It was rather bittersweet being at the Out East Summit by the TIE — don’t get me wrong, it was a fantastic retreat. Could not have asked for a higher-signal crowd. It’s hard not to be bullish when you see the most established financial institutions on Wall Street proclaiming that by 2023, all of Wall Street is coming onchain and trading 24/7 on crypto rails.

Then what’s the problem? The problem is — our industry seems to have resigned itself to the idea that crypto is just Finance 2.0. To say blockchain is only for finance is akin to saying AI is only for developers. The pendulum of capital has swung too far toward “market integration” companies with the arrival of Wall Street.

Let’s start with the framing: there have been two camps of startups in crypto:

> Market creation

> Market integration

Market creation startups are standalone companies that ultimately have an “internal locus of control” over their product and GTM. Examples are the initial batch of DeFi projects and L1s before they became the dirty words they are today from oversaturation, overvaluation, and hacks.

They do not rely on incumbents for product design, GTM, or distribution, nor do they need an institutional BD team wearing suits to translate what they are doing to the procurement or corporate strategy teams of the “Goliaths.”

Market integration startups build with more of an “external locus of control.” Their product roadmap and GTM are influenced by and adapt to existing incumbents, often with an eventual acquisition as the endgame. Think identity, middleware, and infrastructure companies wiring themselves into BlackRock, Stripe, or Mastercard’s rails.

In their purest forms, the underpinning beliefs of the two are:

Market creation: crypto will create its own market.

Market integration: crypto will be subsumed into the existing market.

2020–2022 was when the market-creation imagination dominated — so too did the conviction that crypto was building a new world of economic activity run on a transparent & verifiable engine onchain, where no one was gated by access due to their existing socioeconomic class. A lot of it — undifferentiated L1s, the romanticization of technical esotericism, mercenary liquidity, vaporware, and unsustainable token economics — failed & deserved to. But at least builders in that era dared to think outside the “finance” box.

Then the tables turned. The arrival of Wall Street in digital assets was something too good to be true to have fathomed back then, but came with a Trojan horse that changed the DNA of crypto. Gradual then all of a sudden, “crypto is Wall Street 2.0” became the new consensus.

Is there something wrong with that? Yes — because it constricted the imagination of what the technology really is. I continue to believe blockchain is one of the most important underpinning technologies in a world demanding unprecedented techno-societal velocity, complexity, and integration.

At Portal, Evan and I went to great effort to distill exactly what this technology is good for: the FEIT framework.

1/ Financialization: Wall Street 2.0, tokenization. No explanation needed here.

2/ Efficiency: It comes in two parts:

• Capital efficiency because blockchain’s composability allows builders to borrow liquidity by standing on the shoulders of giants vs. bootstrapping their own.

• Operational efficiency because smart contracts disintermediate middlemen.

While this may sound intangible, it is a direct bottom-line optimization lever for institutions. Wall Street’s newfound obsession with tokenizing equities, stablecoins/CBDCs, and agentic payments hinges entirely on blockchain’s ability to drastically increase the efficiency of financial plumbing through rule-based self-execution and onchain verification.

3/ Incentives: Blockchain can coordinate new networks & incentivize productive actions. Yes, it’s true that DePIN hasn’t worked so far, largely because of a lack of best practices to:

• systematically drive demand and sales motion, something most “crypto-native” teams lack the knowledge and experience to do

• provide institutional-grade service quality / SLAs

• sustain a token price floor through onchain value accrual, where regulatory clarity would also play a role

But no one can deny blockchain’s magic in bootstrapping network supply through incentives. Filecoin amassed 10 exbibytes of storage in under a year through token incentives — something that took AWS more than a decade. Yes, you can argue: what good does supply do if there’s no demand? But you have to start somewhere, and if supply can be solved by blockchain, you have one fewer problem to tackle.

4/ Trust: The entire verification & privacy tech stack: ZK, TEE, FHE.

I’ve always found the framing that “crypto is about trust” too abstract and philosophical to be persuasive in a capitalist sense — but I’ve increasingly felt the commercial impact of trust firsthand:

• In M&A, one of the most critical sources of overhead — and the honeypot where bankers / consultants / auditors make their money — is finding proof that the acquirer can trust the acquiree’s claims about the health of their business.

• In public markets, listed companies spend an average of $2.4M a year auditing their 10-Ks just to verify their reported numbers.

• In enterprises, different departments spend $$$ reconciling data because they lack a verifiable, trusted shared ledger. Per Anaconda’s 2020 survey, data scientists spend nearly half their time just getting data across sources clean enough to trust and use.

• In retail, no trust = no business. Exhibit A: the widespread adoption of Trustpilot, Google Reviews, and Yelp.

• In service industries, customers need to know they can trust vendors’ judgment and ethics to charge fairly before making a deal.

• In crypto, HYPE trades at a 12x premium on FDV/Revenue over PUMP, even though both are profitable businesses directing revenue toward token value accrual. Why? People simply trust the HYPE team more than the PUMP team to 1) keep making money and 2) keep accruing it to the token.

Trust is expensive to outsource — what if it came built in?

Now map FEIT back to market creation vs. market integration.

“FE” (Financialization & Efficiency) is naturally dominated by “market integration” startups because you tokenize someone’s equity; you disintermediate someone’s existing rails.

But the ocean is still blue in “IT” (Incentives & Trust) for “market creation” opportunities. The value proposition is there regardless of whether an incumbent adopts it or not.

Perhaps what we really need is to stop calling our industry “crypto” — to finally shed the FTX and vapor-token baggage that name still drags around. “Crypto” is ONE business model enabled by blockchain, but we should not let it define the limits of the technology.

The risk to our industry isn’t rejection by Wall Street. It’s being defined by it.

Over the weekend, put pen to paper to make my “gut reaction” more legible — now, with facts, data, and napkin math, I stand by it.

Disclaimer: I’m not a macro analyst. I run a crypto fund. However, we live in an environment where situational awareness of the macro setup has material and quantifiable impacts on our career outlook and personal finances. Had to pay attention, so here we go.

Let’s first look at this year’s pre-IPO pipeline and how much these companies plan to raise—see the appendix for a list of pre-IPO companies by region and their respective fundraising targets.

Of course, not every planned IPO will go through, but just three—SpaceX (rumored $1.5T), OpenAI ($1T), and Anthropic (~$0.5-1T)—on top of the pipeline already amount to more than $3 Trillion.

Now let’s look at the supply and demand of liquidity in the public market in the IPO window.

Demand Side: Liquidity Drain in the Next 12-18 Months

1st Bucket: Primary Raise (Day One)

~$300B to be raised from the public market on day one (midpoint of ~$235–$370B from the summary table), based on an assumed 6–8% float. Notably, this already deviates materially from the modern tech IPO norm of 10–15% (see appendix).

2nd Bucket: Insider Selling (6–12 months post cliff)

The math points to - conservatively, ~30% of total supply as sell pressure post-cliff, driven by early VCs selling for DPI, pre-IPO participants (e.g., family offices) selling into the IPO pop, and the team (assuming founders do not sell) selling for liquidity. The calculations and assumptions are outlined below:

For any IPO, depending on day-one float, 85%+ of the stock supply—i.e., “restricted stock”—typically unlocks within the first 6-12 months. A typical emission schedule is below.

• T+0 (IPO Day): 10-15% float released (the “primary offering”)

• T+90 (Quiet period): Underwriter analysts initiate coverage; no new supply

• T+180 (Lockup expiry): 60-80% of total shares become tradable — THE CLIFF

• T+12 months: Secondary offerings become typical

• T+24+ months: Employee RSU vesting cliffs continue periodically

So who’s sitting on that restricted stock — and how motivated are they to sell? This will give us a good sense of the sell pressure / liquidity drain from the public market in the next 6-12 months.

1/ ~55% Early VCs

per Carta, VCs hold on average ~55% of the company at IPO. If you’re a VC in a DPI drought for the past five years, “to sell or not to sell” isn’t exactly a question. Let’s be loose here and assume VCs, once they hit the cliff, offload 50% of their ownership — roughly ~30% of total supply as sell pressure post cliff.

TLDR: Assume 50% sell-through within 6-12 months = ~30% of total sell pressure.

2/ Family offices and HNIs in the pre-IPO mania for a “quick 2x in a year”

Expect decisive selling actions from these pre-IPO subscriptions. Family offices account for ~31% of total startup capital across all stages, mostly concentrated in late-stage rounds. This translates into 2-5% of total supply at IPO. Unlike strategic or sovereign capital, this cohort typically enters for the IPO pop, has no mandate to hold, and faces the least friction to sell.

TLDR: Assume 70–80% sell-through within 6 months post cliff = ~1.5–4% of total sell pressure.

3/ Institutional participants could be longer-term holders

Sovereign wealth funds (GIC, Mubadala, PIF, Temasek), crossover mutual funds (Fidelity, T. Rowe Price, Wellington), and strategic corporates (Microsoft, Nvidia, Google) collectively hold an estimated 40–50% of late-stage pre-IPO ownership. Assuming pre-IPO rounds dilute 10–15%, this translates to roughly 5–8% of the total cap table. For simplicity, assume these institutional holders do NOT sell in the first 1 months.

TLDR: Assume institutional subscribers are long-term holders & not selling in the first 18 months.

4/ Team Selling

Per Blossom Street Ventures, founders and employees combined hold ~30–35% at IPO—roughly 15–20% founders and ~15% employee/option pool.

Let’s be optimistic and assume founders do not sell in the first 12 months - but employees do. Data shows a consistent pattern: typically 30–50% sell-through within the first 90 days post-cliff, as employees have waited years for liquidity while holding concentrated single-stock risk tied to their human capital.

Case in point: Snap employees sold ~40% of vested holdings in the first 6 months post-lockup; Uber ~50%; Robinhood ~60% (down markets tend to accelerate selling).

TLDR: assume ~20% sell-through for the 10–15% employee portion at cliff = ~2–3% of total sell pressure.

In summary, we are looking at roughly ~32–37% of total supply as potential sell pressure — or ~$1.3–1.5T in liquidity demand — within 12-18 months of IPO from this second bucket of restricted stock unlocks. This is driven by ~30% from VCs (assuming 50% sell-through for DPI), ~1.5–4% from family offices (IPO-pop driven selling), and ~2–3% from employees (years of pent-up liquidity demand), with founders and institutional holders assumed to remain largely intact.

