reselling business (more than half of revenue) in 2024, trading at a fraction of the value of its parts. I listed the drivers in order: continued growth of the ad-tech segment (Smadex), a midterm political advertising rebound in 2026, and potential spectrum monetization, with tangible assets as the floor.

For six months the market didn’t really care. The stock was still under $4 in early May 2026. If I had needed the market to agree with me on a schedule, I would have possibly sold (which I considered multiple times). Then Q1 2026 results landed: revenue up 114%, the ad-tech segment up 204% with operating profit going from $6.5 million to $34 million. The stock roughly doubled in a day and kept climbing. It sits around $13 as I write this.

The driver I ranked first is the one that paid far beyond what I modeled. The spectrum, which many considered the sexy part of the story, has still monetized nothing. The Media segment, which I expected to stabilize into the political cycle, actually posted operating losses. One good driver covered every mistake in the write-up. But again, the entry price point left enough room to feel comfortable holding the position.

Guess what? I sold all of them

The returns above belong to the articles, not to my account. I sold most of TASE, BELFB and EVC at some point on the way up, in every case around the moment they reached my fair value estimate. It felt disciplined not holding “overpriced” companies. These kinds of situations are the most expensive mistakes I have made as an investor.

The problem for me is structural, not emotional. I am a value investor, which means I am biased towards cheapness. When I estimate FV, I use conservative assumptions, I haircut the growth and I don’t to pay up for optionality. That is exactly the right stand at the moment of buying. It is exactly the wrong one at the moment of selling, because my fair value estimate usually sits at the lower end of the reasonable range. So, when a quality company reaches my fair value what has usually happened is that the business confirmed the thesis, management kept executing, and the market is just beginning to change its mind, which is the worst possible moment to leave! All this growth investors want your shares, keep them!

Sure, selling a mediocre company at FV is discipline. Selling a quality company at fair value is a category error. The whole point of the cheap entry plus quality management combination is that quality management keeps creating value.

Share Stoic Capital

When I think it doesn’t work: MIPS

If cheapness alone is not the answer, quality alone is not either. MIPS is a great company I’ve been wanting to buy for a while. It is the closest thing to a royalty on head safety: an ingredient brand licensed into other companies’ helmets, with dominant share in rotational impact protection, gross margins above 70%, almost no capital needs, and a long runway as safety standards tighten across bike, motorcycle and construction helmets.

And yet I have never bought a share, because the entry price removes the possibility of the returns this post is about. MIPS almost always trades at 40x-60x earnings. At that price, the market is already recognizing the quality of the business. There is no repricing left to capture, only execution risk/reward. The multiple will actually act as a headwind because the business can perform beautifully and the stock can still go nowhere for years while the valuation normalizes (or maybe this is normal). In the other cases, quality management compounding from a cheap base gave me 4x to 6x. Quality management compounding from 50x earnings gives you, in a good scenario, the growth rate. I would rather wait, possibly forever, for the entry point, which got even further away after they reported great results yesterday and the stock moved +25% (ouch!).

Hopefully I’m applying this lesson with TOI

I think the Oncology Institute is the best business I have in my portfolio right now. I first wrote it up in September 2025 around $3, and it trades around $6 today. A clean double, and comfortably through the fair value I had in mind when I published.

I could perfectly have sold already. I am not, for the same reasons I should not have sold the other three. Value-based oncology has a structural cost advantage over hospital systems that gets more valuable every year healthcare inflation runs. The capitated model aligns TOI with payors in a way fee-for-service oncology never will. The company is still tiny relative to the market it is attacking, the model only recently proved it can generate profits at scale, and insiders were still buying stock in July, well above my original entry. My original (and conservative) fair value has been reached. This is precisely the moment where the error is selling, not holding. I will sell when the thesis breaks, not when my lowball number gets hit.

Lessons learned

1. The combination is a must. Cheap entry without quality management is a value trap. Quality management without a cheap entry is something like MIPS: admirable, ownable in theory, and priced so that the heavy lifting is already done. I know people might disagree here, but I think the risk of buying already recognized great businesses is high for someone looking to generate well-above market returns.

2. The thesis you write is rarely the exact thesis that pays. I underwrote margins at Bel Fuse and got an AI cycle. I ranked spectrum as a major EVC catalyst and Smadex paid instead. I’m not defending sloppy work, but a cheap entry with aligned management is how you stay exposed to surprises you cannot forecast. You want to get into the right businesses at the right prices and then…hope for the best!

3. Multiple expansion does the heavy lifting, and it only exists if you buy cheap. TASE went from mid-teens to 50x earnings. Bel Fuse from 12x to above 60x. That is where most of the return is, and it is mathematically unavailable when you enter at 50x. I don’t think you should model a repricing. You can only own improving businesses before the market agrees, and then not interrupt the process.

4. Sell on thesis damage, not on price targets. My fair value estimates are conservative on purpose, which makes them buy signals, not sell signals. Reaching a lowball fair value while the business keeps improving is evidence the thesis is working but not that it is over. This mistake cost me more than every losing position combined.

5. Be honest about luck. A war de-escalating, an AI capex boom, a political ad cycle. I did not forecast any of these, although I definitely hope for them to happen. The same process also got me into HireQuest and a Duolingo position still underwater. Three multi-baggers in one cohort is not the base rate. What I expect to repeat is the process: cheap entries, quality operators, holding times measured in years.

A train has to be one of the best places to write and do research

Not investment advice. Returns are approximate, measured from each article’s publication date, exclude dividends, and I hold positions in some of the names mentioned. Do your own work.

Why I like Tobila

Every country has a problem with phone scams and AI is only making the situation worse. Elderly people are the more exposed and most scams begin the same way, with a phone call.

Japan’s answer of police campaigns, carrier warnings and, increasingly, software. Sitting quietly at the center of that software layer is a small company from Nagoya called Tobila Systems ($4441), whose entire business is built on a single asset: a proprietary, daily-updated database of fraudulent phone numbers, SMS messages and URLs.

The database is fed by call logs from roughly 15 million monthly users across all three major Japanese mobile carriers, an in-house survey team that sorts more than 5 billion calls, messages and URLs a year, and direct data feeds from the Japanese police. NTT Docomo, SoftBank and KDDI all license Tobila’s filtering technology inside their security option packs. If you have ever had a suspicious call flagged on a Japanese smartphone, there is a reasonable chance Tobila’s database made the call.

If you have read some of my previous work, specially around DUOL, you know that I have a strong belief that the value of proprietary quality datasets have only increased with AI. So, that is 4441 moat’s. The interesting part is what the market is doing with it.

Two businesses moving at different speeds

Tobila reports two segments, and they could hardly be more different.

The Security segment is the original, boring carrier business. It grew just 3.3% in FY2025, which sounds disappointing until you remember what it is: a mature, contracted revenue stream tied to carrier subscriber volumes, running at roughly 70%(!) segment margins. It does not need to grow much, Its job is to fund the other part of the business.

The Solution segment is the key part. Revenue grew 60% in FY2025 and segment profit more than doubled. The engine is TobilaPhone Biz, a box that plugs into a company’s existing phone system and adds automatic call recording, a recorded-call announcement, IVR call routing, nuisance-call blocking from a database of around 30k known bad numbers, and AI transcription and summarization on top. In the first half of FY2026, reported on June 10, the Solution segment grew 69% YoY.

Distribution matters as much as product here. Biz is sold through partners with national reach: the NTT East and West sales channels, where it holds a designated “Select Item” status, plus Otsuka Shokai and Kyocera Document Solutions among others.

In February 2026 the company extended the lineup down-market with TobilaPhone Biz Lite, a stripped-down entry model aimed at small businesses and sole proprietors targeting the potential demand from the regulation coming on October 1.

Regulation tailwinds arriving

On October 1, an amendment to Japan’s Labor Measures Comprehensive Promotion Act takes effect, obligating every employer in the country, regardless of size or industry, to implement countermeasures against customer harassment, known in Japan as kasuhara. Employers must establish policies, set up consultation systems, train staff and respond appropriately to incidents. Failures can draw corrective guidance and public naming by labor authorities. Municipalities, which are themselves covered by the obligation, have in some cases begun subsidizing small businesses that adopt anti-kasuhara equipment.

Most kasuhara arrives by telephone. It is a one-on-one interaction that leaves no record. The practical compliance toolkit for phone-based harassment is call recording for evidence, a recorded-call announcement for deterrence, IVR for screening and deflection, and call-log management for documentation. That list is, item for item, the TobilaPhone Biz feature sheet.

It looks like the market has noticed. Distributors and dealers across Japan are running seminar programs through the summer explicitly framed around the October deadline, pitching Biz as the way to comply, with case studies from municipalities and private companies already deployed. Tobila’s own webinar calendar is dense with kasuhara sessions. Management was attributing record Biz unit sales and growing inquiries from businesses and local governments to this driver before the law was even in force. The regulation now sits less than three months away.

An importatn caveat is that the law mandates countermeasures, not hardware. A company can arguably comply with a manual and a training session. The bet is not that every Japanese employer must buy a Tobila box but that a legal obligation, some national subsidies, an aggressive national sales channel and a genuinely useful product will together pull forward demand from a market of several million small businesses.

Accounting shenanigans

Biz is mostly sold as a bundle pack comprised of the hardware plus a license and maintenance contract running five to seven years, paid upfront (!). Under Japanese GAAP, that cash lands immediately but the revenue is recognized over the life of the contract. You can see in FY2025 an OCF of ¥1,752m against net income of ¥626m, with contract liabilities on the balance sheet swelling by ¥848m in a single year. Customers are basically financing Tobila’s growth by paying years in advance.

Reported FCF of roughly ¥1.6bn overstates the run rate, since part of it is the working capital tailwind from those upfront payments. Strip that out and normalized free cash flow sits closer to ¥900m to ¥1,000m. Still a remarkable figure for a company earning ¥626m of net income,and growing.

Meanwhile the income statement is being deliberately depressed. Tobila’s Medium-Term Management Plan 2028, disclosed in December 2024, targets FY2028 net sales of ¥6.0bn and operating profit of ¥1.7bn, roughly a doubling of both from FY2025, funded by a front-loaded investment program of more than ¥7.2bn over four years, all from OCF. FY2026 guidance therefore shows the awkward optics of 20% revenue growth alongside an operating profit decline, from ¥899m to a guided ¥785m. H1 delivered exactly that: revenue of ¥1,674m, +22%, with profits down on investment spend and guidance maintained.

So the market is being asked to value a company on earnings that are understated twice over: once by revenue recognition that defers 5 years of collected cash, and once by an investment program that management chose to run through the P&L now.

B/S and capital allocation

Tobila carries ¥3.0bn in cash plus another ¥0.7bn in time deposits against ¥146m of debt. Net cash of roughly ¥3.6bn against a market capitalization of about ¥16bn.

Capital allocation is not bad for a Japanese small cap. The dividend policy is a 35% payout with a ¥20 per share floor that management has committed to holding even through this investment driven profit dip. Buybacks were executed in December 2023 and again from December 2024 to March 2025, the latter retiring about 3.3% of shares, and treasury stock now sits near 5% of issued shares. The founder-CEO remains at the helm, and employee incentives were deepened with RSUs vesting in March 2029. The plan itself states an explicit ambition of a ¥25bn market capitalization or more by FY2028.

What you pay today

The stock has been rewarded for the first half, and then some. From around ¥1,230 in mid-June to around ¥1,545 now. The market cap now sits at about ¥16.5bn.

On the surface that looks expensive: about 25x trailing earnings for a company guiding profits down this year. But once you adjust the numbers for the measures that matter, EV is around ¥13bn. Against FY2025 operating profit of ¥899m, before the investment program bit, that is roughly 14x. Against normalized free cash flow of ¥900m to ¥1,000m, roughly 13x to 14x . And against the FY2028 operating profit target of ¥1.7bn, still under 8x.

One more thing to hold in mind. Management’s mid-term plan states an ambition of a ¥25bn market cap or more by FY2028. At ¥16.5bn, the market has already traveled two thirds of that distance, leaving roughly 50% to reach the goal. The question is not whether the current P/E is high but whether you believe a founder-led company with a defensible data asset, a regulatory tailwind arriving in 12 weeks, customers who pay 5-7 years upfront and a segment compounding at 60% to 70% can hit numbers it is currently spending its own reported profits to reach, because at this price you are paying for a meaningful part of that outcome in advance.

What can go wrong

Carrier concentration remains the foundation risk. The Security segment’s economics depend on option-pack relationships with Docomo, SoftBank and KDDI. Those contracts are sticky but not permanent, and any renegotiation of terms would hit the highest margin revenue in the company.