3rd Bucket: Second-order Liquidity Demand

There are two additional, structurally hidden liquidity vacuums that are less obvious but extend over a much longer horizon, continuing to absorb liquidity from the market.

1/ LP Recycling: When VCs distribute IPO proceeds, LPs receive cash back. Most of that is already captured in the insider / VC sell pressure above, but the story doesn’t end there. Successful IPOs validate track records and drive fund size expansion — this is how Sequoia XV at $5B becomes Sequoia XVI at $10B.

LPs that want to maintain their target venture exposure now need to recommit more capital than they received, and that gap has to come from somewhere. In practice, it is funded by selling other liquid holdings. ~$90B was raised by VCs from LPs in 2025 — assume that post-IPO wave, LPs source another ~$100B (to be very conservative) to redeploy into venture over the next 12 months.

2/ Ongoing liquidity vacuum: The drain doesn’t stop at the primary raise or the unlock. Once these names list, the rest of the market still needs to absorb them across index funds, ETFs, benchmark-aware mutual funds, and active portfolios that use indexes as building blocks. These mega IPOs are not a one-time event. Index inclusion, float updates, secondaries, and benchmark reweights create a staggered sequence of forced buying in the new names over multiple quarters — well beyond the initial 12–18 month window.

Summary View

We are looking at roughly ~$1.4–1.6T of new liquidity demand within 12-18 months of these IPOs in the pipeline.

The State of Liquidity Supply

Start with the obvious but misleading number: total U.S. equity market cap sits around ~$69T. That tells you how much is the aggregated market cap of all stocks in the US, but it’s not particularly useful — almost none of it actually moves on any given day.

Flow - or daily trading volume of U.S. equities sits at roughly ~$1.1T of notional daily (Cboe). This is more indicative of the real absorption capacity — the pool of capital that can actually respond to supply on a day-to-day basis.

Against that, the top IPOs are looking at ~$300B of primary raise on day one (assuming ~8% float — and in some cases like SpaceX, closer to ~2%), followed by another ~$1.4–1.6T over the next 12–18 months. Not trivial — but on the surface, it doesn’t look extreme either. So how should we think about it?

Some historical context helps.

At the peak of the 2000 dot-com wave, ~$97B of IPOs hit a market doing ~$130–150B of daily notional — roughly ~0.7 days to absorb. That was the top. The Nasdaq peaked in March 2000 and went on to lose 78% over the next 30 months.

In 2021, the SPAC + IPO wave raised ~$317B (~$155B IPOs +$162B SPACs) against ~$550B of daily notional — about ~0.6 days to absorb. The Renaissance IPO ETF peaked in February 2021 and dropped 57% over the following ~16 months.

Now fast forward to 2026: ~$300B of primary raise against ~$1.1T of daily notional — ~0.27 days to absorb (per Cboe). On the primary-raise metric, 2026 looks the most absorbable of the three, but...

1. Float compression (<10% in 2026 vs. 15–25% in 2000) means cliff supply is magnitudes larger than in 2000 and 2021: ~$1.4–1.6T of potential sell pressure within the next 12 months per the math above. The right comparison is not days-to-absorb-the-raise — it’s days-to-absorb-the-cliff. And that’s where 2026 starts to look shaky.

2. 53%/96% AI concentration means the unwind would be highly correlated and domino-like. By valuation, ~53% is pure AI — OpenAI, Anthropic, Databricks, ByteDance, MiniMax, Zhipu, Pony AI, Hesai, Cerebras, and Anduril total $2.25T. Adding AI-adjacent names (SpaceX/xAI, Stripe, Canva, Jio) brings AI exposure to ~$4.1T, or 96% of the entire pipeline. By raise amount, the number is ~71% is pure AI ($197B) and ~90% is AI-adjacent ($249B).

Stack that on top of Nvidia’s existing ~$5T market cap, and AI-exposed names end up representing ~17-22% of total U.S. equity market cap by end of 2026 — comparable to financials’ peak weighting in the S&P 500 pre-Global Financial Crisis (~22% in late 2006).

Then add in the capex spend: the top five hyperscalers (Microsoft, Meta, Google, Amazon, Oracle) are guiding to $660-700B in 2026 capex, up ~70% YoY. Capex-to-revenue ratios have hit historically unthinkable high — Oracle 86%, Meta 54%, Microsoft 47%, Alphabet 46% - really hoping that ROI math starts working soon.

AI is great — but is it just a tad too great?

Other Looming Factors Not Talked About Enough

I have no idea how or when this plays out, but the setup doesn’t look too pretty — not to mention the several confounding variables below:

1. The Domino Effect of the Sequencing Risk

Every major IPO is targeting the same window between Q2 2026 and Q1 2027, with all eyes on SpaceX first.

If it does well, great — and that won’t surprise anyone. But if it doesn’t, the entire pipeline gets repriced down. PitchBook’s Q1 2026 analyst note flags this explicitly: “These mega-offerings absorb available capital and investor attention, squeezing out other companies and pushing liquidity timelines into 2027 or beyond.”

SpaceX turns out to be a pretty high-stake bellwether.

2. ~70% U.S. Listing Concentration

~70% of the ~$370B global pre-IPO raise is expected to land on U.S. exchanges. Last time I checked…

• U.S. fiscal deficit also needs liquidity — running ~6–7% of GDP; Treasury needs to issue ~$2T+ in net new debt over the next 12 months, going after the same liquidity pool.

• Still a war backdrop — U.S.–Iran talks stalled; Pentagon shakeups signal more turbulence ahead.

• Oil supply chain disruption hasn’t fully fed through — crude still ~$94.

• Private credit still shaky — “software-mageddon” already wiped ~$800B from SaaS; default rates rising in shadow banking, concentrated precisely where the IPO pipeline sits.

• Fed not cutting and inflation not fully contained — CPI choppy; oil shock and tariff effects feeding through.

• Consumer sentiment at all-time lows — UMich 49.8, lowest on record, broad-based across parties, income groups, age cohorts, and education levels.

• VIX signaling fragility — MS flagged “complacency masking underlying macro fragility”; 2018 analog saw VIX spike to 50 within weeks

• Treasury market depth still below pre-tariff-shock levels — the foundational liquidity layer for all risk asset pricing is structurally weaker than a year ago.

Ya you get the picture.

3. Recent IPO Underperformance

Every major 2025 IPO has traded below its private round valuation post-listing. The 2026 mania is being priced as if this pattern doesn’t exist.

4. Historic Under-Float

1990s tech IPOs floated 25–35%; modern mega-IPOs float 8–15%. The 2026 cohort is coming in at 2–5% — historic lows. Low float = artificial scarcity = guaranteed pop = “successful IPO” headlines = retail FOMO into the stock — then the cliff hits and 95% of the company becomes tradable in a short window.

We all know what happens after.

The POINT

All in all, the music hasn’t stopped and will probably keep playing for a while. And yes, it’s a global liquidity supercycle for strange reasons.

But can someone answer this one question: WHERE IS THIS LIQUIDITY COMING FROM to support and sustain $4T+ of market creation and $1.5T+ of net liquidity demand, all within the next 12–18 months?

I can’t tell you where we’re going — but I know where I’m not.

Appendix: Stats for Sanity

AI & Tech IPOs: $174–245B estimated raise

Crypto IPOs: $8.65–14B estimated raise

China/HK Tech: $30–66B Estimated raise

Pure AI IPO Pipeline

AI-adjacent IPO Pipeline

Float at IPO to-date

First off — I’m not denying HYPE is 100% one of the most impressive projects out there. None of the below is a comment on HYPE’s fundamentals as a business. It’s only an objective assessment of its price trajectory.

And here it goes: HYPE is NOT undervalued. I don’t see it doing more than 2x from here ($40 today). There may be temporary wicks above 2x when animal spirits return — but within one standard deviation, price should plateau under $80 over the long haul. And that’s being generous. Here’s why:

1. 3x sell pressure from its 75% unvested token supply: HYPE today is trading at $40B FDV with only 25% float with vesting until 2028.

   Let’s look at comps for a sanity check: ATH of

   1. CME = $118B, ~3× HYPE’s FDV today

   2. ICE = $107B, ~2.5×

   3. Nasdaq = $57B, ~1.5×

   For a HYPE buyer today to get a 2×, they need to assume:

   1. Either the HYPE team never sells, or there’s 6x marginal buying from today to absorb the 3x net-new sell pressure from future team vesting — with the net price effect landing somewhere in between.

   2. HYPE eclipses Nasdaq and closes in on CME’s ATH — the world’s largest derivatives marketplace, processing over $1 quadrillion in notional volume annually, with a 130-year regulatory moat.

2. So who are the buyers? The two sides are retail vs. institutional.

   1. Retail: What’s the net-new “secret” or “alpha” that triggers the buy when literally everyone and their grandma knows HYPE is the posterchild of crypto? HYPE is no one’s secret. There may be some buyers on the sidelines — but how many?

   2. RE institutional — this one could move the needle. Two types of institutional buyers here:

      1. Wall street hedge funds or mutual funds

      2. Crypto hedge funds

   For (i), I highly doubt web2 hedge funds would touch HYPE — because it’s illicit. The KYC-less setup and regulatory arbitrage are features for HYPE’s blitzscaling, but are endowments and mutual funds really going to get comfortable underwriting that? You could argue PURR is the way in — but that’s up to their risk tolerance and I can’t comment there.

   Even if web2 mutual funds and endowments do buy, what do you think these Wall Street veterans would comp HYPE to? Nasdaq (1.5x HYPE at its ATH) and — by a long stretch — CME (less than 3x HYPE at its ATH).

   For (ii), ask any sensible hedge fund managers on CT — when HYPE gets to $80B, will they continue to hold? Even if they want to, they’d need to sell some to meet fiduciary duty for their LPs. Expecting HYPE, an illicit perp exchange, to be more valuable than CME — with 130 years of regulatory moat and lindy effect — would make them look delusional to LPs.

   So tell me: where are these marginal buyers coming from? And we need A LOT of them.

3. Escalating hack risk: between Lazarus Group and Mythos-equivalent AI hack capabilities on the horizon — if someone were targeting crypto, where do you think the biggest honeypot is?

4. Not to mention the keyman risk: Hyperliquid’s is as bad as it gets in crypto. Jeff Yan leads an 11-person team with a product-first, no-VC philosophy, and every part of the infrastructure — matching engine, validator code, execution logic — was built internally. Can anyone point to documentation of backend knowledge transfer if something were to go wrong (knock on wood)?

   Yan is the sole strategic and technical anchor of a protocol processing $10B+/day. No named co-founders, no visible succession plan, no governance oversight. I genuinely hope Jeff is spending a good chunk of that profit on security and protection

5. Crypto traders are mercenary by nature, and blockchain interop is definitely a bug, not a feature, for projects trying to retain users. At some point, traders will decide HYPE’s upside is growing too slowly and rotate their liquidity into the next 10x. Network effects do not hold up in crypto — I’ve written about this extensively.

Net, if I’m a retail investor looking to:

Buying crypto for a multi-year hold — why would I choose HYPE with a ~3x net supply overhang from the 76% yet to unlock, over BTC at $70K: the only digital gold, 100% vested, no central control, completely irreplaceable? HYPE needs to find 3x net-new buyers just to maintain its current price.

Buying crypto for a 10x — why would I pick a token already at $40B FDV, as consensus as it gets, with competing DEXs multiplying and every KOL shilling their bags — over a lindy memecoin that’s 90%+ down from its peak, memetic, and volatile in all the right way?

And for those of you comparing HYPE to BNB: it’s a dubious comp because

1. BNB’s 100% vested so no net-new sell pressure - unlike HYPE at 24%.

2. Binance’s willing to do ALOT more to “kingmake” BNB vs. HYPE’s laissez-faire

3. the most important Q to answer for price action remains “who are the marginal buyers” - for rational institutional buyers, the comp will be CEM/Nasdaq and not BNB

Go ask the KOL shillers if they'd buy HYPE at this price. It made a lot of them rich — so ask yourself: who's their exit liquidity?

Crypto venture is at a watershed moment. Over the past 3 cycles, token exits have been the primary driver of outsized returns but are now undergoing a meaningful reset. The definition of what makes a token valuable is being rewritten in real time, yet an industry-standard evaluation framework has not emerged.

So what exactly happened?

Crypto’s market structure was disrupted by a confluence of forces unseen in any of the prior cycles:

1/ HYPE came out of left field and jolted the token market awake by proving that token prices can have real revenue support, with 97%+ of its nine- to ten-figure revenue accrued onchain. This triggered a sharp disillusionment with narrative-driven but fundamentally weak governance tokens—think the L1s and “governance tokens” that primarily existed to skirt security law ambiguity back then that made direct revenue sharing infeasible. HYPE reset market expectations almost overnight: revenue are now scrutinized far more aggressively and have become table stakes.

2/ Knock-on whiplash to other token projects: prior to 2025, if you have onchain revenue, you are a security; post-HYPE, if you ask most hedge funds, if you don’t have onchain revenue, you are going to 0. This put most projects, especially the non-defi ones, between a rock and a hard place scrambling to adapt

3/ PUMP delivered a spectacular supply shock to the system. The explosion of supply from memecoin mania fundamentally disrupted market structure by fragmenting attention and liquidity. On Solana alone, the number of new tokens generated ballooned from roughly 2–4k per year to 40–50k at peak. This effectively divided the liquidity pie, which did not grow much, by roughly 20x. Attention and capital from the same buyer base seeking outsized upside rotated toward flipping memecoins rather than holding altcoins.

4/ Alternative outlets for retail risk capital proliferated. Prediction markets, equity perps, and leveraged ETF trades now compete directly for the same pool of retail risk capital that would previously have gone to alts. In parallel, the maturation of tokenization technology makes it possible to lever up blue-chip stocks that do not carry the same risk of going to zero as most alts and are far more regulated and transparent, with lower information-disadvantage risk.

The result is the compression of the token lifecycle: the lead time from peak to trough has shrunk dramatically, and retail appetite to “hold” tokens has dropped precipitously, replaced by faster rotation.

Some of the biggest questions every VC is asking themselves and their peers:

1/ Are we underwriting equity, tokens, or some combination of the two? The biggest challenge here is that we have no new playbook of best practice in value accrual for token projects—even some of the most successful projects like Aave still face controversy across DAO and equity.

2/ What is best practice for onchain value accrual? Most common is token buyback but doesn’t mean it’s right. We’ve been long against the prevailing token buyback trend: it is toxic & puts founders with real revenue between a rock and a hard place.

• The motivation is completely wrong: stock buyback happens when they are done investing in growth vs. crypto buyback is increasingly forced to be immediate by retail/public perception—something completely mercurial & irrational. You could be literally burning 10M of what could have been reinvested and the next day it gets wiped because some random MM gets liquidated.

• Public companies buy back stocks when they are undervalued. Token buybacks get front-run left and right and so often executed at local peaks.

• You are spitting into the wind especially if you are a B2B biz generating offchain revenue. Imo there is 0 reason to buy back when you are doin less than 20M in rev just to please retail instead of reinvesting in growth. I liked this report/screenshot by fourpillars showing how 10-fig buyback barely made a dent to help projects set a long term price floor.

  

• On top of that, to make retail and hedge funds happy, you also gotta buy back consistently and transparently like HYPE. Anything short of that, you get punished like PUMP trading at 6x P/F (fully diluted) because the public “doesn’t trust” them—despite the fact they’ve literally burned $1.4B in revenue that could have gone to the treasury. Here’s further read of what COULD work as a onchain value-accrual mechanism without burning money:

3/ Will “crypto premium” completely go away? What this means is that going forward, everything will be valued based on a similar multiple (something between 2–30x revenue) as public equity. Take a second to sit on what this means—if true, we will see a 95%+ further price drop for most L1s from here, with exceptions like TRON, HYPE, and other revenue-generating DeFi projects. This is before even taking vesting into consideration.

https://tokenterminal.com/explorer/markets/blockchains-l1

Personally, I do not believe this would be the case—HYPE has set an outlier expectation that made many investor impatient about “day one revenue/user traction” for early-stage companies. For sustaining innovations like payments and DeFi companies—yes, that’s a fair expectation. But disruptive innovations take time to build, ship, grow, and then hit the rev hockey stick. We went from having way too much patience and hopium for “disruptive technology” through 8–9 funding rounds for new L1s, Flashbots/MEV esotericism over the past 2 cycles, to now an overcorrection and only backing DeFi projects.

The pendulum will swing back. While it’s indeed a net positive for the maturity of the industry to value on “quantitative” fundamentals for DeFi projects, but for non-DeFi categories, “qualitative” fundamentals also need to be accounted for : culture, technological innovation, disruptive concepts, security, decentralization, brand equity, and industry connectivity. And these qualities will not reflect purely on TVL and onchain buy-back.

So... what now?

Return expectations for token projects have meaningfully compressed, while equity businesses have not seen the same degree of dampening. This nuance is playing out differently across early-stage versus growth.

Early stage investor have become far more price-sensitive when underwriting projects that are likely to exit via a token. Concurrently, the appetite for equity businesses has grown, especially given the favorable M&A environment. This is a complete departure from 2022–2024, when token exits were the preferred liquidity path, underpinned by the assumption that token valuation premiums would persist.

Late stage investor whose brand equity and value-add are strongest within crypto-native contexts, are increasingly drifting away from purely “crypto-native” deals. Instead, they are backing more “web2.5” companies where underwriting is anchored to revenue traction. This puts them in uncharted territory, competing directly with crossover and Web2 fintech funds—firms like Ribbit or Founders Fund—that have deeper context in traditional fintech, stronger portfolio synergies, and better visibility into early-stage deal flow outside of crypto.

The crypto VC landscape is heading into an attribution period. Right to survival hinges on VCs finding their own “product-market fit” among founders, with the product being a combination of capital, brand identity, and value-add.

For the best deals, VCs need to sell themselves to founders to win the right to be on the cap table, especially given that some of the most successful outcomes in recent years required very little institutional capital (e.g., Axiom) or none at all (e.g., HYPE). If capital is the only thing a VC offers, it will almost certainly get shed.

The VCs that deserve a right to stay in the game need to be crystal clear about what they have to offer in terms of brand identity—which drives why the best founders would engage in the first place—and value-add, which ultimately determines their right to win the deal.

Further reading on

• My mental model for “value-add”

• Why pre-seed investing is a different sport

VCs pattern match for a living, but the diversity of what makes someone a killer founder truly defies all existing stereotypes and frameworks — especially when there isn’t much data to go off: young, first-time founders, outside the network, with no blue-chip pedigree or track record in a new sector. No, MBTI won’t help here — that’s a personality tool, useful for assessing a friend, coworker, or life partner.

I don’t care about compatibility. My questions are more binary: is this founder a killer or not? And how “raw” vs. “polished” is this gem?

Reading the “On-paper”

Founder Success = Obsession + Ambition + Aptitude

Obsession (the #1 quality I look for)

As I collect more data points on founders since the start of my career in venture, founder obsession has become the landslide leading indicator of how far they can take a project. Obsession can't be faked: it’s the unique insight no one else saw; the stamina to climb an uphill battle with no one cheerleading; and the grit to keep building when momentum-chasing VCs call you 'too early.'

Here’s a non-negotiable question I ask myself before making any investment decision: if no one gives this founder another cent, are they obsessed enough to keep building anyway?

Good companies come from opportunists; generational ones come from the obsessed.

Ambition:

Signals for ambition are relatively easy to spot. A classic one is "chip on the shoulder puts chips in the pocket." Another is a record of being somewhat "unemployable / relentlessly entrepreneurial," never quite fitting inside others' systems. A third — and a rather unusual question I ask some founders — is how they would rank money, power, and prestige in importance to them. There's no right answer, but it's always telling.

Aptitude:

Signals for aptitude can show up in many forms. One is “unfair advantage,” even if manufactured. Not every startup can be king-made like Snowflake — incubated by VCs and inheriting the Rolodex — but founders can manufacture their own edge by becoming the expert in their respective field through deep research and obsession (hence the importance discussed above).

Another signal is evidence of "punching above their weight." It's not about the cards someone is dealt, but how they play them. Not every killer is Ivy or ex-FAANG blessed. Sure, getting into a top school likely means someone did well in high school (excluding legacy kids), but that says little about adult life. Many strong founders just weren't lucky enough to be picked up by Ivy admissions, couldn't afford the tuition, or were late bloomers who coasted through high school, went to non-target colleges, and then had to work doubly hard to break into Wall Street. Landing a first job at Goldman from a non-target school is a very green flag to me.