The kasuhara catalyst can disappoint in degree. The obligation is procedural more than technological. If most small employers satisfy it with a policy document and a poster, the demand pull-forward will be a fraction of what the channel enthusiasm implies. The first clean read on this arrives with Q3 results in September and, more importantly, the quarters after the law takes effect.

TobilaPhone Cloud, the cloud PBX product, remains an incubation cost. It took a ¥41m impairment in FY2025 and is still loss-making. The mid-term plan leans on it eventually working.

Execution risk on the plan itself is non-trivial. Doubling revenue in three years while nearly doubling operating profit requires the Solution segment to keep compounding at rates that are obviously challenging to sustain for any business, and the front-loaded spending only pays off if it does.

As always, sell-side coverage is essentially absent, liquidity is thin, and momentum in names like this cuts both ways. The market is clearly taking notice of the name but this could move hands before October. I have hold some small japanese companies that stayed cheap for way longer than I originally thought. We’ll see how the catalyst in October affects the name-

What is my bet

This is a monopoly-adjacent data asset that collects cash years before it reports the revenue, is reporting a depressed P&L by investing todays profits and does so 3 months before a national law obligates every employer in Japan to solve the exact problem its flagship product is built for.

The market has noticed, which is why the easy first leg is gone and part of the second is too. What remains is the question of whether the FY2028 numbers are real or even understated. If they are, today’s EV is still paying single digits for them and the company’s own market cap target implies another 50% from here. If they are not, there is a net cash balance sheet, a dividend floor and a buyback-willing management team underneath.

When you are thinking about buying something obscure, a small cap nobody knows, it can be difficult to know what the market actually thinks about it, or even if someone is thinking about the company at all. Big swings in price are sometimes just someone buying or selling and moving the volume significantly. It happened just last Friday with TVA.B a +13% move with no news at all.

You check the filings, nothing. Check the news, nothing. The obvious question is whether someone out there knows something that you don’t.

Large caps don’t have this problem. When Apple moves, a thousand analysts explain why within the hour, and they’re usually wrong in well documented ways. With small caps, price movements are not so much a matter of consensus as of presence. Whether the crowd is arriving or leaving is the key thing you want to know about your illiquid positions.

Here is how I think about temperature.

I look at temperature in three different ways

The first is flow. Actual money changing hands. The second is attention. People reading, writing, and talking about the name. The third is positioning. Who owns it, who’s accumulating, who’s stuck.

A stock can be lacking all three, which is where you might find a bargain, but it won’t matter if no one ever finds out. If it’s hot on attention and cold on flow, the story hasn’t been believed yet. If it’s hot on flow and cold on attention, it might be that a small group of people knows something that you don’t.

Confusing these three is how you misread markets. A stock going up on no volume isn’t demand. A stock going sideways on huge volume isn’t apathy. You have to know what’s going on.

Flow comes from dollars, not percentages

In a small cap, the percentage move means nothing without the dollar volume behind it. A 25% jump on 40k dollars traded is just one impatient buyer. The same jump on two million dollars traded is an institution, or several people who read the same thing. Be sure to find out what they know that you don’t.

So the first question is always: how much money actually moved, compared to a normal day? Volume at 3x the 90 day avg is something to pay attention to. Ten times is an alarm ringing. Then look at where the trades printed. Buyers who keep paying the ask are reaching. Sellers who keep hitting the bid are leaving. A tape that grinds higher on the ask, day after quiet day, with no news, is a key pattern in markets because patient, price insensitive accumulation in an illiquid stock is almost never random. Someone knows something.

Canada, for example (where the latest idea I talked about is), has a special characteristic that most markets don’t: broker attribution is public. You can see which firms executed the trades. When one broker does 80% of the buying for three weeks, that’s not a crowd. It’s a person with an idea, which is obviously different from a crowd with a mood.

Attention as the leading indicator nobody charts

Here’s the sequence that plays out over and over in small caps. Someone publishes a writeup. VIC, a Substack, a long fintwit thread. Nothing happens for a week because the audience for obscure ideas is small and thorough (normally). Then a few readers finish their own work, agree with the thesis and start buying.

Volume triples about two weeks after the writeup, and everyone watching the stock wonders what happened. What happened was published fourteen days earlier, in public and for free (most times).

Attention is the closest thing small cap investors have to a leading indicator of flow.

Ask your preferred AI tool to search for new writeups on the name, check whether the forum threads have woken up, watch whether influential people (measured by followers) are suddenly showing interest, look for engagement on a tweet or Substack post. None of this is precise, but it can help you be early.

The inverse matters too. A stock nobody has written about in three years, with a dead message board and a single uninterested analyst, is cold. Cold isn’t bad because you might get a good price, but you should know you’re buying in winter and acknowledge that it might last a long time (sometimes years).

Positioning as the paper trail

The third way to measure temperature is the slowest and the most accurate, because it comes from filings. Insiders have to report their purchases within days. Anyone crossing ownership thresholds has to announce it and short interest gets published on a schedule. Funds disclose quarterly.

In a big liquid stock this data is noise, but in a small cap it’s a small cast of key shareholders, and the most important is the one that already controls the company. When a controlling shareholder starts buying the minority shares in the open market, that is a signal.

Paying attention to silence

In small caps, the absence of news can itself be news. A sharp move on real volume with no disclosure is a key event.

Uninformed buyers need a reason, and reasons are public. So a reasonless move narrows the field to buyers who don’t need a public reason, which is a short list, and it’s a list of people who tend to be required to explain themselves shortly afterward.

Leaks before takeovers and going-private deals are one of the best documented patterns in market history. That doesn’t mean every unexplained move is something, but you should be trying really hard to know for sure.

One more thing

Temperature taking sounds like trading, and as a value investor you might be trained to move away from that. That’s a mistake I made for a number of years.

Your analysis tells you what a thing is worth, and temperature tells you when the market starts agreeing with you. Undervaluation in small caps can stay unresolved for years, so you’d better pay attention to these signals. The cheapest stock in the world, owned alone, will continue to stay cheap. When an idea gains traction, things can move quickly (quicker than your ability to increase a position).

So watch the value, but keep a hand on the thermometer. And always remember that you are buying from other people, and that you need other people to agree with you to make money.

I enjoy the idea of working analog and AI doing the digital work for me.

Why I like TVA Group - written by AI (Claude Fable 5)

The situation

TVA Group is the leading French language broadcaster in North America, based in Montreal and roughly 65 to 68 percent owned by Quebecor, the C$15bn Quebec telecom and media group, which holds all voting shares. The minority trades as non voting Class B stock. At C$1.98 the market values the company at about C$86mn, or 0.25 times book equity of C$348mn, against C$110mn to C$165mn of net cash, pension surplus and real estate alone, and a sum of the parts between roughly C$3.90 and C$6.50 per share depending on how aggressively the pieces are marked. The thesis rests on three pillars: a hard asset base worth more than the entire market cap independent of operations, a regulatory reset that could structurally lift broadcasting profitability, and a controlling shareholder with clear economic logic to buy in the minority. The stock has risen roughly 130 percent over the past year and sits near its 52 week high, so the market has begun to recognize the value; the question at this price is how much of the remaining return depends on the controlling shareholder actually acting.

The business

TVA’s durability comes from language. Quebec is a market of about nine million French speakers where global streamers command far less cultural share than in anglophone markets, and TVA’s network and specialty channels hold roughly 39 percent of provincial viewership. Broadcasting, about C$460mn of revenue, monetizes that audience through advertising and carriage fees, the per subscriber payments distributors remit for carrying its channels. The segment produced roughly C$9mn of adjusted EBITDA in 2025 excluding one time items, moving to a run rate near C$16mn after an arbitration victory over Bell repriced carriage fees upward by C$7.5mn annually and delivered a C$36mn retroactive payment. The first quarter of 2026 was the segment’s first profitable Q1 since 2021, though results were helped by a one off tax reversal and restructuring savings that will not repeat, and advertising continues its secular decline. MELS rents 20 soundstages, production equipment and post production services across more than 500,000 square feet of Montreal area property; it earned C$7.4mn of EBITDA in 2025 but is currently EBITDA negative on weak production volumes, so its value is best anchored to a real estate floor of roughly C$38mn. Magazines contributes about C$4mn of EBITDA in structural decline and is treated as residual.

The asset base

As of Q1 2026 the company held C$26.7mn of cash with no debt drawn on its C$120mn Quebecor facility, and generated C$12.1mn of operating cash flow in the quarter. The pension plan is overfunded, with a gross surplus of approximately C$83mn per the audited notes, carried on the balance sheet at only C$48.8mn because of the IAS 19 asset ceiling test. Quebec law allows a sponsor to extract such a surplus by purchasing annuities that guarantee all member benefits, waiting through a three year discharge period, and keeping the remainder, with TVA’s tax losses available to shelter the income. The company also owns its vacated 285,000 square foot former headquarters occupying a full Montreal city block, under study for residential conversion, plus an estimated 344,000 square feet of legacy property around the province whose last full disclosure dates from 2001. Marked conservatively for annuity friction, a likely social housing conversion and disclosure uncertainty, the non operating assets total about C$109mn, or C$2.52 per share; marked at face, they approach C$165mn, or C$3.82.

The catalysts and where they stand

The regulatory catalyst is live but suspended. Canadian broadcasters have long been required to spend roughly 30 percent of revenue producing Canadian content while foreign streamers escaped the obligation entirely. In May 2026 the CRTC, the federal broadcast regulator, cut traditional broadcasters to 25 percent with added flexibility and imposed 15 percent on streamers. For TVA the relief is worth roughly C$20mn annually, most of it falling straight to EBITDA, enough to triple the segment’s run rate. In June the federal government directed the CRTC to reconsider the framework amid trade friction with the United States, substituting C$600mn per year of direct sector funding and explicitly committing to protect French language production. Policy direction remains favorable; the final form is unresolved, and a full rescission could revert obligations higher.

The hockey question is unresolved and near term. TVA Sports’ NHL sublicense expired with the 2025-26 season, and national rights holder Rogers renewed its league deal at C$11bn, two and a half times the prior contract. TVA was paying about C$73mn a year for programming rights and was still negotiating renewal as of April, with a credible report suggesting it remains an NHL broadcaster. Renewal at the new benchmark would impair broadcasting economics for years; losing the rights would undermine the channel and the newly won carriage fees.

The privatization logic is straightforward. Quebecor has resolved the leverage constraints that once argued against spending, keeps integrating TVA functions through related party transactions, most recently transferring the TVA Films distribution assets to itself for C$2mn in January 2026, and could acquire the roughly 35 percent minority for C$45mn to C$75mn at plausible offer levels, more than half funded by TVA’s own cash upon consolidation. Canadian minority protection rules under MI 61-101 require an independent formal valuation and majority of the minority approval for any insider buyout, which constrains a lowball offer. A parent buy in of an already controlled subsidiary should not trigger CRTC change of control review, removing a common friction in broadcast deals. The window logic favors moving before the repriced carriage fees and any content spending relief are fully visible in reported results, though Quebecor’s attention is partly occupied by its pursuit of Corus Entertainment.

Return potential at C$1.98

Scenario analysis across combinations of NHL and CRTC outcomes keeps the sum of the parts between roughly C$4.50 and C$7.25 per share in every cell, because the non operating assets dominate the total. The operating variable that matters is Broadcasting EBITDA, which ranges from roughly zero, if the NHL is renewed at benchmark pricing while content obligations revert, to about C$30mn if relief holds and hockey is renewed on sustainable terms. What varies far more is the fraction of intrinsic value the market pays for a non voting, controlled stub. A privatization between C$2.75 and C$4.25, spanning an opportunistic offer up to the non operating asset value alone, returns 39 to 115 percent, with a C$3.50 midpoint returning 77 percent. Absent a deal, a favorable stub scenario with visible C$30mn broadcasting EBITDA supports a re rating toward C$3.25, about 65 percent upside, while a muddle through scenario sits near C$1.80, slightly below today. The bear case, combining an adverse hockey outcome, regulatory reversion, conservative asset marks and continued transfers to the parent, points toward C$0.85, a loss of roughly 55 percent. Weighting a deal within two years at 40 percent produces an expected value around C$2.50, or about 28 percent above the current price. The distribution is still favorable, but at this level the buyout is the trade; recognition alone no longer carries the return.