A third signal is evidence of a "hard pivot." Pivoting from accounting to IB is a moonshot. Even within consulting, firms love to pigeonhole people into a service line or industry because retraining is expensive. Breaking out of that mold takes courage, grit, ambition, and curiosity — all of which translate directly to founder quality.

Reading the “Off-Paper”

Reading what’s “off paper” is the real craft. What if founders weren’t “trying” at the moment you met them? How do you tell someone is impressive when you meet them in non-work mode — when they aren’t putting on a show because they only care about getting a16z on the cap table? When they’re quiet “sigmas” sitting in a corner with zero urge to tell you what they know? When they’re introverts, or intentionally keeping things on the DL?

A few tells I’ve come to notice when observing founders in an open social setting:

• Killer founders are not earnest to please: they aren't approval-seekers in conversation, nor are they typically the smoothest talkers. They aren’t eager contributors to every conversation, but when they do speak, it’s substantive and thoughtful.

• Killer founders don’t pitch — they tell: here’s the reality you don’t see, and I’m showing it to you. The tone is never salesy, but measured and firm. The best founders take you on a reality distortion into the future. Some of my best investments happened after conversations that felt like a nutritious meal for my mind.

• Killers have a certain intensity and stillness: there’s a low hum of energy that doesn’t need an outlet. They aren’t fidgety, they aren’t performing. It’s the stillness of someone with a quiet self-knowing that they’re among the most interesting people in the room, whether or not anyone else has figured that out yet.

But here’s the catch: that stillness is easy to overlook — especially when someone louder, smoother, and more eager to charm you is standing right next to them.

That’s most confounding variable of all — the trap of charisma.

The Trap of Founder Charisma

An articulate and charismatic founder like Sam Altman or Jensen Huang is every VC’s dream. To the left of the founder charm spectrum you have Elon Musk and Alex Karp; too far right and you’re veering into Adam Neumann and Elizabeth Holmes territory.

Interestingly, the “curve of diminishing returns on founder charm” follows the same trajectory as the Yerkes-Dodson curve — which states that performance increases with stimulation, but only up to a point. After that peak, more charisma, like more stimulation or stress, drives outcomes off a cliff.

This is where most false positives and false negatives happen.

• False positive: founders with dazzling charm and eloquence that masks thin aptitude and a shakier execution track record.

• False negative: introverted, unassuming, sometimes “awkward” founders who struggle to compress their thinking into a 30–60 minute pitch — not because they’re poor communicators (which is definitely a dealbreaker), but because they’re deep in a complex, multi-layered mind palace and need help zooming out to the big picture.

  In fact, this is consistently the most sought-after help our portfolio founders ask us for. They often have the right materials — they just need a trusted thought partner and sounding board to help package them into a digestible format for fundraise or user audience.

Being a great VC — especially at the earliest stage — is not easy: constrained data points, extreme variance across founder backgrounds and sectors, multi-year delayed gratification, and the ever-present possibility that the founder you backed pivots in a completely different direction tomorrow.

It is a world of 10% constant, 90% variable. But here’s the secret: that 90% is entirely anchored in people, and human nature hasn’t changed for millennia. Master that, and the rest is just pattern recognition.

Inflection

A change in the status quo — technological, social, or regulatory — that enables a break from the past. The “why now?”

We’ve already seen the transition from centralized content creation — dominated by traditional media studios — to the rise of user-generated content across platforms like TikTok and Instagram.

Layered on top is the AI content explosion. As of late 2025, more than half (52%) of all written content on the internet is AI-generated, versus ~1% in the pre-GPT era. In aggregate, we now have a trifecta of forces driving a parabolic rise in content availability:

• AI/bots generating net-new content

• AI accelerating both amateur and professional creator output

• Existing studios and agencies continuing to scale production

The result: explosive content from studios, individual creators, and AI is flowing down the same outlets — dominated by big-tech incumbents — competing for the single scarcity: our attention.

Insight

An unconventional truth derived from an inflection that translates into opportunity.

The Opportunity to Financialize Attention as a New Asset Class

$3.18 trillion in revenue has been generated over the past decade by big tech productizing our eyeballs. On the micro level, every time I open Twitter, Instagram, or Little Red Book (the top GMV driver in China), everyone is fighting for likes, reposts, and comments in hopes of reaching KOL status with a respectable following.

Among these aspiring KOLs, some are doing it for clout (think ct VC posts), others — and I’d bet the majority — are doing it for the prospect of monetization one day. Case in point: there are now 6,939 influencer marketing-focused service providers in 2025, up from 1,120 in 2019.

Now let’s look at the status quo market structure more closely. Players in the attention economy are

1. Attention Producer: audience viewership

2. Attention Wholesaler: KOLs and social accounts looking to monetize their followership

3. Attention Buyer: individuals or enterprises (brands, any company with a marketing budget) seeking distribution of their creatives, written or video

The existing market structure is painful for everyone.

“Attention Wholesalers” — the KOLs, nano-influencers — still do not have a liquid, “go-to” venue to monetize their distribution capacity on social media. The numbers are stark:

• There are 207 million content creators globally, yet only 4% earn more than $100K/year. Nearly half of all monetizing creators earned less than $500 in 2025.

• 59% of beginner creators have never monetized at all, even though nano-influencers account for 75.9% of Instagram's influencer base.

• The mid-tier is a dead zone — white space. NeoReach’s 2025 Creator Earnings Report identifies an “under-monetized middle tier” of creators too small for agency representation, too large for nano-gifting programs. This segment has no venue, no price discovery, and no liquidity.

• 57% of full-time creators earn below a living wage from platform revenue alone, which is why brand deals account for 68.8% of creator primary income.

This is exactly the problem: the process of making these brand deals is extraordinarily inefficient. The experience for the "buy side" — anyone looking to pay for distribution of their content, video or written — is a high-friction structure akin to an OTC market: pitching via DMs, negotiating rates with no transparent pricing, and waiting weeks for a single partnership to materialize — if it materializes at all. Brands report 10-20% positive response rates on outreach, meaning 80-90% of attempted transactions fail to clear.

What About Existing Attempts to Solve?

Sure, platforms like Whop, Noise, Vyro, and Clipping.net are disintermediating the agency model — replacing DMs with self-serve portals where buyers post bounties and creators race to fulfill them. But these are job boards, not exchanges. A buyer sets a fixed CPM, creators manually opt in, and settlement happens after views are verified. No price discovery — the buyer dictates the rate. No liquidity — creators can only monetize when a campaign matches their niche. No continuous market — supply and demand clear in discrete, asynchronous batches.

Attention is still priced by fiat, not by the market. A true spot exchange would let distributors post ask prices for their distribution capacity, let buyers bid against live inventory, and let supply and demand set the clearing price — the way Google’s search auctions and The Trade Desk’s programmatic display already operate.

The gap isn’t “can I post a clipping bounty online” — it’s “can I get a real-time, market-determined price for 10,000 views of organic distribution on TikTok and execute that trade in seconds?” That’s the white space.

The Holy Grail: a “Monetization-native Social Media”

Long term, this is a bold but recurring vision in my head: what if this platform evolves beyond a “Nasdaq of attention” and horizontally expands to become a “monetization-native social platform” — one that:

1. Aggregates commercializable content of all formats — written and video — across existing platforms (TikTok, Instagram, X, LinkedIn, media outlets, etc.)

2. Directly rewards viewers for their attention — a "watch/read-to-earn" model, so to speak. Crypto is a natural fit here to bootstrap the initial supply of attention (viewership) via a bonding curve. The reward for any individual piece of content will be negligible, but in aggregate it becomes sizable. Imagine TikTok paying viewers for the time they spend on the platform.

3. Becomes a Monetization-native Social Platform that evolves from content aggregator to native home for UGC with embedded economic value. Every post is a monetary unit. Content lives where it monetizes — no handoff to ad networks, no separate sponsorship negotiations, just native economic rails built into the content layer itself.

How Big Does It Get?

To give a sense of the opportunity size: paid social media marketing is now the single largest advertising channel in the world — $247.3 billion in 2025, surpassing paid search for the first time. That's what the market is already willing to pay for attention distribution through today's fragmented, high-friction pipes.

Revenue = (Global Ads Spend) x (% to paid social media marketing) x (platform take rate)

• Global ad spend in 2024: $1 trillion, at 10-15% CAGR

• TAM = ~20%-25% being paid social media marketing, at 11% CAGR

• Assume platform take rate: 20% (AppLovin is 20-35%)

• Revenue opportunity on a 5-year horizon across scenarios

All in all

Meteoric rise in market demand. Inefficient market structure akin to OTC. No price discovery mechanism. The opportunity to financialize attention as an asset class is screaming at me.

And this is why I've been tracking Floodify as a close expression of this.

And in case you’re wondering, I’m quite social (or so I’m told) and haven’t had the urge to take an autism test.

I’ll keep this one brief—and the same for all future posts, because attention really is the one scarcity on the internet. Here are my daily AI hacks, as a typical VC whose life is hectic by default, to solve my “death of a thousand cuts” problem.

My AI Productivity Hack

The obvious use cases aren’t listed here—everyone already knows them.

Gemini:

1. News intake: label all newsletters into a single folder in Gmail. Every morning, schedule a daily summary from my subscriptions

2. Personal CRM: this is not perfect but pretty real-time. Create a CRM table inside Gemini chat, and as you walk out of a meeting, use the audio feature to talk to Gemini and log details.

   Imo there still isn’t a perfect personal CRM for most investors and founders. I’ve tried:

   • Affinity: way too much friction due to having to use their website; their mobile app is unusable; website is way too complex/inaccessible for real-time add/notes

   • Notion: decent mobile UX but still “too much” and too many clicks because of their multi-layer formatting. This is not exactly a bug — being able to support robust container blocks of note-taking necessarily renders “one-click” or speed (like Apple Notes) impossible.

   
   Another big one here is that I’m personally paranoid about Notion because of the portability of data. The isomorphic formatting that contributes to its success by allowing infinite nesting make data export an absolute nightmare, and I need to keep my CRM close to my chest. Gemini, on the other hand, just puts it in a table form that I can export to Google Sheets and copy-paste back any day.