Risks

Quebecor has no obligation to bid, already captures most benefits of full ownership, and can transfer assets piecemeal below materiality thresholds without minority approval, as the TVA Films sale demonstrated. The NHL renewal is a binary near term event with negative tails in both directions. The CRTC relief may be diluted or rescinded on reconsideration. The pension surplus is interest rate sensitive, slow to extract, exposed to annuity pricing and potential legal contest, and accrues to minorities in a buyout only through the formal valuation. Legacy real estate square footage is unverified since 2001, and a social housing conversion of the headquarters would likely realize below market residential value. Shares carry no votes and no dividend, a single analyst covers the name, and the same individual serves as chief executive of both companies. Entry near the 52 week high offers less cushion than at any point in the past year.

What to watch

Second quarter results around July 30 provide the first clean post carriage adjustment quarter and should confirm renewal of the Quebecor credit facility that came due June 30. The NHL French rights announcement is expected before the fall season. The Cabinet and CRTC reconsideration of the content spending framework, and the allocation of the C$600mn federal fund, land in the second half of 2026. Every quarterly related party disclosure deserves close reading.

Conclusion

At C$1.98 the market pays roughly a third of a base sum of the parts, and about half of even the most conservative marks, for a debt free, cash generative company with a single well capitalized natural acquirer. The downside is protected by assets rather than earnings; the upside requires the controlling shareholder’s incentives to convert into action on a timeline no minority holder controls. It is an attractive asymmetric special situation, appropriately sized as such rather than as a core position

Competitive advantages are evolving. Companies capable of collecting unique, high-quality data can compound value through increasingly personalized and effective AI systems.

At the same time, exponential improvements in component technologies drive the cost of delivering intelligence toward zero.

When I wrote about Duolingo a couple of weeks ago I argued that the current LTV of their subscriber base (paid users) is worth approximately $6bn–$7bn, which is close to today’s EV.

I also made a conservative estimation of their LTV/CAC ratio to understand how strong their unit economics are. You must make some assumptions, but it’s easy to get to a number above 4x, meaning that the value of a new customer is at least 4 times the cost of acquiring it.

Those numbers represent a safe net for investors but say nothing about the future. Given only 9% of users are paid subscribers and that user growth continues at strong rates, it’s fair to assume that most of the company’s value will come from increased paid conversion and continuous user growth.
If I’ve learned something about Luis von Ahn, it’s that he will not be focusing on increasing conversion any time soon. You could see what Wall Street thinks about that strategic decision when looking at the 6M price chart.

Continuous user growth it is.

I believe that is the correct path for two reasons.

1) Network effects

Proprietary data will be the most valuable asset in an AI-dominated world. More data means more value for users because it allows for the creation of a better product. Duolingo has years of iterations that allow them to optimize engagement and quality by understanding what their users want and how they learn.

The main critique to the product is that it’s so gamified that it impedes learning. This is a common case of reverse causation, where an outcome is incorrectly treated as the cause. The product is set as it is after thousands of iterations through A/B tests to see what works and what doesn’t. I believe the gamification of the product is, for the most part, the result of short attention spans and users being accustomed to high dopamine hits from social media apps.

The app is a reflection of the user’s activity and preferences over the years. The result of proprietary data and AI working their magic.

Network effects describe how a product or service becomes more valuable to users as more people use it, creating a positive feedback loop. In the 80s, Metcalfe introduced the idea that the value of a network is proportional to the square of the number of connected users, which was first applied to telecom networks and then to the first platforms on the internet.

AI introduces a new kind of network effect where Metcalfe’s Law still stands. The proprietary data created from the existing customers allows AI to improve the experience of the product, which in turn attracts new users. The additional data allows for the creation of an even better and more engaging product, creative that same positive feedback loop.

1) Curve convergence1

Duolingo is set to target the private tutor market, a TAM of approximately $125bn. I believe they can do so at a high pace of adoption because of the idea of curve convergence.

Curve convergence states that when multiple component technologies improve on exponential curves while industry architecture remains frozen around outdated assumptions, it creates the opportunity for disruption with a better product.

Teaching technology works around two pivotal points:

1. The teaching figure

2. The teaching methodology and content

Over the last thousand years, teaching progressed from Scholastic text and logic instruction to Humanist learning focused on languages and the whole person, then to Comenius’s structured universal education, followed by the Enlightenment’s standardized school systems. The Industrial era scaled this into mass schooling, which later sparked Progressive, child-centered learning. Mid-century Behaviorism introduced reinforcement-based instruction, then the Cognitive Revolution shifted focus to how the mind processes information. Finally, Constructivism emphasized active meaning-making, leading to modern digital and data-driven adaptive learning.

It’s easy to see the common denominator of all these improvements. Everything orbits around the idea of changing the form and the content, but not the figure delivering it. To this point we’ve been able to improve what we teach and how we teach it, and we depend on the teaching figure to learn and apply new improvements. This progress is limited by human cognition—not by that of the student, but by the teacher!

AI being able to generate a conversational tutor personalized to each student is a change of paradigm. It tries to solve the never-ending problem of teaching quality by looking at the teacher instead of what is being taught.

I don’t believe that human connection will ever be substituted. Most of the value of education institutions relies on socialization, interchange of ideas, and teaching students the ability to learn in itself. Once that is done, I think most people will choose an AI tutor because of a) proven efficiency and b) dramatically lower cost.

This is where the curve convergence idea comes into play. Both the teaching methodology/content and the teaching figure will improve exponentially, supported by AI, while old technologies will remain stagnant or instantly introduced into the system.
Following Moore’s Law, obsolescence will be evident when the components keep improving (LLM capabilities double in capacity every 6 months) while the original architecture stays frozen.

“A” is the point of maximum tension—the point where component technologies vastly outperform existing architecture, leading to rapid displacement of the old and rapid adoption of the new.

The opportunity lies in rebuilding the system in its entirety, with fundamentally superior economics and performance from day one. Disruption at a high pace.

While the market seems fixed in playing the AI wave through AI enablers, I believe it’s much safer to bet on AI users, especially one like Duolingo, which already has a winning position in this race.

Subscribe now

The content of this webpage is not investment advice. It is for general purposes only and does not take into account your individual needs, investment objectives, and specific financial circumstances. Investment involves risk.

Syncrolift remains the core of the company and still the global leader in shiplift and ship-transfer systems, especially in naval and defense-oriented yards. Tender activity is booming; the pipeline now sits at NOK 7.4 billion — up 85% year-over-year — with roughly 70% tied to defense rearmament. These projects take time to close, but they are large, sticky, politically durable, and tied to long-term NATO modernization. Once margins normalize back toward the historical 20% range, Syncrolift alone could produce MNOK 40–90 in annual EBITDA. The striking part is that, on conservative industrial multiples, Syncrolift by itself could reasonably justify a valuation equal to or even above Nekkar’s entire current market cap. The market still treats it as a slow industrial, but the setup looks more like a multi-year defense capex boom.

The more growthy opportunity sits in FiiZK, where Nekkar owns 39%. Norway’s new Miljøfleksordningen policy is changing the economics of salmon farming almost overnight. Farmers can reclaim biomass — effectively increasing production — if they shift into closed or semi-closed systems. And FiiZK is the leader in precisely that technology. Their revenue trajectory is already steepening, new orders (including four Protectus systems post-Q3) confirm demand is accelerating, and their installed base now provides the biological proof the industry needs: stable growth cycles, no escapes, and strong lice control. If adoption continues at this rate, FiiZK could grow from roughly MNOK 100 in 2025 to MNOK 400–500+ by 2027. For Nekkar, that means a meaningful share of the upside without carrying the full capital burden. This is the part of the story the market still prices as optionality, even though regulation has now turned it into inevitability.

Around these two pillars is a set of well-positioned businesses that add both stability and upside.

Intellilift continues to compound — steady revenues, healthy margins, and growing digital traction in offshore automation through its JV with Transocean. Techano Oceanlift is moving through the typical “first project pain” phase: early cranes had cost overruns, but repeat orders of the same model now give management far better margin visibility. As floating wind and aquaculture support vessels shift from one-off prototypes into serial production, Techano is positioned to benefit. Globetech remains the dependable cash generator with high margins and recurring digital maritime work. And Skywalker — though early — keeps Nekkar exposed to the eventual automation push in offshore wind installation.

Put together, the portfolio is much stronger than the Q3 numbers imply. The company has no debt, substantial cash, and clear visibility into multi-year drivers across defense, aquaculture, and offshore digitalization.

If Syncrolift secures even a couple of the medium-sized defense contracts currently in play and margins drift back toward historical norms, the earnings power changes quickly. Layer on FiiZK’s policy-driven acceleration — which now looks more like the start of an industry transition than a niche technology bet — and Nekkar’s EBITDA could climb meaningfully, potentially 40–60% over the next 24–36 months. With that kind of earnings profile, a blended 10x multiple looks conservative rather than optimistic. On that basis, the implied value lands closer to MNOK 2.2–2.8 billion, suggesting solid upside from today’s pricing and a real chance the market begins to re-rate the company as these growth drivers start to show up in the reported numbers.

The short version is simple: Q3 looked flat, but the underlying business is anything but. Syncrolift may already be worth the entire company, FiiZK sits at the front of a regulatory-driven capex wave, and the rest of the portfolio gives Nekkar multiple ways to grow as the ocean economy modernizes. The market still views this as a slow-moving industrial. It isn’t. It’s a platform with several inflection points lining up at once.

Subscribe now

Duolingo has been properly beaten up over the last six months. The worries: slowing user growth, the company prioritizing teaching quality and long-term user growth over near-term monetization, and of course the fear that ChatGPT might disrupt language learning.

The business operates a freemium model, generating revenue from around 9% of users who convert to paid subscribers. The strategy is to build the largest pool of learners possible, leveraging more data and insights to drive efficiencies and engagement. This should attract more paid subscribers, creating a virtuous cycle of growth and investment. And they’re heavy AI users—recently on a list of top OpenAI customers.

Duolingo has historically bet heavily on technology to drive growth. For example, they invested early in text-to-speech technology, despite its robotic quality at the time, anticipating improvements in audio.

Major AI investments have accelerated content creation. They mentioned that this allowed them to introduce a new subject—Chess—in just 8 months, a process that would have otherwise taken years. AI also powers features like video calls with the character Lily, something once only achievable with a human tutor. They now teach not only languages but also Music, Math, and Chess. I believe the market is overlooking the hidden optionality of the company to introduce new verticals that have the potential to open entirely new TAMs and compete at scale.

But the real value here is their massive proprietary dataset. With over a billion exercises completed daily, they can leverage that information to build personalized and adaptive lessons. Machine learning evaluates every learner’s answer and predicts probabilities of correctness to construct “just right” lessons. A core part of the strategy is constant A/B testing—hundreds of experiments per quarter—to make small product tweaks that rapidly optimize learning and engagement.

Duolingo currently reaches ~50m daily active users (DAUs), with an estimated potential learner base of around 2 billion globally. English learners, who make up 80% of global language learners, account for 46% of DAUs—highlighting massive growth potential.

In Q3 they shifted toward long-term initiatives—prioritizing user growth over monetization—which produced a mixed market reaction. Despite what the market seem to be thinking, marketing brings people in, but product quality keeps them around.

FY25 guidance: bookings of $1,151–1,157m, revenue of $1,027.5–1,031.5m, a small 80bp gross-margin dip (AI costs), and adjusted EBITDA of $270–300m (29% margin). Metrics to watch: DAUs, MAU→DAU, ARPU, churn, LTV, CAC. Long-term EBITDA target: 30–35%.

Duolingo sees its real competition not as other apps but as tutors and institutions. Their bet is that by being cheaper, more engaging, and more effective, they can capture a huge chunk of the market. Duolingo Max is their premium layer, with AI features like video calls, “explain my answer,” and roleplay.

For a company like this, where most of the value likely depends on future cash flow generation, the key is a qualitative assessment. I normally look at what the numbers say today and how the market is valuing their current and potential cash flows. From there, we can infer their position and speculate about under/overvaluation.

Subscribe now

What the numbers are telling us

Duolingo has an EV of around $7,530m. That’s roughly 22x their LTM FCF or 8x LTM revenue. Normally that multiple would make me pass before opening a single filing, but something is clearly working: they have almost quadrupled FCF over the last four years. If they can compound FCF at around 30% for the next few years, we’d be looking at a much less painful ~10x FCF in 2028.

But 30% FCF CAGR is a big ask.

The market is worried about their focus on long-term user growth instead of user-to-subscriber conversion, which would improve metrics today. They have ~11.5m paid subs—about 9% of total MAUs.

Subscription bookings, the revenue generated by those subs, is $936m LTM, implying ARPU of around $81. A key aspect here is estimating today’s LTV of those customers. That’s tricky because they don’t disclose churn and the average contract is one year (so theoretically churn would be 100%). But if we assume no growth and no churn—meaning the subscriber number stays flat—we can calculate LTV as ARPU * gross margin / discount rate. Note that Von Ahn stated in the Q3 call that the deceleration in paid subs comes from shifting focus to user growth, not from rising churn.

This gives an LTV per subscriber of roughly $550–$600. Multiply by 11.5m subs and you get an LTV of current subscribers close to $7,000m—very close to today’s EV. The market is valuing Duolingo almost entirely on the LTV of its existing subscriber base. Six months ago the market valued the company at 3x the LTV of its subs.

Everything else—MAU base, future subs, price increases, expansion into other verticals, conversion improvements—is essentially “free” optionality.

That’s not typical for a hyper-scaling consumer SaaS brand with strong retention and massive top-of-funnel. Duolingo is expensive on a traditional FCF multiple, but not expensive when you look at underlying (and superior) unit economics. The market is basically pricing in low monetization forever.

Looking at LTM spending on sales & marketing plus R&D, you get around $422m. If you treat all of that as cost to acquire new subs (a conservative assumption, since it pushes CAC higher) and divide by the 2.9m subs gained over the LTM, CAC comes out to ~$140–$150. LTV is 4–5x CAC.

Those are outstanding unit economics and fully justify management’s focus on long-term user growth. If they increase conversion or ARPU, that’s pure upside.

Their KPI's tell a story from an initial user downloading the app, to an engaged user using it every day, to a paying sub, using the full capabilities of their product: MAU→DAU→sub conversion.

Subs/DAUs has been steadily rising, though it has slowed a bit. The market seems laser-focused on this metric because it flows straight into revenue. As usual, people miss the bigger picture. Growing conversions is less about squeezing a higher percentage of users into subs and more about growing MAUs, because scale improves their ability to iterate and make the product better. If they keep working on quality, conversion will naturally follow.

The key metric is DAUs/MAUs—a direct measure of their engagement power. In a recent interview, von Ahn said they’re close to engagement levels approaching those of social networks. If they create a product as engaging as social media while giving users the “I’m being productive” feeling, this app will take over.

Also, slowdowns have happened before (2021). Von Ahn has said their investment cycle rotates across three priorities:

1. Growing users

2. Improving teaching quality

3. Expanding paid subscriptions

The first two—where they’re focused now—hurt short-term financials but based on margin potential and unit economics, these investments likely generate the highest long-term returns on capital.

We can see how FCF margin has fluctuated over time depending on their emphasis on conversion vs user growth.

Then there’s the icing on the cake: new verticals. AI lets them launch new subjects, like Chess, in 8 months. Expect more modular subjects and new learners. They’re also trying to create a universal proficiency standard to compete with TOEFL. They expect to surpass TOEFL revenue this year, and “Duolingo score” is gaining acceptance at top universities.

Additionally, China is only 5–6% of the business but is the fastest-growing DAU market. Once Max gets regulatory approval there, subs should jump. Partnerships like Luckin Coffee boosted awareness. Given that English is the most learned language globally, China should meaningfully expand the subscriber base.

So why does this business work and might continue to do so?

I think the main drivers are:

1. Proprietary data is gold, and they own a goldmine.

It’s without question their most valuable asset. By combining user engagement data with AI, they create a self-reinforcing loop: more users → more data → better AI-driven learning and engagement → more users.

They argue generic AI models like ChatGPT cannot replicate the experience because they lack years of engagement data and a structured curriculum. Duolingo’s AI now matches or exceeds generic models while benefiting from proprietary insights, gamification, and an optimized learning loop. This sets up a winner-takes-all dynamic in AI-driven education. Generic AI can’t match their structured curriculum, years of engagement data, and gamification loop.

On top of that, Duolingo is an AI consumer, not trying to out-develop foundational models. The market may be confusing technology development with technology consumption. DUOL is an AI consumer leveraging its dataset to improve engagement. They’ll use any future AI developments, and their head start is a meaningful moat relative to non-specialized tools like ChatGPT.

2. People aren’t inherently autodidacts.

Schools exist because most people struggle with self-directed learning. Guidance, interaction, and stimuli improve comprehension.

People also don’t like to read, and that makes learning new things very hard. The motivation behind learning a new language is inherently social—a much stronger motivator than curiosity. For non-English speakers, it also meaningfully increases their economic value and opportunities.

Short attention spans and declining reading habits don’t eliminate the willingness to learn. Duolingo has found a key market necessity: enabling people to learn something traditionally requiring years of consistency and deep focus through short term stimulus and high interaction. It’s not the prettiest reason behind their success, but I think it’s the main one. This debunks the idea that people will use other AI tools to learn as its based on the principle of minimum effort. Why would I interact with ChatGPT to learn a language when I already have an app that takes care of it for me and does so in a very engaging way?

Gamification and high-stimulus engagement keep learners motivated, even as attention spans fall. Individual outcomes vary, but overall the approach systematically improves learning efficacy. They’ve shown that improvements in teaching quality lead to user growth, usually with a 6–12 month lag—much like rising NPS scores as a leading indicator of future subscriber additions.

3. A true AI teaching tutor opens an absolutely massive market for Duolingo.

Private tutoring is a ~$200–250 billion global industry, with language tutoring alone accounting for roughly half of that. An AI tutor that delivers human-level guidance at near-zero marginal cost could let Duolingo tap into a category where customers today pay hundreds to thousands of dollars per year—far above Duolingo’s current ~$80 ARPU. Even modest adoption of AI-tutor features could meaningfully lift ARPU and subscription tiers, while mass adoption could redefine their revenue model entirely, shifting Duolingo from a low-price consumer app to a scalable, AI-powered tutoring service with significantly higher monetization potential.

Long-term outlook

Duolingo is building a scalable, AI-driven education platform. Its engagement loop, dataset, and AI integration give it a durable moat. The market is confusing a temporary DAU slowdown with an existential AI threat, even as management says retention is unchanged.

Its focus on engagement, teaching quality, and user growth—supported by a massive proprietary dataset and deep AI integration—creates a compounding cycle of improvement. By prioritizing long-term value over short-term revenue, Duolingo is positioning itself to dominate global learning markets, build a durable competitive moat, and transform how billions of people learn languages—and potentially many other subjects—over the coming decade.

Taking all this into account, Duolingo looks meaningfully undervalued relative to the scale and quality of what it is building. The market is confusing a temporary, entirely normal slowdown in DAU growth with an existential threat from ChatGPT, despite management clearly stating that platform-wide retention remains strong and unchanged. Duolingo has repeatedly demonstrated an exceptional organizational ability to identify the single metric that drives the most value and to move it—something Luis reiterated in Q3 when noting they more than doubled spoken words per video call simply by focusing on it. This cyclic growth pattern is exactly how Duolingo turned itself into a cash machine, and that same capability is intact today. With just 50.5M DAUs, Duolingo is pointing at a multi-trillion-dollar opportunity as the operating system for global education.

Under even reasonable assumptions—mid-teens DAU growth, modest ARPU uplift from AI-tutor and other features, continued teaching-quality improvements that historically lead user growth by 6–12 months—Duolingo could be generating $700–900m of FCF by 2028. Applying a 20x multiple, not at all stretched for a data-advantaged, AI-native platform with best-in-class retention and a massive runway, yields a valuation in the $14–18bn range. And if Duolingo captures even a sliver of the ~$200–250bn private tutoring market through a scalable AI tutor—something their proprietary dataset uniquely positions them to do—FCF could exceed $1bn, justifying a $20–25bn valuation. In other words, the current ~$7.5bn EV prices in a no-growth, no-monetization-further future, while realistic execution returns 100%–200% upside with long-dated optionality essentially free.

If you are curious about Luis, here are a couple of recent interviews that offer a glimpse about how he sees the company and plans to positioned it for the future.

Also, I’ve gotten up to speed and had a good understanding of the potential here reading substack. Highly recommend to take a look at his DUOL deep dives.

The content of this webpage is not investment advice. It is for general purposes only and does not take into account your individual needs, investment objectives, and specific financial circumstances. Investment involves risk.

These were very predictable but solid results. Checking PRVA’s commentary on value-based practices—which they normally report a week earlier—continues to be a good leading indicator of what TOI might deliver.

Capitation markets remain a clear success, especially in Florida. September was their first positive EBITDA month, and they continue to guide toward being free-cash-flow positive by mid-2026.

What truly surprised the market to the upside was dispensary revenue. It turns out that attracting a higher volume of patients also increases pharmacy sales (obviously), but more than compensates from not selling the higher price range drugs.
This reveals a considerably under-appreciated source of upside. The industry earns roughly a ASP+ 6% margin on the drugs they dispense. Meanwhile, value-based practices aim to reduce overall costs and align incentives with payors so that:

1. Risk is shared, creating attractive demand from payors who want to offload downside risk. This is visible in TOI’s continued flow of new delegated contracts.

2. The highest-cost drugs often aren’t the most effective in improving total-care outcomes. What actually moves the needle is continuous monitoring and avoiding preventable ER visits. Their MLR continues to improve and now sits in the high-60% range, which is dramatically better than major U.S. players reporting MLRs in the high-80s to low-90s. As I’ve written before, TOI’s network of clinics operates with a cost-focused culture that allows much tighter control over total cost of care.

Remember: an MLR of ~60% means 60 cents of every capitation dollar goes to medical costs, versus ~90 cents for traditional payors.

But—and this is a big BUT—you’re still selling drugs and still earning that 6% margin. Even if you’re not pushing the highest-cost drugs, revenue is volume × price, and TOI is driving dispensary revenue through the volume side.

And just like that, dispensary revenue grew 57% YoY, now representing 55% of total revenue, driven by higher prescription volume and better pharmacy attachment within the network. One could argue that the difference in drug pricing isn’t compelling enough to sacrifice volume, especially when the business model rewards cost-effective care.

The TAM here is enormous, around $200bn, and TOI still has a lot of room to grow. Fee-for-service remains 66% of service revenue, while value-based capitation is 34%—but capitation revenue grew 39% YoY, driving 21% overall growth in patient-service revenue. The momentum is clearly coming from the value-based side and will continue to do so.

They also updated 2025 guidance, now expecting revenue around the $495–505M range—basically the $500M mark. it’s not crazy to think they could generate $600M in revenue next year, especially considering their comments on increasing demand in existing (under-utilized) markets, where costs should not rise as sharply as volume. SG&A has largely stabilized around $100M per year, so we’ll keep seeing operating leverage playing a part here. I now expect roughly $20M of EBITDA in 2026.

The big question open question here is what a normalize EBITDA margin is. I suspect it is probably in the 8%-10% range.

As a reminder, if they can sustain ~20% revenue growth (which they guided to on the call), by 2028 they could generate +$70M of EBITDA. Growth potential should still exist beyond that, given the number of markets left to expand into and recent comments about seeing interest from payors in markets where they are still not present. Expansion will impact margins, of course, but we’ve already seen very rapid operating leverage and, arguably, a very high ROIC from new-market launches. They’ve also mentioned that they would explore expansion opportunities where they can get a cash-on-cash return within a year. With that potential, applying a somewhat conservative 15× 2028 EBITDA multiple gets you to more than 3× where the company trades today.

But the biggest driver is the underlying thesis: value-based care works in oncology—and works better than fee-for-service. It aligns incentives for both payors and patients, drives demand, and delivers lower MLRs, which in turn attracts more patient populations. And with the latest OCM/EOM frameworks, we know this model only truly works at scale. TOI is now putting distance between themselves and competitors and seems like it will continue to do so.

Feel free to subscribe. Think of it as value-based care for your portfolio.

Subscribe now

In my latest post, I highlighted how cheap the company is — trading around 4x EBITDA — and how there’s a clear path toward value creation, even without a re-rating (which it will probably get in a significant way if/when the war ends). This is the only pure-play Ukrainian investment vehicle available to U.S. and international investors.

The value creation path rests on two pillars: Convergence and Digital Ecosystem.

If KYIV can leverage its mobile dominance to cross-sell fixed broadband, TV, and digital services, it can drive meaningfully higher ARPU, lower churn, and stronger margins.

Well, it turns out they’re executing faster than I anticipated. Rebased revenue was up 20%, with a massive 526% growth in digital services (Uklon acquisition), and MAUs up 49% YoY. This growth comes both from new customers and, importantly, a strong 25% increase in multiplay users. This perfectly supports my thesis that KYIV is building a super app in Ukraine — one that will generate significant cash flow and enjoy a ring-fenced competitive position in the country, comparable to Gojek, Kaspi, or WeChat.