   
   Personal CRM needs to be somewhat between Apple Notes and Notion, and more importantly, allow for real-time intake without changing modality. Gemini is a pretty solid option for now and let’s see how this evolves.

3. Self-pattern analysis: this is more of a periodic exercise. Sometimes I upload my journal for an LLM to observe — well, perhaps NotebookLM is a better container for this, as I’m not so thrilled to have my personal journals show up in their training dataset. Dump the data, start asking Gemini questions about patterns, recurring themes, what I haven’t really thought about, etc.

4. Real-time therapist & “what would _ say”: a rather obvious one. AI can be an on-demand therapist and friend when you need to process something in real time. Something I layer on is prompting “you are [an author or school of thought]” before your question if you want advice in the style of “what would Ray Dalio say in this situation.”

5. Travel itinerary log: ask AI to reconcile your itinerary with hotel, flight, and conference info based on your email confirmations.

6. Self-brainwashing / subconscious hack: I set my personal context to include certain goals I have in life/long term so that the LLM directs all its answers toward such.

GPT: well, pretty much the only thing I do there is to help with writing at this point.

My intuitive prediction on the AI competitive landscape

Won’t pretend to be an AI expert, but sharing my observations as a retail power user.

1. I see Google winning retail due to its indisputable data-integration moat across Gmail (most and foremost), YouTube (also huge in tailoring personalized content), and task automation based on Gmail

2. The majority of “normie” users like myself will choose convenience over privacy. After all, most don’t have much to hide, and we don’t put our bank passwords in Gmail.

3. Contrary to my prior post calling GPT the Western WeChat for its ecosystem play, I’m turning bearish on OpenAI’s retail strength as I—and my circle—use Gemini more and more. There are a couple of sub-threads undergirding this.

   • GPT’s first-mover advantage still can’t beat its competitors’ data moat via Google’s ecosystem lock-in across platforms where historic context and preference data already reside

   • GPT’s business model has been in flux. I’ve also learned to take Altman’s words with more grains of salt: he long argued ads posed a conflict of interest—and now, boom, GPT is testing ads, adding fuel to the fire over its privacy controversy. He’s vacillated between ads and no ads for too long, despite ads being the obvious model for his user base.
     A code-red desperation?

   • That said, GPT’s “AI App Store” play could still work given its head start with GPT “Apps” and the confluence of startups building on GPT wrappers. Conservative estimates put the number of GPT wrapper startups in the high thousands and counting. There’s no reason for OpenAI not to rent-seek and win from there - but unclear if that’s even the strategic focus.

4. What will happen to the OpenRouter of the world? Intuitively, if what’s winning long-term users from GPT to Gemini is integration and data moat, then OpenRouter’s retention is in a dicey place. I can think of two lasting use cases:

   • Developers wanting to try out different models before committing to one

   • Retail users getting too much of a “sycophancy” issue from GPT/Gemini’s contextual knowledge and wanting a clean slate

   Neither are super long-term per se. I see OpenRouter as more of a “on-ramp” - then the question becomes… how big is the TAM?

5. Apple Intelligence doesn’t even get to be “in the game” of AI as it stands today, but long term—especially after its partnership to use Gemini to train Siri—I’m still bullish on the Apple ecosystem because it has three of the most impervious moats of all, at least in Web2:

   1. network effect

   2. vertical integration / ecosystem lock-in

   3. foremost, distribution via hardware (so so hard to disrupt user inertia - for one I’m not thrilled by Iphone upgrades but I’m not switching to Android)

   On the AI battle field, Apple really just has one job going forward: make Apple Intelligence not suck.

Curious to see all your hacks - DM!

A native wechat user myself, I’ve been secreting placing bets on who has the best chance. Contenders have been

Instagram

It tries, but the brand itself is too tainted with the image of a “social media” app.

Let’s first examine WeChat’s value propositions in ranking of importance:

1. A necessity: Wechat is perceived as a neutral and indispensable communication tool. This is the stickiest and most important trait for any attempted superapp.

2. A fintech tool built on its user base: P2P and merchant payments, issuing payment cards, wealth management.

3. A social network: a high-filter, closed-circle social product: you can only see the “Moments” posted by people you know.

4. A services gateway: pay utility bills, shop in-app, buy railway tickets, concert tickets, order Uber, top up phone bills, and even donate to charity

5. A discovery platform for content and services: mini pages and mini merchants in the ecosystem.

6. An entertaining “brainrot”: the integration of TikTok-like reels and conversion to GMV & ads.

Note the entertainment piece comes last. No one associates WeChat with a mindrot or the corollary aversion to it. WeChat is neutral and universal, yet Instagram is often associated with being “girly,” “vain,” and “extractive of user attention,” driven by ads and “evil algo” designed to hook users and create negative externalities. From the get-go, Instagram was never going to make it as the “universal must-have” because it’s a mere “nice-to-have.”

WhatsApp

It remains shocking how bad its UX is for an app so entrenched in people’s daily life, speaking volumes about how powerful the network effects of a walled-garden incumbent really are. Still, does WhatsApp even have a UX designer or a growth team? It should have been the most obvious thing for them to integrate CRM or note-taking features into the “Contacts” screen, and it’s still completely absent. Innovation has stalled for years.

Details aside, what a waste of their first-mover advantage as the de facto chat app of the Western world.

Telegram

I’ve been immensely impressed by Telegram’s execution. The UX is sleek and continuously improving with the introductions of mini apps, merchant pages, and services, wallet integrations, and Instagram-like Stories. Its trajectory comes as close as it gets to becoming the first WeChat of the West—until I noticed the true ambition of OpenAI.

The GPT Takeover

I did not see this coming. All of my comps were somewhat chat-app related, but the age of AI changed things. When WeChat got started, instant and smooth communication was the “new trend” when everyone was still being charged per text-basis by telco providers.

Inflection vs. Insight

As articulated in one of my favorite venture books, Pattern Breaker by , a breakout startup needs two things:

1. An inflection (an event that creates the potential for radical change in how people think, feel, and act), and

2. An insight (a nonobvious truth about how one or more inflections can be harnessed to radically change human capacities and behaviors).

Wechat’s “inflection” was the advent of the internet, enabling texting over data rather than through cellular providers, and the “insight” was that people do not want to be rate-limited and charged per text. They want to communicate freely.

This was a major point missed when I limited my scope in searching for the “superapp” of the West because the old “inflection” that worked for WeChat has long been gone. But a new one has arrived: the maturation of AI, and a new “insight” that AI will increasingly become the entry point for consumers’ online journeys, to the point that it’s very fathomable to me that humans will grow more reliant on AI for decisions and recommendations across many aspects of their lives.

SuperApps need Super User Base + Distribution

Going back to the foundational blocks of a “super app”—that is, a lot of people have to use your app in the first place for it to become “super.” In other words, whoever wins distribution has the highest probability of winning. This has become obvious for GPT. By the numbers, here’s a comparison generated by GPT based on reported and estimated figures.

ChatGPT’s MAU is already leading, coupled with the strongest growth trajectory, putting it on track to accumulate the distribution and user base needed as the foundation to become a de facto “superapp.”

Now Lets Compare the 6 Value Propositions of WeChat vs. GPT

1. A necessity: ✅ GPT is increasingly becoming one, provided it continues to dominate general-purpose LLM. Competition from Gemini is fierce but the “superapp-store” vision is not glaring from Google’s side.

2. A fintech tool: ✅ OpenAI recently announced partnership with PayPal allow users to make purchases directly within the chat interface. Without a shadow of doubt, more will come - even crypto if I’m making predictions.

3. A social network: ✅ Many will disagree with this, but what if, in the future, we socialize with AI more than with humans? The very definition of a social network is on the verge of being rewritten. Personally, I already spend a significant amount of time talking to AI, assigning it multiple personas—personal assistant, executive coach, psychiatrist, book summarizer, tour guide, etc. The more it knows about me, the more personalized and effective it becomes. GPT and Gemini are already taking up an increasing share of my “free” time, and the ROI on time is often higher than in many human-to-human interactions.

   Even if you disagree and still bet on “real human” chat as the core of social connection, GPT already supports group collaboration and sharing of queries and conversations.

4. A services gateway: ✅ This is super obvious. More people are already turning to GPT for what used to be Google Reviews or Yelp. Sam Altman has discussed the conflict-of-interest risks in trying to monetize this layer and does not plan to do so. He’s smart. The north star is to get people to trust your platform & stay—if credible neutrality is what it takes, there will be a thousand other ways GPT can monetize its user base.

5. A discovery platform for content and services: ✅ “GPTs” is a blatant attempt to replicate the success of the Apple App Store in the age of AI. Conservative estimates put the number of GPT wrapper startups in the high thousands, and that number is only growing. Even if GPT were to remain free (which it is not), these startups will monetize on top of it—at which point there’s no reason for OpenAI not to rent-seek. The B2B angle here will be pretty fat.

6. An entertaining “brainrot”: ⏳ My guess is that either “Pulse” will evolve into this, or new tabs will be introduced for media and entertainment like “trending” queries. Once a platform aggregates attention and brand out to become the “discovery” platform, the path to ads monetization is direct and inevitable. Entertainment has always been the final layer built on top of utility and distribution - this is a natural evolution.

Net, the new “superapp” of the AI era will emerge at the convergence of two forces: first, who owns the largest user base as the broadest distribution platform; and second, who can then vertically expand the depth of services to subsume every GMV source in form of consumer demand - which by the way, is pretty predictable and timeless: 1. how they eat 2. where they go 3. what they wear.

Can other major LLMs like Gemini, Kimi, or Grok do this? Maybe—if they 1) succeed in winning public perception as general-purpose, neutral LLMs, 2) move fast enough, and 3) make their intentions more obvious to properly prime their customers. My prediction is that 2-3 general-purpose LLMs will emerge as winners, while everyone else pivots to verticalized or use-case–specific LLMs (my guess is Grok will belong here).

Now, the most relevant question for founders: in a world where GPT or Gemini becomes the sweeping everyday app/“WeChat,” how will innovations evolve, and how should you position yourself?

1. DEX-Only Viability: As projects become increasingly allergic to CEX listing token allocation asks, more are floating the idea of a DEX-only launch—similar to the “HYPE” style. But can project teams drive 9–10 figure outcomes with a DEX-only launch? How replicable is the HYPE model?