The result? Multiplay users now contribute over 45% of total revenues.

I think the market is missing this story because super apps are typically built from a social platform that layers in payments, vendors, and services over time. Kyivstar, on the other hand, is starting from a telecom foundation — and adding mobile apps, digital healthcare, media streaming, and even ride-hailing services. The latter, through its Uklon acquisition, has been a game-changer and a key driver of recent growth (alongside strong traction across other verticals).

My thesis is playing out: ARPU continues to grow while subscriber numbers remain stable despite the ongoing war. Naturally, as users adopt more of Kyivstar’s digital offerings, churn keeps trending down.

If you compare the subscriber base of each service to Kyivstar’s 22.5 million mobile users, the cross-sell opportunity is crystal clear:

• Fixed services: 6.6m

• Uklon: 3.6m (integrated in 2025)

• Kyivstar TV: 2.1m (+130% growth)

• Helsi (healthcare): 2.5m (+50% growth)

• MyKyivstar (app): 5.2m

• B2B: 600 Cloud clients and 2,500 AdTech clients (+70% growth)

When asked about future M&A moves during the call, management said they’re interested in verticals like fintech, e-commerce, and AdTech. Still doubting this is turning into a super app?

With this level of growth, trading at under 10x FCF and sitting on a $372m net cash position, there’s clearly significant upside for investors here.

This is one to keep a close eye on. And of course, if the war ends, and hopefully it does soon, we will have both repopulation and a re-rate of the multiple.

You’ve already scrolled this far — may as well commit.
Subscribe now and pretend you discovered Nekkar before it was cool

Subscribe now

The content of this webpage is not investment advice. It is for general purposes only and does not take into account your individual needs, investment objectives, and specific financial circumstances. Investment involves risk.

For years, Entravision Communications (EVC) was seen as a small, cyclical Hispanic broadcaster, dependent on local advertising budgets and election-year tailwinds. Today, the company is emerging as a hybrid digital advertising and media platform with genuine growth momentum (triple digits in the ad-tech business), tangible asset backing (worth multiples of the current enterprise value), and management incentives directly tied to shareholder value creation.

I believe we’ll see a meaningful re-rating through Q4 2025 and 2026, driven by three main forces:
• Ongoing growth of the Advertising Technology & Services (ATS) segment
• A midterm election–driven rebound in the Media segment
• Potential monetization of the spectrum portfolio

Add to that a stronger balance sheet, ongoing debt reduction (which lowers interest costs), and $5 million in annual savings from the recent restructuring, and the setup looks very attractive.

And while we wait — we’re getting $18 million in annual dividends, translating to a 7% yield.

From Broadcaster to Ad-Tech Platform

EVC’s Q3 2025 results made the company’s transformation undeniable. The advertising segment (Smadex + Adwake) is taking over, with 104% revenue growth and operating profit up nearly 400% — the fourth consecutive quarter of accelerating performance. ATS now represents 2/3 of total revenue and virtually all of Entravision’s operating profit.

This business operates globally in programmatic advertising and AI-driven campaign optimization — essentially running digital ads automatically using data and AI. It competes with major players like The Trade Desk and AppLovin (possible acquirers down the line?). The company continues to expand its salesforce and increase engineering capacity, and scalability should improve as cloud infrastructure costs grow more slowly than revenue.

Media Segment: 2026 Midterm Tailwinds Coming In

The traditional Media business — which includes television and radio properties serving Hispanic audiences across the U.S. — faced a difficult 2025, with a 25% year-over-year revenue decline, reflecting tough comparables from the record 2024 election cycle.

The key now is that EVC is present in politically vital, Hispanic-heavy states such as Nevada, Arizona, and Florida, where midterm years have historically delivered a substantial boost to ad sales. In 2022, Media revenue grew roughly 20–25% year-over-year. If 2026 follows a similar pattern, that could mean an additional $40–45 million in revenue, most of which should flow to the bottom line given the largely fixed cost structure. Management is already investing in local and digital assets to capture that opportunity.

Spectrum: A Not-So-Hidden Asset With Real Value

It’s “not-so-hidden” because attentive investors know it’s there — but the timing of monetization remains uncertain. The value lies in the redeployment of Entravision’s TV spectrum (the recent Dish Network transaction offers a good reference point).

According to the FCC, Entravision’s licenses cover 49.5 million people. If valued using the $0.93 per MHz-pop benchmark from the 2017 FCC incentive auction, these holdings could be worth $275–300 million — a substantial figure relative to Entravision’s sub–$400 million enterprise value.

There’s no active FCC auction today, but the company stands to benefit from any renewed spectrum demand under a more industry-friendly regulatory environment. Even if it takes time, the spectrum provides a tangible valuation floor for EVC.

Management Is Taking the Right Steps

In 2025, Entravision repaid $15 million in debt, reducing annual interest expense by around $1 million. Management emphasized that deleveraging will remain a priority through 2026. With debt now around $170 million, and interest costs at roughly SOFR + 2.5–3%, buybacks may ultimately offer a higher return — but for now, this derisking phase makes sense until leverage reaches more comfortable levels.

The company also restructured its workforce and leases within the Media segment, incurring a $3.2 million one-time charge that will deliver $5 million in annualized savings, fully realized by 2026.

Together, these actions should lift EBITDA by roughly $6 million even before any revenue growth.

What’s the Value?

There are plenty of moving parts here, but even on conservative assumptions, the upside looks compelling. Treating the spectrum as a valuation floor, incremental EBITDA growth alone can drive a substantial re-rating.

Using cautious assumptions: if advertising (ATS) grows around 30%, it could generate $40–50 million in EBITDA. If the Media segment rebounds 10–15% with midterm tailwinds, that adds another $25–30 million. Layer in roughly $6 million in structural savings, and total EBITDA could reach about $80 million in 2026 — up meaningfully from the low-$60 million range in 2025.

You don’t need to apply aggressive multiples for this to work — modest re-rating alone could drive attractive returns.

Catalysts I’m Watching

1. ATS expansion – Continued growth should help the market recognize what EVC is becoming and re-rate it closer to ad-tech peers.

2. Political ad surge in 2026 – Midterm tailwinds could sharply improve Media EBITDA.

3. Spectrum monetization – Any movement here would be a MAJOR catalyst.

4. Operational leverage – Ongoing cost control and debt reduction should enhance margins.

5. Strategic actions – A spin-off or separation could highlight the disconnect between parts and whole.

Closing

The asymmetric opportunity here is very real — growth potential on one side, tangible assets on the other. Entravision looks mispriced, and if the company executes even modestly — or if a spin-off or spectrum sale materializes — the returns could be substantial.

has a good, free, and comprehensive write-up that helped me get up to speed. You should check it out.

Big Picture

Nekkar ASA is basically a bet on the future of the ocean economy — defense, digitalization, renewables, and aquaculture. It’s a small-cap Norwegian tech platform with real industrial depth: the company owns or controls several niche, world-leading businesses and has zero debt, a pile of cash, and exposure to multiple high-growth trends that are starting to move at once.

They’re sitting in a sweet spot: the world is rearming, the aquaculture industry is being forced to go sustainable, and offshore operations are going digital. Nekkar has a piece of each through its subsidiaries — Syncrolift, Intellilift, Techano, Globetech, and FiiZK.

What They Do

Syncrolift (90%) – builds giant ship-lift and docking systems for shipyards. They’re the #1 player globally, with around 90% share in naval/defense shipyard projects. Right now, tender activity is booming because of defense spending, but projects are taking longer than expected to close.

• They’ve got a NOK 7.4 billion pipeline, up 85% year-over-year. Around 70% of that is defense-related.

• Think of contracts like the Haakonsvern naval project (NOK 164m) as the medium size. If they land two or three of those over the next couple years (base case), revenue jumps sharply.

• Once margins normalize (back to 20%), Syncrolift alone could add MNOK 40–90 in EBITDA.

• Upside case: expansion into export markets as other NATO countries start modernizing naval infrastructure (+ headwinds from target of 2% GDP going towards defense expense)

Intellilift (51%) – the digital arm of the group. They make industrial automation and control systems for offshore drilling and load handling.

• Solid business with MNOK 11–19 per quarter in revenue and around 15% EBITDA margins.

• Their joint venture with Transocean (InteliWell) is getting strong traction.

Techano (90%) – designs high-end cranes and load-handling systems for renewables, subsea, and aquaculture vessels.

• The first projects had some cost overruns (classic “first project” problem), but the setup is now in place.

• New crane projects are moving ahead with better cost visibility, and they should start to show margin recovery once deliveries hit milestones. They have delivered 2 first cranes and now they have an order for two more, exactly the same, so pricing should be accurate generating an attractive margin.

• Positioned nicely for growth in offshore wind and aquaculture support vessels.

Globetech (67%) – the “steady Eddie” of the group. Provides IT and digitalization services for maritime customers.

• Pulling in about MNOK 25 per quarter with fat margins (21–28%).

FiiZK (39%) – this is the most exciting piece. They’re the market leader in closed and semi-closed cage systems for salmon farming — the key technology for making fish farming environmentally compliant.

• Norway’s new Miljøfleksordningen policy is a game-changer. It lets fish farmers reclaim previously restricted biomass if they switch to closed farming systems — and FiiZK makes exactly that.

• Their revenue’s already exploding: MNOK Q4 11 → Q1 24 → Q2 41 in recent quarters.

• In a base case, FiiZK could grow from ~MNOK 100 in 2025 to MNOK 400+ by 2027, doubling year after year as adoption takes off. Nekkar’s 39% stake means it gets meaningful upside exposure without having to fund it fully.

• In an upside case (fast adoption + export orders), FiiZK’s revenue could easily hit MNOK 500+ by 2027.

Skywalker – early-stage wind turbine installation system. Still in R&D, but could open another growth angle in offshore renewables.

The Numbers

• Market cap: MNOK 1,200

• Cash: MNOK 225

• Undrawn credit: MNOK 200

• No debt

So, you’ve got a debt-free balance sheet, growing pipeline, and multiple business lines that could all inflect higher within 18–24 months.

What Could Happen (Base Case)

If Syncrolift lands just two mid-sized defense projects (~MNOK 200 each) and margins normalize to 20%, you’re looking at MNOK 80+ incremental EBITDA there.
Add in a steady ramp at FiiZK (maybe MNOK 80–160 in Nekkar’s share of revenue by 2026–27), and suddenly you’re seeing 30–50% EBITDA growth over the medium term.

At a 10x multiple, which is what you’d expect for a capital-light tech-industrial mix, that points to MNOK 2.0–2.5 billion in fair value — roughly 50–80% upside from where it trades today.

Why It Works

• Defense tailwinds: rearmament = more naval shipyard upgrades, and Syncrolift owns that niche.

• Aquaculture reform: regulation forces innovation, and FiiZK is the market leader.

• Digitalization trend: Intellilift and Globetech keep the tech angle alive.

• Balance sheet: debt-free, cash-rich, flexible. M&A as part of the strategy for growth (target of NOK 2bn+ in 2027 through existing companies + 1-3 new buys). Important to mentioned very high ROIC on past M&A decisions.

• Execution leverage: when projects start to flow again, margins swing fast.

Risks to Watch

• Timing risk: Defense contracts are unpredictable — projects can take forever to close.

• Execution risk: Techano and early FiiZK scale-up could still have cost bumps.

• Minority exposure: FiiZK is only 39%-owned, so Nekkar doesn’t fully control it.

Bottom Line

Nekkar is a clean, cash-heavy platform with multiple structural growth drivers lining up — defense, aquaculture, renewables, and offshore tech. 2025 might still be a transition year, but from 2026 onward, there’s real potential for revenue and EBITDA to inflect.

At this valuation, you’re basically getting the optionality on FiiZK’s regulatory windfall and Syncrolift’s defense boom for free.

If execution stays on track, Nekkar looks like a re-rating story waiting to happen.

You’ve already scrolled this far — may as well commit.
Subscribe now and pretend you discovered Nekkar before it was cool

Subscribe now

With profitability on the cusp and a vast growth runway ahead, TOI stands out as one of the most compelling long-term investment opportunities I’ve encountered in a while.

Challenges in the US Healthcare System: Focus on Oncology

The US healthcare system places a significant financial burden on families, with unexpected medical bills being the top financial concern for American adults. Approximately 46% of US adults find it difficult to afford healthcare costs, and 25% report that they or a family member struggled to pay for healthcare in the past year. Prescription drug costs exacerbate this issue, with 21% of adults unable to fill prescriptions due to cost and 23% opting for over-the-counter alternatives.

Among healthcare challenges, cancer care is the most expensive. Experts project that cancer care costs will rise from approximately $200 billion today to $245 billion by 2030, raising critical questions about affordability and access for patients nationwide. Oncology drug costs, which account for roughly 75% of oncology expenses, have been growing at an unsustainable rate for over 16 years, driving high out-of-pocket costs for patients and unsustainable expenses for payers (insurers, Medicare, Medicaid).

Specialty care, such as oncology, has emerged as a key driver in reducing healthcare spending. Unlike primary care, which addresses common illnesses and prevention, specialty care focuses on specific diseases or body systems. Value-based care (VBC) models, which aim to deliver high-quality care at lower costs compared to traditional fee-for-service (FFS) models, are gaining traction. In FFS, physicians are reimbursed based on the volume of services provided, often incentivizing overutilization. Currently, only a small fraction of specialty care operates under risk-bearing VBC arrangements, but oncology, as a high-cost specialty, stands to benefit significantly from VBC’s cost-saving potential through:

• Reduced variation in diagnostic testing

• Standardized chemotherapy, procedures, and radiation choices

• Shifting appropriate infusions (e.g., lower-risk chemotherapy) to home settings

• Earlier diagnosis and intervention to reduce avoidable ER visits, hospitalizations, and readmissions

This thesis posits that VBC models will dominate the oncology care market over the next 10–20 years, with The Oncology Institute (TOI) positioned to capture significant growth.

Origins of Value-Based Care (VBC) in Oncology

Oncology spending in the US surged in the early 2000s due to expensive drugs, complex care coordination, and rising cancer incidence. Studies revealed wide variations in cost and outcomes across patients. The Affordable Care Act (ACA) of 2010 addressed these challenges by establishing the Center for Medicare and Medicaid Innovation (CMMI) within CMS to test alternative payment models focused on improving quality and reducing costs. The ACA emphasized accountable care organizations (ACOs) and bundled payments, laying the groundwork for value-based reimbursement.

The Oncology Care Model (OCM), running from 2016 to 2022, was the first major national effort to apply VBC principles to cancer care. It aimed to improve care coordination, patient experience, and outcomes while controlling costs for Medicare beneficiaries receiving chemotherapy. The OCM included monthly care management payments (fixed per patient per month) and performance-based payments tied to quality metrics. Approximately 200 oncology practices and 16 commercial payers participated. The program showed mixed results: it improved care coordination and patient satisfaction but failed to achieve sufficient cost reductions, resulting in net losses for Medicare. High drug costs, a fundamental flaw in traditional oncology care models, were a key challenge.

In mid-2023, the Enhancing Oncology Model (EOM) succeeded the OCM, focusing on seven common cancer types with stricter requirements for health equity, data reporting, and patient navigation. Monthly payments were initially reduced from $160 to $70 but later increased to $110 in an ongoing second round, with the program set to run until 2030. Early reports indicate net losses for Medicare in the first six months for only some participants, but with potential for positive results. The primary source of immediate cost reduction is Part B systemic cancer therapy spending (injectable or infused drugs, not self-administered).

Drug costs remain the primary pain point in oncology care. In the FFS model, providers purchase drugs and are reimbursed at the average sales price (ASP) plus 6%, creating an economic incentive to overprescribe, especially newer, more expensive drugs. In contrast, VBC models incentivize cost reduction by aligning performance incentives with overall cost savings, benefiting both patients and payers. Major FFS players like AON, OneOncology, and the US Oncology Network rely heavily on drug margins, with physician autonomy and referral-driven growth leading to overutilization, including off-label or off-pathway medication use that increases costs without evidence-based benefits.

The end of the OCM and the launch of the EOM are landmarks that highlight the urgent need for a sustainable solution to oncology care costs in the US. My thesis is that TOI has now acquired enough scale that it’s gaining distance from competitors as they win big populations of patients through already existing satisfied payors.

How The Oncology Institute (TOI) is Positioned for VBC Success

The Oncology Institute (TOI), the largest oncology community practice in the US with over 60 practices in Southern California, Arizona, Nevada, and Florida, has demonstrated success in VBC. For 18 years (four as a public company), TOI has reduced costs by adhering to evidence-based prescribing, following National Comprehensive Cancer Network (NCCN) guidelines, and closely managing patients to avoid unnecessary ER visits. TOI’s 41% revenue CAGR reflects strong market feedback. As one of the most successful participants in the OCM, TOI serves payers (insurers, Medicaid, Medicare) through fixed-payment contracts, actively managing medical costs and sharing risk with payers, eliminating incentives to overprescribe.

TOI’s scale gives it a competitive edge, securing large patient populations through satisfied payers. Unlike other VBC providers, TOI unifies care delivery and management services. Other providers fall into two categories:

1. Benefit Managers (e.g., Evolent Health): These manage independent practices but face misaligned incentives, as physicians earn ASP +6%, limiting cost savings to 4–5%. Evolent recently sold its Care Partners unit, which supported practices in the broader Medicare Shared Savings Program (MSSP), to Privia Health for $100 million in cash and $13 million contingent on 2025 MSSP performance, implying a valuation of $940 per attributed life or 10x EBITDA. This move signals Evolent’s focus on complex specialty VBC.

2. Care Delivery Providers (e.g., AON, OneOncology, US Oncology Network): These FFS players face overutilization due to financial incentives to prescribe expensive drugs.

TOI employs clinicians and operates its own medical and radiation oncology practices while wrapping management services organizations (MSOs) around independent providers. This structure enables:

• Higher compliance with value-based prescribing (NCCN guidelines)

• Enhanced delivery of ancillary services like clinical trials and transfusions in-house

• Engagement with independent practices, appealing to health plans

This results in more competitive rates, predictable cost savings, and lower medical loss ratios (MLR) in risk-bearing contracts, outperforming industry benchmarks.

TOI’s Journey and Growth Strategy

Founded 18 years ago in California, TOI spent 15 years building its VBC business. While 35% of its revenue still comes from FFS clients, this segment is shrinking as VBC grows. Three years ago, TOI expanded into Florida, Nevada, and Arizona, targeting markets with heavy Medicare presence to capitalize on capitated (risk-sharing) contracts. This expansion required significant capital expenditure, leading to negative EBITDA in 2021 and 2022 and a 2022 restructuring that cut $11 million in SG&A costs.

Despite these challenges, TOI achieved record success in securing capitated partnerships in new markets, with revenues growing rapidly. With infrastructure now in place, new partnerships require minimal additional capital, enabling TOI to guide for break-even or positive EBITDA in Q4 and beyond. In its October 2025 investor presentation, TOI noted ongoing interest from payers in new markets but emphasized its focus on existing markets, particularly Florida, where incremental margins are high. New market entries will only occur if cash recovery is achievable within a year, reflecting lessons from the 2022 restructuring.

Financial Drivers and Outlook

TOI’s investment thesis centers on transitioning from a non-profitable to a profitable company while maintaining strong growth. Key drivers include:

1. New Value-Based Contracts: In 2024, TOI added $48 million in annualized revenue from new contracts and expects similar growth moving forward.

2. Dispensary Revenue Growth: TOI’s pharmacy business, tied to Part D oral drugs, accounts for ~46% of revenue. Historically, every $1 of capitated revenue generates $0.9 in dispensary revenue. Patients prefer TOI’s pharmacies for accessibility, and TOI earns a 6% margin on drugs sold without incentivizing expensive prescriptions.

3. Volume Productivity: With capacity built, new revenue significantly boosts gross margins, driven by high payroll productivity per visit.

4. Clinical Trials: This smaller segment provides access to underserved demographics, reduces reliance on capitated drugs, and lowers overall business risk.

TOI’s business comprises two main segments:

• Patient Services: Includes FFS revenue (35%, growing at ~10% annually, in line with the industry) and capitated revenue (17%, with higher margins and faster growth).

• Drug Dispensing: Contributes 46% of revenue, growing in line with patient services ($0.9 per $1 of patient service revenue).

• Clinical trials, though smaller, are highly accretive with ~90% margins.

The average per-member-per-month (PMPM) payment is higher in Florida and Nevada than in California, allowing TOI to achieve strong revenue growth with fewer new customers. In the last four quarters, TOI grew its value-based lives by 18%!.

Impact of Drug Pricing Reforms

Drug pricing reforms (e.g., Medicare negotiation, manufacturer rebates, portals like “TrumpRx”) will impact FFS and VBC differently. FFS practices, reliant on “buy and bill” margins, will face immediate challenges. VBC practices like TOI, which must meet spending benchmarks to earn shared savings, are better positioned but could face pressure if benchmarks reset too quickly.

Valuation

If TOI achieves 7% market penetration in Florida and Nevada (matching California) over the next few years, it could add ~$150 million in EBITDA. Dispensary revenues could contribute another $100–130 million, given historical trends. Assuming SG&A stabilizes at $100 million without new market expansion, total segment margins could reach 20% (10% in California, 30% in Florida, 20% in Nevada). Patient services margins trend toward 20% (increasing with capitation), dispensary margins hold at 20%, and clinical trial margins remain at 90%.

With a conservative 20% growth rate, TOI could achieve ~$70 million in EBITDA by 2028. With cash creation by then and assuming no net debt, applying a 15x EBITDA multiple (compared to OneOncology’s 19x in 2023 and Evolent’s 10x for its Care Partners sale), TOI’s valuation could reach $1,000–1,050 million by 2028, significantly above its current $480 million EV.

The content of this webpage is not investment advice. It is for general purposes only and does not take into account your individual needs, investment objectives, and specific financial circumstances. Investment involves risk.

The company is a U.S.-based OEM specializing in avionics. It designs, develops, manufactures, and supports a wide range of systems for commercial, business/general aviation, and defense platforms.

Its core focus is integrated fleet modernization and retrofit—upgrading older aircraft to improve functionality, meet regulations, extend useful life, reduce downtime, and lower total cost of ownership. In other words, instead of replacing entire fleets, operators can modernize with ISSC.

The stock has been volatile in recent years due to acquisitions and the resulting swings in growth and margins. The thesis here is basically margin expansion plus further inorganic growth, leading to a rapid compression of the forward multiple and positioning ISSC as a potentially underappreciated, cheap, high-quality company.

Macro tailwinds:

• Fuel costs are a huge airline expense. Higher prices make older, less efficient aircraft costly to operate, which accelerates retirements or retrofits. Instead of replacing entire planes, operators often upgrade avionics for efficiency and compliance. That’s a win for retrofit OEMs like ISSC.

• On the defense side, with the F-16 products and military exposure, you’re essentially long global conflict and defense spending.

Recent acquisitions

A key turning point was June 2023, when ISSC entered an Asset Purchase and License Agreement with Honeywell. They acquired inventory, equipment, and customer docs, plus the assignment of certain contracts. Practically, this was Honeywell exiting certain products while ISSC picked up stranded assets still valuable to customers. The relationship worked—there have since been two more Honeywell deals (July and Sept 2024), the latest tied to commercializing F-16 products (a boost to the military segment). Honeywell still needs its customers to be supported, so ISSC fills that gap.

The September transaction hit gross margins because military now makes up more of the mix, but management guided margins normalizing to the mid-40s. By FY25 (ending Sept), ISSC should be at ~$80m sales, gross margins above 40%, and EBIT around $15m—about 13x.

Valuation looks cheap, especially since they secured new $100m financing terms in July 2025. The 2023 Honeywell deal was $36m, with follow-on deals of $4m and $14m in 2024, fueling significant growth. Sales are expected to climb ~130% from FY23 to FY25E. Deleveraging has been rapid—already down to 1.1x net leverage (vs. the 3x they are willing go-up to).

Future growth opportunities

Management says they’re actively scouting more stranded assets and small avionics acquisitions. They’re also expanding their Exton facility, which will support up to $250m in sales (~3x current levels). Integration costs and interest will tick up, but the P&L has room, and scale benefits are already showing in SG&A and R&D.

I think another acquisition is coming soon and could re-rate the stock once the market catches onto their scaling strategy. Integration risk is the main concern—management admitted the first Honeywell deal was more complicated than expected. But they’ve since built stronger capabilities, and remember, these aren’t “traditional” acquisitions; they’re buying product lines and inventory, then manufacturing sub-assemblies locally for existing customers.