Note: A DEX-only launch = mentally preparing for shallow depth, especially in the early days, and potentially never being able to tap into CEX liquidity due to stringent day-two listing requirements

1.1.[Extension from Q1] Feasibility of DEX Tokens for Institutional Buyers: How can liquidity constraints on DEX launches be sufficiently solved by time and market adoption alone—without relying on CEXs—to eventually support institutional size (for both buy + sell)?

2. The “Token Comeback”: What earns a token the right to “come back” from the dead beyond the obvious answer of “revenue?” Rev ≠ guarantee of survival. What are the qualitative drivers & how they weigh against quantitative across different types of tokens?

3. Off-Chain Revenue Impact: How effective is off-chain revenue (think DePIN or B2B rev) in driving a token valuation premium?

i.e., If an L1/L2/DePIN network secures 9-figure licensing contracts (B2B revenue), but such revenue remains largely off-chain (accruing to equity vs. the token), how much does it actually affect the token market cap?

4. Crypto startups’ right to win in AI: I’m of the belief that the “Davids” in crypto AI building “general-purpose LLMs” do not have a right to win against the “Goliaths” of Big Tech incumbents (happy to debate this).

With that premise, what are the specific verticals or use-case specific models where crypto AI companies actually possess a unique edge?

5. Longtermism in Crypto: [Perhaps the most important Q] What motivates founders to remain long-term oriented in crypto? What is the VC’s role in fostering it?

Let’s call it: the “Leakage Theory.”

The core idea: minimize unnoticed energy leaks to maximize your “Goal Throughput.” Leakage shows up in many forms

🟣 Ambiguity Leakage

Not knowing what you want is the biggest energy leak. In both work and life, if you're unclear, every step you take might be wasted.

Meditation helps, so does ruthless clarity.

The faster you figure out what exactly you want:

→ the faster you get to your goal

→ the more goals you can set

→ the higher your "achievement throughput"

🟣 Thinking Leakage

Know the difference between thoughtfulness and rumination.

Here’s a simple check I use:

Will this extra thought lead to incremental action?

If not—it’s leakage.

Clear your mental “cache.”

Every thought that surfaces, convert into something tangible:

→ What action can come from this?

→ If none, what does the thought reveal about me?

Journal the learning. Then move on.

Brains run faster with no cache.

🟣 Interpersonal Leakage

Internal politics. People management. Being calculating.

Worrying about perception. Arguing for ego. “Proving a point.”

All of these burn energy that could be spent on creation.

🟣 Emotional Leakage

Namely: worry, doubt, fear.

These are consuming emotions that can be mitigated by action.

- If you’re worried about something, take preventative steps.

- If you’ve done absolutely everything in your power to prevent or mitigate it (like giving your all toward a job you want), there will be no regret— then why worry about it?

🟣 Attention Leakage

Where you pay attention is where you spend your energy.

So many self-help books emphasize the importance of self-love and self-improvement:

- Pay attention to yourself.

- Know what YOU want.

- Only try to control what you can.

In work:

Instead of stalking competitors, take your attention back.

Improve your product.

In life:

Stop trying to “change” others.

People only change when they want to.

Take your attention back & influence by example. Walk the talk yourself.

-

This is how I do it daily. Hope this thread gifts its audience a clearer mind and a lighter brain

In my experience, these comparisons tend to fall into one of two buckets:

1. Crypto GOATs Born from Web2 Duds: a breakout success in a stagnant Web2 sector, where crypto-native mechanisms unlock entirely new growth dynamics.

2. Crypto Duds Born from Web2 GOATs: a failed attempt to replicate a dominant Web2 business model, where fundamental structural differences in crypto break the very assumptions that made the original work.

First Bucket: Crypto GOATs Born from Web2 Duds

In crypto, some of the most compelling use cases emerge when a Web3-native approach redefines a plateaued Web2 category. These aren’t just crypto clones of existing Web2 products—they’re fundamentally different bootstrapping machines that unlock entirely new demand curves. By embedding crypto incentives into go-to-market strategies, these projects solve the cold-start problem and activate users in ways Web2 cannot.

Take Worldcoin as an example.

The identity sector in Web2 is mature and heavily consolidated. Enterprise platforms like Okta, Microsoft Entra, and Google dominate. Power is too concentrated in platforms & access to user data is almost entirely walled-off. Any startup aiming to build a user-owned identity graph from scratch faces near-insurmountable barriers.

Yet Worldcoin did it. Within its first 6–8 months, it onboarded over 10 million users across 160+ countries, collecting one of the most sensitive data types—retina scans—at global scale. For a Web2 identity startup, this level of adoption would require years of compliance vetting, trust-building, cross-border partnerships, and user acquisition spend.

So why did it work? The answer lies in a user-incentivization loop that only crypto can create.

Worldcoin designed its reward structure to front-load user reward through a tapering emission schedule. At launch, users received up to 25 WLD as a “genesis grant,” with ongoing monthly rewards starting at ~3 WLD and decreasing over time. This created urgency—early users earned more for the same action—and a clear countdown effect that pulled forward demand.

One might argue that Web2 startups could also just pay users like this. But there are two critical distinctions:

1. User Subsidies Funded Externally (by the Liquid Market) in Web3 vs. Internally (by the Company) in Web2
   
   Worldcoin distributed over 525 million WLD tokens. During the early distribution window, WLD traded between $3–$10, equating to $1.5B–$5B in value delivered to users—without the company spending from its balance sheet. These weren’t coupons or ad credits. They were liquid assets, priced and distributed by the open crypto market. In Web2, only hyper-capitalized companies like Uber have executed user acquisition at this scale, and it was entirely funded through venture dollars from investors, costing billions in cash burn.

Crypto uniquely gives startups access to tap external liquidity from the crypto capital markets at previously unfathomable scale to overcome the cold start problem—without diluting equity or draining treasury.

2. Crypto Automatically Turns Users into Stakeholders
   A user on Instagram doesn’t necessarily care about Meta’s share price unless he/she bought the stock separately. But when a user receives a Worldcoin airdrop, they are automatically financially tied to the network’s success. That alignment changes behavior: users become active promoters, referring friends, defending the project, and contributing to its growth.
   A key reason this works is that token incentives are dynamic, not fixed. In Web2, referral rewards are typically static—$10 is always $10. But in crypto, 15 WLD could be worth $5-$150, depending on market performance. This price variability amplifies emotional and financial engagement. Users don’t just earn—they speculate, evangelize, and stay invested. Their upside is no longer capped.

In essence, tokens are more than just rewards—they are a viral, financial coordination mechanism, turning users into an organic, decentralized marketing engine fueled by ownership.

The “Self-Fulfilling Effect” of Token Networks

Beyond a customer-acquisition hack, the crypto component can also be a powerful engine for sustainable user growth past the initial phase. A token network creates a reflexive loop between belief, participation, and value—a phenomenon I call the Self-Fulfilling Effect. In this loop, belief itself becomes an active economic force: it bootstraps supply, demand, and legitimacy simultaneously.

Take Offline Protocol as a current example. It aims to build infrastructure and applications that function without internet or traditional telecom—a radically contrarian bet on the future of sovereign, offline connectivity.

Here’s how the self-fulfilling dynamic plays out:

If I believe that offline-first infrastructure will be essential
→ Then my Offline ID and token could become valuable
→ That belief drives me to participate early—to farm, test, refer
→ More users = more local nodes = more utility
→ That utility reinforces belief, and belief accelerates adoption

The result is a self-reinforcing flywheel leading up to 300k+ organic signup with zero marketing dollar. Tokens don’t just represent value—they create it, by giving users a financial stake in a future they actively build.

This isn’t new. Bitcoin is the canonical example of the “Token Self-fulfilling Prophecy”. In its early days, miners believed Bitcoin could become a store of value → they mined despite low rewards → belief spread → more miners joined → hashpower increased → security improved → institutional belief followed → price rose. A belief loop became a value loop.

Tokens are more than incentives—they are economic coordination tools that drive belief, behavior, and growth at the same time.

Second Bucket: Crypto Duds Born from Web2 GOATs

I have yet to see many crypto skeuomorphs of successful Web2 companies achieve resounding success. Projects that set out to become the “Google, Snowflake, Amazon, Facebook of Web3” have largely fallen short.

Maybe Binance comes close to playing the “Nasdaq of Web3,” but that’s a rare exception.

Why is that?

My hypothesis is that what made the magnificent 7 so dominant was Business-to-Business-Model Fit as below. Let’s examine three dominant business models from Web2—Marketplace, Data & Attention, and SaaS—and why their Web3 counterparts faltered.

1. Marketplace Business Model

Web2 Example: Amazon, Airbnb, Etsy, Uber

Monetization:

• B2C: Transaction fees, markups on sales

• B2B: Ads, platform fees from sellers

Success Factor: Strong network effects—more buyers attract more sellers, creating scale-driven defensibility. Such network effects create sticky, winner-take-most platforms. More users = more value = higher switching costs = lower acquisition costs.

However, in crypto, Network Effect does not work due to below three structural differences:

• Crypto Users Tend to Be Mercenary: Developers treat chains like products—choosing based on fees, speed, liquidity, and grants. Retail users chase yield and slippage, not UX or brand. Validators stake where reward is highest. Loyalty doesn’t factor in.

• Crypto is open-source by default, which drastically lowers the barrier to entry for copycats: Anyone can fork. SushiSwap copied Uniswap’s code and UX, then added better token incentives to drain users—a “vampire attack” that would be impossible in Web2 without stealing Facebook’s entire backend and bribing users to move.

• Crypto is interoperable by default, which minimizes switching costs for both developers and retail users: Swapping USDC to USDT or PYUSD takes seconds. On the other hand, switching a card from the Visa network to Mastercard is much more onerous.

❗️In some cases, it can even be argued that crypto has the opposite of the Network Effect: more users = lower yield (in DeFi) or higher gas fees (on congested chains).

As a result, while the marketplace model has been responsible for some of the most successful web2 tech giants, it is not a defensible business model in Web3—because it relies heavily on the network effect, which fundamentally does not hold in crypto. For a deeper look at why the network effect breaks down in crypto, see this deep-dive piece.