That’s why I think this is “easy” growth. There’s plenty of stranded asset opportunity in the market. Each acquisition not only adds product lines but also drives cost synergies and better utilization of ISSC’s engineering team, while opening up cross-sell potential.

What if I told you there is a dominant player in the Ukrainian telecom industry (!!) that is growing at high double digits (37% in H1 YoY), while maintaining one of the highest EBITDA margins in global telecom (58% in Q2 2025), and building a digital services platform that already represents 10.7% of total revenue, growing 460% YoY?

Of course, it’s cheap — around 3.5x EV/EBITDA — for obvious reasons, like the ongoing war that challenges the very existence of the country where this company operates (talk about risk!). Still, there are plenty of reasons to believe it will do well if/when the war ends, or even if the conflict stabilizes. I’m no portfolio manager, but this is the kind of idea that should be sized carefully within a portfolio. And just to be clear: this is not investment advice.

The Company

Kyivstar is the leading telecom operator and one of the most recognizable brands in Ukraine. With ~22 million mobile subscribers (~47% share) and 1.1 million broadband subscribers (~14% share), it is the country’s #1 telecom provider and a digital services leader through Kyivstar TV, Helsi (digital health, 29m registered patients), cloud & big data services, and most recently, Uklon (ride-hailing and delivery).

Market share has been stable over time.

The company generated $589m of Adj. EBITDA LTM on a $2.5bn market cap with net cash. They recently listed on Nasdaq (via Cohen Circle Acquisition Corp. SPAC), making it the only pure-play Ukrainian investment vehicle available to U.S. and international investors.

The public float is just $227m, so don’t expect it to be very liquid. The parent company, VEON, is a multinational telecom and digital services operator with businesses across several countries.

History

Kyivstar was founded in 1994 under the name Bridge and was one of the first mobile networks in Ukraine. Initially focused on mobile services, it broadened its offerings throughout the 2000s by investing in mobile data and modernizing infrastructure to support the rollout of 3G and later 4G LTE. In 2005, it was acquired by VEON, which has been listed on Nasdaq since 1996.

In recent years, Kyivstar has diversified through strategic acquisitions. In August 2022, it bought a stake in Helsi, a digital healthcare platform, and in May 2025, it increased its stake from 70% to 98%. Helsi generated $5.1m of revenue in Dec 2024, growing 42% YoY.

In December 2022, Kyivstar spun off — while retaining full control — Kyivstar.Tech, which provides services from UX and product development to vendor management and infrastructure. These are mainly intra-company services ($24m out of $24.2m in 2024 were intra-group).

In December 2023, it acquired LanTrace for $2m to expand broadband services.

In April 2025, it acquired 97% of Uklon for $155.2m. More on Uklon later.

Why This Might Work: Convergence + Digital Ecosystem = ARPU Growth

Kyivstar’s strategy is clear: leverage its mobile dominance to cross-sell fixed broadband, TV, and digital services — a model proven across Europe to drive higher ARPU, lower churn, and stronger margins.

• Multiplay penetration is accelerating: 6.5m multiplay users (+24% YoY), now 31.7% of MAU.

• Digital users are expanding rapidly: 13.4m in Q2 2025 (+51% YoY, +12.5% organic).

• Fixed-mobile convergence is deep: 83% of fixed clients also use mobile; 34% also take TV.

• Recent acquisitions strengthen the ecosystem: Helsi (healthcare), LanTrace (broadband), Uklon (ride-hailing), alongside Kyivstar TV and cloud/AdTech offerings.

This is essentially a super-app strategy built on a telecom base — mirroring models like Kaspi or Tencent. Traditionally, super apps emerged from banks (payments, credit), but Tencent shows it’s possible to grow from other foundations (like social). Kyivstar demonstrated this with Uklon, a ride-hailing and delivery platform that can now be offered to its 24m existing telecom customers.

ARPU Gap = The Core Upside

Ukraine’s mobile ARPU remains far below regional peers — $2.8/month vs. $10.7/month across CEE operators.

Even modest convergence would imply 3x+ revenue uplift potential.

Structural drivers are now in Kyivstar’s favor: GDP recovery post-war, the shift from prepaid (76% of subs) to postpaid contracts, and multiplay expansion.

Pricing power also remains largely untapped — management acknowledges that ARPU has been “held back by the war.”

Kyivstar’s ARPU already grew 34% YoY in H1 2025 to $3.4, even with customer loyalty discounts in the prior period. Once macro conditions stabilize, Kyivstar can deploy pricing as an additional lever on top of convergence.

“We believe there is significant room for additional ARPU growth that has been held back by the war.”

“We believe the average mobile ARPU in Ukraine is relatively low and has the potential to increase to levels comparable to ARPU levels in Central and Eastern European nations. In the nine months ended September 30, 2024, our average monthly ARPU in our mobile services business was $2.8, compared with $10.7 across MNOs in selected Central and Eastern European countries. We believe that with favorable economic developments and renewed GDP growth in Ukraine, this ARPU increase is possible in the medium term.”

Why ARPU Has Historically Been Lower in Ukraine

The short answer: Low GDP and high prepaid use.

ARPU has long been significantly below Eastern European peers because of systemic structural factors:

• Lower purchasing power

• Heavy prepaid & multi-SIM usage

• Later, lower-value data monetization

• Intense price competition

• FX effects that worsen USD comparisons

These factors only shifted incrementally in 2020–21 (data adoption rose, but prices stayed low), and by 2025 the structural picture remains similar. Operators are pushing higher-value bundles and fixed convergence, but income gaps, customer behavior, and competition still constrain ARPU growth.

That said, this is the same starting point seen in many other countries. Heavy prepaid use inflated SIM counts (the denominator of ARPU) and dragged ARPUs down historically. As postpaid adoption grows and bundled services spread, ARPU should climb significantly — as it is now — but there’s still a long runway. Keep in mind: this transition has been a global telecom trend.

Financial Strength and Profitability

• H1 2025 revenue growth: +37% YoY (+23% adjusted for cyberattack impact).

• H1 2025 Adj. EBITDA growth: +40% YoY (+17% adjusted).

• Margins: 57% EBITDA margin, 58% in Q2 2025 — among the highest globally.

• Capex: EBITDA – Capex = $309m.

• Balance sheet: $450m cash, no external debt.

• Digital: 10.7% of revenue, growing 460% YoY, led by Helsi and Uklon.

Growth Strategy

1. Telecom

• Maintain leadership and expand multiplay.

• Drive consistent ARPU growth through postpaid conversion and pricing.

• Expand fixed broadband organically and via M&A in Ukraine’s fragmented market (~3,000 providers).

Mobile market:
On the consumer side, customers are either postpaid (monthly contracts) or prepaid (pay-as-you-go). Postpaid entails higher stickiness and recurring revenue, while prepaid is more volatile. As of Dec 2024, 76% of B2C customers were prepaid, highlighting a significant opportunity to transition users to recurring postpaid contracts.

On the B2B side, 11% of 3-month active subs were B2B (as of Dec 2024), with 8% from large enterprises and 3% from SMEs.

Infrastructure is owned by VEON’s tower company, UTC, which was created through a Kyivstar carve-out and is now Ukraine’s largest towerco. Together they own 15,500 sites, with only 868 in occupied territories.

Kyivstar is the anchor tenant for nearly all UTC sites under a 7-year Material Lease Agreement (with renewal options). Terms are favorable: a 20% anchor fee discount if a third party joins, and an additional 30% discount if leasing over 5,000 towers.

Broadband market:
The market is extremely fragmented (~3,000 providers), with the top three holding ~24% share. Global telecom trends suggest consolidation is inevitable, as larger players invest in networks and roll out bundled services.

“This fragmentation presents a significant opportunity for consolidation, allowing us to expand our market share and extend our reach to both B2B and B2C customers. We are well positioned to capture this opportunity by leveraging our mobile network and our established presence across almost 43,500 broadband-connected buildings.”

Kyivstar leads with 1.1m customers (14.1% share) versus Ukrtelecom (5.4%) and Lifecell (4.8%). Recent consolidation has already seen four providers acquired since 2021.

Kyivstar also has strong FMC (fixed-mobile convergence): 83% of fixed clients also use mobile; 34% also take TV. In 2024, it made its first FTTH acquisition (LanTrace for $2m). As of June 2025, it had 44.4k connected buildings.

2. Digital

• Monetize vast customer base by cross-selling healthcare, TV, ride-hailing, cloud, and AdTech.

• Continue strategic acquisitions in complementary industries (healthcare, cloud, mobility, broadband).

• Use data aggregation and personalization to create network effects.

Capex / Infrastructure:

• 96% LTE coverage with 16k sites; partnership with Starlink for satellite direct-to-cell.

• $1bn investment plan (2023–27) with parent VEON to strengthen infrastructure and resilience.

“We are focused on sustaining market leadership and expanding market share by enhancing our digital offerings. A key component of this strategy is driving digital revenue growth through strategic acquisitions and the development of additional products. We are particularly focused on M&A in industries such as broadband, fiber optics, online healthcare, cloud technologies, and ride-hailing.”

Uklon

Kyivstar acquired 97% of Uklon for $155m. The company generated $67m in 2024 revenue and grew at a 29% CAGR from 2021–2024. It has over 100k drivers and is a leading delivery platform. In 2023, Uklon also expanded into Uzbekistan, a market with 37m people.

Uklon’s ecosystem impact is underappreciated. Beyond being a new revenue stream, it generates valuable data that can feed into product development, personalization, and cross-selling across Kyivstar’s platform.

• MAU growth: +40% YoY

• Rides booked: +19.6% YoY (H1 2025)

• Deliveries completed: +35% YoY (H1 2025)

• Revenue contribution: $21.7m, making digital 10.7% of total revenue (+460% YoY)

Helsi

Kyivstar now owns 98% of Helsi, a digital healthcare platform that supports doctors and institutions while improving patients’ access to care through remote consultations, appointment booking, and medical data storage.

• As of Jan 2025: 29m users

• 2024 revenue: $5.1m (+42% YoY)

Impact of War: Headwinds Now, Optionality Later

• 3.1m subscribers lost, but 1.2m retained abroad through roaming offers

• 5.3% of network destroyed, with 82% already restored

• Inflation, energy tariffs, and FX volatility pressured results

Yet macro recovery has begun: GDP grew +5.3% in 2023 and +2.9% in 2024. Longer term, post-war reconstruction will accelerate demand for both connectivity and digital services.

Investment Case

1. Structural ARPU Re-rating

   • Path to ~3x revenue uplift as Ukraine converges with CEE peers.

2. Super-App Ambition

   • Telecom foundation plus Helsi, Uklon, Kyivstar TV, and cloud drive ecosystem stickiness.

3. Market Consolidation Tailwinds

   • Highly fragmented broadband sector where Kyivstar has the balance sheet and scale to lead.

4. Best-in-Class Profitability

   • 57–58% EBITDA margins, with further upside as ARPU expands.

5. Unique Pure-Play Exposure

   • The only U.S.-listed Ukrainian operator — a rare, asymmetric way to invest in post-war recovery.

6. An end of the war will mean a substantial rerating of EBITDA multiple

   • Trading at comparable eastern European telecom multiples would mean a 2-3x.

Conclusion

Kyivstar is more than just a telecom company — it is building a digital super-app for Ukraine. With unmatched scale, industry-leading profitability, and a balance sheet ready to support acquisitions, Kyivstar is positioned to turn Ukraine’s economic recovery into exponential ARPU growth.

If ARPU converges even halfway to regional peers, revenue and EBITDA could double — delivering one of the most compelling margin-expansion and re-rating stories in emerging markets telecom today.

The content of this webpage is not investment advice. It is for general purposes only and does not take into account your individual needs, investment objectives, and specific financial circumstances. Investment involves risk.

TOI was founded in 2007 as a collection of oncology practices in California. Today, it operates as an oncology benefit manager with the care delivery infrastructure of both a provider and a pharmacy. The company partners with managed-care providers and other risk-bearing entities, mainly insurance companies (payors), to treat cancer patients.

Its model is based on gain/loss sharing or capitation contracts, where payors pay a fixed per-member-per-month (PMPM) fee. TOI’s network consists of 66 clinics owned by affiliated physicians, 14 independently owned clinics, and several unaffiliated clinics. Most are physician-owned because, in most states, laws require physician-related treatment companies to be owned by physicians. TOI actively manages both its patients and its clinics.

The oncology treatment model in the U.S. has been flawed for decades. It relies on a Fee-for-Service (FFS) system in which providers are reimbursed per service delivered. This structure added to the recent surge of oncology treatment costs to unsustainable levels has created a national problem that TOI directly addresses.