2. Data & Attention Business Model

Web2 Example: Google, Facebook, ad-driven publishers
Monetization: B2B: Sell access to user data via targeted ads
Success Factor: Proprietary data and ownership over user relationships

However, proprietary data or “data moat” simply doesn’t exist in the onchain world. Everything on-chain is fair game and accessible to everyone. One could argue that there are data moats to be had from:

• Organizing and structuring on-chain data, and competing on the quality of such service (e.g., indexing protocols)

• Combining on-chain and off-chain data (since most metadata—such as customer information, content, etc.—is stored off-chain), which can be proprietary

Sure, but to monetize data, this on-chain + off-chain combination would need to be made “proprietary” and closed (e.g., Chainalysis), which inherently contradicts Web3’s ethos of transparency and decentralization.

Also, the degree of exclusivity or proprietariness of data—which often serves as a proxy for its implied value to potential data buyers—is much lower in Web3. Facebook has complete monopoly over its customer data, giving it absolute pricing power.

By contrast, even one of the most lucrative data companies in Web3, Chainalysis, faces fierce competition from TRM Labs. Chainalysis may maintain competitive presence due to feature depth, oversight, and UX—but it still works off the same on-chain data out in the open as its competitors.

3. Software Subscription (SaaS) Business Model

Web2 Example: Snowflake, Salesforce, AWS, OpenAI
Monetization: B2B: Subscription-based recurring revenue
Success Factor: A large user base + a superior, sticky product to generate predictable recurring revenue and low churn.

SaaS has been the crown jewel of software business models over the past decade. Its success—valued at $384B in 2024, projected to exceed $1.8T by 2034—comes from:

• Massive TAM: virtually every company across sectors—finance, logistics, healthcare, media, government—benefits from SaaS. It addresses nearly the entire GDP.

• Accounting benefits: predictable recurring revenue, OPEX vs. CAPEX shift, easier budgeting.

• Operational efficiency: reduced IT overhead, faster onboarding, better infra management.

• Financial clarity: subscription-based revenue, high margins, and cost predictability at scale.

In contrast, Web3-facing B2B SaaS serves a much smaller economy. For a Web3 SaaS company going to market, the pool of potential buyers is hyper limited. Outside of giants like Coinbase or Binance—who are actually more likely using mature Web2 software stacks for IT management, most Web3 companies lack enterprise-grade spending power. These teams are lean, often pre-revenue, and operate with limited software budgets. Formal procurement processes, such as multi-year contracts, are nearly impossible.

Of course, one could argue that a Web3 B2B SaaS company can still target the broader Web2 enterprise space—yes, theoretically. Several counterarguments that make it infeasible are:

1. Web2 enterprises sales requires a completely different skill set and salesforce than selling to Web3 startups.

2. Few Web2 enterprises are adopting truly crypto-native or on-chain products, as blockchain is not yet core to their business operations.

3. Competition in the Web2 SaaS space is far more intense, with a full suite of well-established startups that often deliver superior UX and enterprise readiness.

Now, what are unique moats in web3 warrant a separate article for another day.

Broadly, there are two triggers for both joy and distress: Objective causes vs. Subjective Causes

Objective Joy & Jitter

Objective causes refer to experiences that impose a sensation directly, without needing to be witnessed, validated, or interpreted by others. They are self-contained and inherent, arising from the experience itself — unmediated, intrinsic, and not dependent on external observers.

Objective Joy

This is joy that exists on its own terms. It doesn't require a third party to validate it or reflect it back to you. Ask yourself: if no one else were around to witness this, would it still bring the same satisfaction?

Examples:

• Pleasure in taste regardless of who you're dining with or what they think

• Spotting a beautiful rainbow

• Freedom — financial, time, or location

• The dopamine hit from a great workout or long run

• Solving a problem you’ve been stuck on for months

• Using high-quality goods and materials

• Self-actualization (without the intent to become famous)

• Presence and peace

• Confidence rooted in internal validation, not comparison

Objective Distress

Suffering that arises directly from a state of the body or emotion. It does not rely on social context or perception to become real.

Examples:

• Physical discomfort

• Financial pressure to meet basic needs

• Fatigue or burnout

• Grief from loss

• Overwhelming noise or sensory overload

Subjective Joy & Distress

Subjective experiences are defined by external inputs, societal norms, peer comparison, or the perception of others. The feeling itself is dependent on the presence of an observer or an implied audience.

Subjective Joy

Examples:

• Buying luxury goods, like Gucci sunglasses — a non-luxury pair may use the exact same materials, but the joy comes from the symbol, not the object. If the world no longer recognized Gucci as a prestige brand, would the emotional return still exist? Likely not.

• Fame and vanity — admiration that fades if the crowd stops watching

• Social media followership and likes

• Perceived popularity or status

• Peer approval, affirmation, applause

These are forms of joy whose emotional payoff evaporates if no one notices — the pleasure is borrowed from another’s gaze.

Subjective Distress

Examples:

• Pressure from not conforming to societal “norms” (e.g., not immediately finding a job after graduation or pursuing one of the “socially-proof” careers like consulting/IB after graduation). Distress doesn’t stem from internal misalignment, but from perceived disappointment or non-acceptance by others (parents, peers)

• Insecurity that originates from external comparison: feeling “less than” because someone else appears more successful, even if you're aligned with your own values

• Imposter syndrome — one may feel distressed not because they objectively lack ability, but because of perceived standards set by others — imagined scrutiny, peer comparison, and fear of being "found out."

These are illusory joys and distresses — fluctuating with social winds.

The Mixed Layer (Objective + Subjective Overlap)

Many real-life experiences blend both objective/intrinsic and subjective/dependent causes. It's rarely black and white.

Examples:

• Buying a Ferrari for the thrill of its speed and build (objective joy), and for the status symbol it signals (subjective joy)

• Achievement: deep fulfillment from realizing a personal goal (objective), while also enjoying the spotlight of praise and recognition (subjective)

• Insecurity: rooted both in internal doubt and external comparison — a hybrid emotion

Cross Currents: When Joy and Distress Clash

Some situations contain opposing emotional vectors — where joy and distress come from opposite sources at the same time.

Objective Joy + Subjective Distress

Example: Eating alone at a restaurant or traveling solo
→ The activity brings real autonomy, freedom, and mental clarity (objective joy) — yet may carry low-level social discomfort from the perception of being “alone” in a space designed for groups (subjective distress).

Subjective Joy + Objective Distress

Example: Wearing painful heels to a formal event
→ Socially, heels are flattering (subjective joy), but physically, you suffer real, undeniable discomfort (objective distress). The pain is intrinsic — the joy is external.

So What’s the Point?

Subjective joy and distress are shaped by mind — not reality. Do not let them dominate. They're ephemeral, amorphous, and often illusory. Their grip on you is based on how much you care about how others see you.

And get this - people don’t pay attention to you nearly as much as you think they do — it’s one of the fundamental human cognitive biases — called the Spotlight Effect — to overestimate how much attention others are paying to us.

Most people are too absorbed in their own inner world to judge yours — yet our brain tricks us into acting like we’re always being watched, evaluated, or compared.

The Real Pursuit of Happiness

So here’s my working principle for happiness:
Prioritize objective joy — that’s the only kind that’s real and that matters. They don’t need witnesses. They hold steady when the crowd leaves.

Track how much time, money, and mental energy are spent chasing subjective joy or avoiding subjective distress — and start cutting back.

The more you invest in what’s real, the more stable your inner core. Own your happiness, don’t borrow it.

Now - what is your Objective / Subjective Joy Ratio?

Since then, the field has advanced dramatically, with major breakthroughs making headlines in 2024 and 2025:

• D-Wave Claims 'Quantum Supremacy' in Practical Problem (March 12, 2025)​

• Microsoft Introduces Majorana 1 Quantum Chip (February 19, 2025)​

• Google Unveils 105-Qubit Willow Processor (December 9, 2024)​

• IBM Launches Modular Quantum System Two (November 2024)​

• ​Quantinuum's H-Series Achieves Record Quantum Volume (October 2024)

This piece is a follow-up to that original article—grounded in refreshed examples in detail —to illustrate just how transformative quantum computing will be across some of the world’s most critical sectors.

Please note, this article is heavily assisted by ChatGPT.

Digital Assets (Cryptocurrency & Blockchain)

• Near-term: Quantum computers threaten current cryptographic safeguards underpinning digital assets. Shor’s algorithm could factor large encryption keys, undermining RSA and ECC security that protect crypto wallets – a task infeasible for classical computers. This is driving an industry push toward quantum-resistant cryptography long before large quantum machines arrive, ensuring cryptocurrencies remain secure as quantum hardware progresses.

• Long-term: Fully scalable quantum computers could break today’s blockchain encryption in a practical timeframe, enabling theft or forgery of digital assets that would have been impossible with classical computing​. For instance, deriving a private key from a public key (which would take a classical computer billions of years) may be done in minutes or hours on a future quantum machine. Such an exponential speedup would force a complete overhaul of digital asset security. On the upside, quantum technology may also introduce ultra-secure transaction methods (like quantum key distribution) and new cryptographic algorithms, unlocking unprecedented security paradigms for digital asset exchanges.

Financial Services & Banking

• Near-term: Quantum computing offers powerful tools for risk management and trading optimization. Banks are testing quantum algorithms (e.g. quantum Monte Carlo methods) to price complex derivatives and simulate market scenarios faster than classical HPC allows. This yields more accurate risk assessments and portfolio optimization by considering a broader range of variables in real time​.

  Early prototypes show quadratic speedups in Monte Carlo simulations – problems taking days on classical systems could run in hours – giving firms a competitive edge in decision-making. Major institutions like JPMorgan and Goldman Sachs have even formed in-house quantum teams to explore these advantages.

  Additionally, financial firms are “quantum-proofing” their infrastructure by upgrading encryption now, ensuring secure transactions before quantum attacks become feasible.

• Long-term: In a decade or more, full-scale quantum computers could solve financial optimization problems of staggering complexity that classical computers can’t handle at all.

  • Global portfolio optimization with thousands of assets, real-time arbitrage across markets, or instantaneous fraud detection on vast data sets could become reality

  • Quantum algorithms promise to find optimal investment strategies or loan portfolios in polynomial time – an enormous leap over today’s NP-hard brute-force limits. The magnitude of improvement could be transformative: tasks like risk simulations that would take classical supercomputers 10,000 years might be done in seconds, enabling near–instantaneous stress tests and economic forecasts. As quantum hardware matures (with fault tolerance beyond 2030), using it will move from a novelty to a necessity in finance​.