The Market Problem

The U.S. oncology market is roughly split 50/50 between hospitals and community clinics. Of the latter, about 40% of the market remains unconsolidated.

The flaw lies in drug prescriptions. Roughly 75% of oncology costs are attributable to drugs, and providers can earn a 6% margin on them—creating an incentive to overprescribe. Oncologists are reimbursed on a “cost-plus” basis, meaning the higher the price of the drug, the higher their earnings, regardless of whether it is the most appropriate treatment for the patient.

Major players such as AON, OneOncology, and USOncology allow their physician partners significant autonomy and depend heavily on the 6% markup on drugs. Their business growth is driven by physician referrals, while TOI, in contrast, secures patients directly through payors.

By introducing fixed-payment contracts and actively managing medical costs, TOI removes the incentive to overprescribe. One might assume this would reduce quality of care, but in practice, the opposite is true. Evidence-based treatment and close patient management are more efficient and effective. TOI follows the National Comprehensive Cancer Network (NCCN) guidelines, which are strictly evidence-based.

Instead of relying on physician referrals, TOI wins contracts directly from payors, securing populations of patients.

Why TOI’s Model Works

TOI’s model aligns incentives across all stakeholders:

• Payors benefit from gain/loss sharing contracts, which typically evolve into capitation contracts once trust is established. These arrangements transfer the risk of cost overruns from insurers to TOI, making the model attractive to HMOs and insurance groups.

• Patients receive closer care management, resulting in higher satisfaction.

As a result, payors—including Anthem, CareMore, Heritage Provider Network, and OptumCare—are steadily shifting from traditional FFS to value-based care. More than 50% of TOI’s 2024 revenue came from payors, with TOI serving as the exclusive or preferred provider for many of them. Its largest customer accounts for 13% of revenue.

The company’s recent expansion into Florida has validated its model, driving strong revenue growth over the last four years. Medicare, its primary customer in Florida, offers higher reimbursement rates and greater flexibility than Medicaid in California, making Florida a more profitable and attractive market.

The system is working at a very rapid pace. They have gain up to 13% of their client base in a single quarter.

In contrast, FFS providers often hire third-party benefit managers to review and authorize treatment requests. However, these managers lack the tools and alignment to achieve cost control on the scale TOI provides. Only a value-based culture established at the physician level can consistently align costs and care.

As TOI explains:

“We distinguish ourselves from other managed oncology practices and specialty benefit managers in our ability to align incentives across the care continuum, including physicians and payors, in delivering high-quality care at lower costs. We believe there are currently no other value-based oncology management companies of meaningful scale in the U.S.”

Although roughly 50% of patient services revenue still comes from FFS contracts, TOI is steadily transitioning them to value-based models.

On top of that, TOI also has the pharmacy revenue source. As you might expect, patients will normally go to your pharmacies, if you have them, because of convenience.

As mentioned, while FFS models encourage overprescription because of drug margins, TOI’s pharmacy business remains strong even under evidence-based prescribing, since patients still purchase medications from TOI-owned pharmacies. This creates a healthy correlation between patient services revenue and pharmacy revenue. In fact, in Q2 2025, pharmacy revenue exceeded patient services revenue, though historically the ratio has been closer to 0.9:1.

Clinical trial revenue remains immaterial and is being shifted to equity method accounting, so it will no longer appear in P&L results.

History and Financial Background

Despite steady revenue growth, TOI has faced financial turbulence.

In 2018, it secured outside funding, and in 2019, investor and board member Brad Hively became CEO. The company began expanding aggressively, particularly in California. In 2021, TOI entered Florida, though 90% of revenues still came from California at the time. SG&A expenses ballooned, forcing the issuance of $110 million in convertible senior secured debt in August 2022.

The debt carried strict covenants: minimum quarterly annualized net revenue of $200M in 2023, $300M in 2024, and $400M in 2025, as well as a minimum cash balance of $40M.

Continuous expansion efforts into Texas and Florida in 2023 drove OPEX higher due to new clinic buildouts and marketing. That year, Dan Virnich, replaced Hively as CEO and became Chairman. Hively moved back to the board. The change reduced costs tied to having both a President and CEO.

Meanwhile, revenues and gross profit continued to grow, while SG&A began to scale down. In Q3 2023, TOI announced its first full-risk capitated contract in Florida with a national payor, a major validation of its model.

In 2024, cash burn reached $10–20M per year, but management maintained a clear path to profitability, with revenue growing more than 20% annually. Although rising drug prices pressured pharmacy margins and lower utilization in Florida hurt patient services, revenues still increased 24% YoY. Cash fell to $46M by mid-2024, raising covenant concerns and driving the stock down.

In early 2025, TOI paid down $20M of the convertible debt to renegotiate and remove the $40M cash covenant, and raised $16.5M in equity. With default risk reduced and management guiding for positive adjusted EBITDA and free cash flow (FCF) in Q4 2025, investor confidence has returned.

Notably, revenue growth never slowed during these challenges.

Why the Opportunity Exists

At ~$325M market cap, TOI does not appear optically cheap. It is expected to generate ~$480M in revenue in 2025 with around negative $13M in adjusted EBITDA.

The opportunity lies in scalability. For years, high SG&A relative to revenue has kept TOI unprofitable and forced shareholder dilution. Today, diluted shares stand at ~120M, including convertibles maturing in 2027.

Yet revenue growth has remained strong, and payors continue to trust TOI. Q4 2025 will mark the company’s first quarter of positive adjusted EBITDA and FCF. From then on, TOI will prove that value-based oncology care works at scale, unlocking significant margin potential.

When TOI enters a new market, capitation margins take time to materialize. During the transition, PMPM fees are received, but services may still be provided by the legacy provider, limiting cost control. Margins improve as TOI gradually takes over care management, so early results understate long-term profitability.

• Pre-inflection: SG&A remains high while new contracts ramp.

Financials

• Post-inflection: SG&A drops to 10–15% of revenue, and margins expand.

Florida alone could add $300M in revenue if TOI achieves penetration levels similar to California. Given the existing trust with payors, future growth should remain robust.

Management has suggested an 8% adjusted EBITDA margin by 2028 is achievable, even while maintaining growth investments. With $800–850M in projected revenue by then, EBITDA could reach $64–68M. At 10–15x multiples, TOI could be worth 2–3x its current valuation.

Comparable transactions support this view:

• June 2023: TPG acquired OneOncology at 19x EBITDA.

• August 2024: FCS sold its Management Service Organization division for ~$7M per practitioner, implying ~$900M valuation for TOI.

• PRVA, a comparable care provider, currently trades at ~15x.

Investment Thesis

The case for TOI is not only about growth and profitability, which seem increasingly likely as scale takes hold. More importantly, TOI represents a new model of cancer treatment that aligns incentives between patients and payors.

While legacy health providers struggle under rising costs and FFS-driven incentives to overprescribe, TOI offers an evidence-based, cost-efficient model that benefits all parties. This positions the company at the forefront of a major shift in U.S. oncology care.

The content of this webpage is not investment advice. It is for general purposes only and does not take into account your individual needs, investment objectives, and specific financial circumstances. Investment involves risk.

Culture is set from the top, so the quality of management is the main contributor to how a company will behave and operate at a group level.

Identifying good or bad quality management and culture from the outside, as an investor, is a notoriously difficult thing. Most of the time, people talk about it in retrospect.

“We just can’t think of another company that exhibits that level of paranoia after such a track record of success or that shows such a willingness to experiment, tweak, and rapidly change. Of course, those experiments often fail, and that willingness to risk failure to change is the point. Truly exceptional people and cultures seem to understand that reinvention is critical in order to continue to thrive.”

Danaher, from the book: Lessons from the titans. What companies in the new economy can learn from the great industrial giants to drive sustainable success.

To discover management quality, investors usually try to build a relationship with management, look at how the compensation is structured, how they have reacted to difficult times… so they can understand their motives, incentives and personalities.

I have identified and used a simple framework to identify three different ways of how a business operates. They are interconnected but offer a glance into the quality of management and what your expectations as an investor can be.

There are three main ways in which a management team operates:

1.       Executing according to a desired outcome.

On September 2022, some activist investors appointed their own board members in Medical Facilities Corporation. The Company went form a high dividend and some low ROIC acquisitions, to setting the following goals:

·       Suspend acquisitions

·       Divest its non-core assets

·       Pursue overhead cost reductions

·       Evaluate and implement strategies to return capital to shareholders

The business was valued at around 5x adj.FCF. That was a sensible valuation given they were investing their capital in value destroying opportunities. A change of objectives and the consequent implementation of those has generated a revaluation of the business. Quality has been increased.

2.       The second one will be executing according to established processes. Objectives are still set, but they will be achieved focusing on the process to get there.

George Koenigsaecker was a general manager of the Jake Brake operation, a business part of the Danaher portfolio. When he met two of the architects of the Toyota Production System he was so impressed that he implemented similar process-oriented measures in the company. The Danaher Business System was born.

“DBS defines the culture at Danaher. Its excellence is best illustrated through the company’s ability to maximize cash flow, profit margins, and the customer experience. DBS creates process where it’s needed, freeing up time for the most value-added functions. It also provides the company with differentiation in its M&A and portfolio decisions. Once people are brought in, Danaher teaches them how to be more productive, more focused, and more customer-centric. DBS pushes the organization out of its comfort zone. Employees are empowered to achieve autonomously, equipped with mandatory tools. Lean manufacturing becomes a religion to follow. Wasteful processes are tossed aside. Expectations rise. Those who embrace this rigorous culture thrive; others are filtered out. The metrics are clear, with no place to hide.”

Danaher, from the book: Lessons from the titans. What companies in the new economy can learn from the great industrial giants to drive sustainable success.

A more to-date example is some of the Nordic serial acquirers. Their sobriety and ability to stay consistent with their playbook is driving considerable shareholder returns. They acquire companies at cheap prices, integrate (driving cost synergies) and never overleverage.

An overlooked one is Green Landscaping group:

Another example is Bel Fuse Inc. When Farouq Tuweiq joined the company in January 2021. He went from being a sell side analyst to the CFO of a struggling electrical components business, with a big pay cut and big challenges ahead. The reason? He clearly saw multiple ways of optimizing the current state of the company. He increased EBITDA margins from 9% to 24%, forming today’s Bel Fuse Inc $BELFB. A change in the way the company operates produced significant shareholder returns.

Bill Bennett spent 3 years at Walmart and later Kroger where he served as Vice President / Head of E-Commerce. He quit his job to join, at the time, a $13m market cap company. When you implement top-level quality processes into a small and unknown microcap with potential, results speak for themselves.

See an article about the idea written by @treasurehunting

3.       The third one is to execute based on identity.

This level of operation is the true essence of quality. Some founder-led businesses and family-owned companies can be the best examples of companies operating based on identity.

You have probably heard of the Founders podcast by David Senra. He reads biographies and talks about the life of some of the most famous founders. Obsessiveness and goal-oriented are common traits among them, but maybe my biggest insight is how their identities are tied to the companies they found and run. This has a powerful effect both on their life and on their company’s life.

I believe some great investors owe much of their returns to being able to identify the correct business identities.

When a founder/CEO with long tenure sells the company but decides to stay as head of it, the outcome is a millionaire set for life that just loves his work and wants to keep working. This has been in Berkshire’s playbook for a very long time and contributes to Buffet’s style of running a decentralized organization.

Looking for founder-led and family businesses where the correct identity has been implemented can be very profitable. It is also tricky.

The current on-going changes in Japan might be a strong generator of shareholders return after some activist shareholders are pushing for change. Traditionally, very conservative companies were run with cash rich balance sheets and ignoring shareholders returns. The difficult thing with identity led businesses is that a change implies a collective variation of frameworks at an individual level that the people running them have. It is a change of paradigm.

Probably the business I like most in this category is Rolex. You can listen to a podcast about it in the Acquirers podcast. Businesses like Apple, Amazon or Costco have a strong identity that comes from a single founder. Diverging from that identity is a true risk. We might be seeing that with Apple now.

The last opportunity I have found and wrote about in this regard is Moury Construct. A 4^th generation construction company with deep relationships with the public sector that makes them able to run a negative working capital operation. You can identify quality from the way they have managed leverage over the centuries, to the way they present their financial statements. They refer to the way they want to see the future as “avec sérénité”. You can guarantee a bigger than normal cash position, but their focus on quality outputs is slowly and consistently increasing their market share. Plus, a company surviving for four generations  is a statement by itself.