Biotech & Pharmaceuticals

• Near-term: Quantum computing is already being explored to speed up drug discovery and biomedical research. Early quantum processors can simulate small molecules and protein fragments with higher accuracy than classical methods, which struggle with quantum chemistry at scale. For example, pharma companies are using quantum algorithms to model how a candidate drug binds to a target protein – computations that classical computers must approximate (or cannot do at all beyond a certain molecular size)​.

  Even with today’s nascent hardware, there are prototypes: Moderna and IBM have used a quantum algorithm (CVaR-VQE) to predict mRNA molecule structures more accurately than classical solvers, aiming to design stable mRNA vaccines faster​. These near-term experiments show how quantum computers can tackle biomedical problems (like protein folding or genomic pattern recognition) that were previously impractical, potentially cutting initial R&D iterations from months to days.

• Long-term: Quantum computing could revolutionize biotech by making the in silico simulation of complex biological systems routine. Tasks like accurately modeling the folding of large proteins or the interaction of a multi-step biochemical pathway – currently impossible for classical supercomputers – would become feasible.

  This means researchers could virtually test thousands of drug molecules against a disease target with atomic precision, pinpointing the best candidate instantly. The improvement unlocked is massive: imagine reducing the drug development cycle from ~10 years down to a couple of years or less through computational breakthroughs. In medicine, quantum-powered analysis might enable personalized treatment optimizations (by crunching through a patient’s entire genomic and proteomic data to find tailored therapies, something classical AI can’t fully manage).

  In short, quantum computing could yield cures and treatments for complex diseases by exploring biochemical possibilities that are astronomically large in number – a quantum leap in both speed and scope, likely materializing in the 10+ year horizon as more powerful quantum machines come online.

National Economic Planning (Government Modeling & Budgeting)

• Near-term: Governments can begin harnessing quantum computers for enhanced economic simulations and scenario planning on a limited scale. While current quantum hardware is small, it can be used in hybrid quantum-classical algorithms to analyze specific complex problems – for example, optimizing a subset of a national budget or running simplified models with many variables.

  Quantum-inspired simulators might help model supply chain shocks or epidemic impacts on the economy more accurately by sampling many scenario variants in parallel. In the next 5–10 years, we may see pilot projects where quantum algorithms assist in forecasting tax revenues or outcomes of infrastructure investments, offering insights beyond the reach of standard models. These early applications will likely be specialized and run alongside classical supercomputers, but they lay the groundwork for governments to develop expertise in quantum-based decision support tools.

• Long-term: Quantum computing holds the promise to model an entire economy with unprecedented detail, transforming national economic planning and policy-making. Classical economic models are limited by computational complexity – for instance, simulating the full global economy or a national budget with millions of interdependent agents and policies is intractable today. Quantum computers could handle far larger, more complex models: one analysis noted that quantum machines will simulate financial networks at a scope that would require an impractical amount of classical resources (more than atoms in the universe) to replicate​.

  In practice, this means governments could run vast “what-if” simulations of fiscal and monetary policy choices (budgets, taxation, stimulus, etc.) and get results with high fidelity, helping them choose optimal strategies for growth and stability. They could also optimize multi-decade strategic plans – for example, finding the ideal allocation of trillions of dollars across education, defense, healthcare, etc., to maximize GDP growth or social welfare, a combinatorial optimization classical methods can’t fully solve.

  The impact of such capability is enormous: one study predicts quantum computing could boost a nation’s productivity by ~7% by 2045 through better decision-making and innovation​. Long-term (10+ years), quantum computing may become an indispensable tool in government planning, allowing leaders to navigate economic complexities with a clarity and foresight never before possible.

Materials Science & Chemistry

• Near-term: Quantum computers are uniquely suited to simulating chemical and material systems, and early demonstrations are targeting high-value problems like battery chemistry and industrial catalysts. In the next few years, quantum algorithms (like the Variational Quantum Eigensolver) will be used to calculate properties of modest-sized molecules more exactly than classical chemistry codes can.

  For example, automakers such as Daimler and Hyundai are already partnering with quantum computing firms to study new battery materials at the quantum level​. These initial applications handle molecules that are on the edge of classical computing ability (e.g. certain lithium compounds) – classical supercomputers often have to resort to rough approximations for these, whereas quantum processors can in principle model them without loss of fidelity​.

  The benefit in the near term is targeted: even slightly improved accuracy in simulating a battery electrolyte or a catalyst’s reaction pathway can guide researchers to better designs, potentially shaving years off material R&D. We’re seeing the first quantum-guided discoveries (e.g. identifying a promising chemistry for a longer-lasting battery or a more efficient solar cell) on small scales, as a preview of what’s to come.

• Long-term: In the 10+ year horizon, quantum computing could unlock materials and chemical discoveries that elude classical computing entirely. Many critical industrial challenges – like designing a room-temperature superconductor, a lightweight super-alloy, or a new carbon capture material – involve quantum interactions among electrons that are too complex for any classical simulation. A fully capable quantum computer can directly simulate these quantum systems, effectively letting scientists “test” new material structures virtually with exact physics.

  The magnitude of improvement is illustrated by the case of fertilizers: finding a catalyst to produce ammonia (fertilizer) at low energy could save huge amounts of energy worldwide, but the molecular complexity (e.g. the enzyme Ferredoxin’s reaction) is beyond classical modeling. Quantum computers could solve this chemistry, leading to a catalyst that replaces the century-old Haber-Bosch process (which currently consumes about 2–3% of global CO₂ emissions)​ – a groundbreaking environmental and economic win.

  Similarly, quantum-designed catalysts, drugs, or materials could be discovered in days versus decades, fundamentally accelerating innovation in energy, manufacturing, and healthcare. The long-term vision is that quantum computing becomes a standard part of the materials scientist’s toolkit, routinely delivering breakthroughs (new compounds, stronger materials, greener chemical processes) that were impossible to find before.

Logistics & Transportation

• Near-term: Quantum computing excels at complex optimization problems, and logistics is rife with such challenges. Already, specialized quantum annealers have been used in pilot projects to optimize traffic flow and vehicle routing. A notable example is Volkswagen’s trial in Lisbon, where a quantum algorithm calculated the fastest routes for each bus in a fleet in near real-time, reducing commuter travel times and improving city traffic flow. These early successes target scenarios that classical algorithms handle sub-optimally – like coordinating many vehicles or deliveries simultaneously without causing jams or delays, an exponentially hard problem as systems scale.

  In supply chain management, companies are testing quantum solvers for things like warehouse picking routes and delivery scheduling, aiming for even a few percent efficiency improvement. (In the automotive industry, for instance, simulations show that optimizing robotic assembly paths via quantum methods could improve productivity by 2–5%, translating to $10–$25 billion saved per year in manufacturing costs​)

  Over the next 5–10 years, we can expect to see quantum optimization software in use for specific high-value logistics sub-problems – essentially acting as an “optimizer plug-in” alongside classical systems to find better solutions for routing, scheduling, and resource allocation.

• Long-term: The logistics sector could be completely transformed by large-scale quantum optimization. With advanced quantum computers, it becomes feasible to compute optimal solutions to problems that are currently unsolvable: think globally optimized supply chains considering every factory, truck, ship, and warehouse in one calculation, or real-time optimization of all traffic lights and routes in a megacity to eliminate congestion.

  Such tasks involve astronomically many possibilities (combinatorial explosion) that classical computers can only ever approximate. Quantum computers, however, might handle these at scale, finding routes and schedules that minimize fuel usage and transit time far better than today’s heuristics. This could mean

  • cheaper and faster deliveries for consumers (as shipping companies optimize container loading and routing across the world)

  • significant fuel and energy savings (benefiting both profitability and the environment), and

  • much more reliable transportation networks. In air travel, for example, quantum optimization could schedule flights and crews with zero conflicts and minimal delays across an entire airline’s network, something currently beyond reach.

  The unlock here is not just speed but quality of solution – reaching the true optimal or something very close. In the long-term future, every major logistics decision (from urban planning to global distribution) may be vetted through quantum optimization models to ensure maximum efficiency. This promises an industry-wide leap in productivity, cost savings, and capability, making today’s logistical challenges a solved problem in retrospect.

Sources

1. Coinbase – “Is quantum computing a threat for crypto?” (2023) – Explains how quantum algorithms could break cryptocurrency encryption by cracking private keys, and notes it may take a decade or more for practical attacks​coinbase.com​coinbase.com.

2. Foley & Lardner LLP – “Quantum Computing’s Transcendence: Impacts on Industry” (Nov 2024) – Discusses quantum computing’s impact on various sectors. Highlights for finance: enhanced risk modeling and portfolio optimization with broader variables​foley.com. Highlights for pharma/biotech: inability of classical supercomputers to model molecules like caffeine and proteins, and quantum’s potential to simulate complex molecules for drug discovery​foley.com​foley.com.

3. Deloitte Insights – “Quantum computing in financial services” (2023) – Industry report noting key use cases in finance (Monte Carlo simulations, portfolio optimization, etc.) and that major banks (Goldman Sachs, JPMorgan Chase, HSBC, Barclays) have dedicated quantum teams to develop algorithms for these problems​www2.deloitte.com.

4. IBM News/Blog – “How will quantum impact the biotech industry?” (2023) – Describes a collaboration with Moderna using a quantum algorithm (CVaR VQE) to predict mRNA structure stability, outperforming a classical solver and aiming at problems classical computing can’t handle as sequence length grows​ibm.com.

5. Oxford Economics & IBM – Quantum Computing and Economic Growth (Press Release) (Feb 2025) – Study forecasting that quantum computing could boost the UK’s productivity by up to 7% by 2045 (adding ~£212 billion to GDP) if its potential is realized, underscoring significant long-term economic benefits​oxfordeconomics.com.

6. Forbes Tech Council – “Could Quantum Computing Better Predict and Prevent Economic Downturns?” (Aug 2022) – Suggests that quantum computers will be able to simulate financial networks at a scope requiring more resources than there are atoms in the universe for a classical computer​forbes.com, hinting at their superior capacity for modeling extremely complex economic systems.

7. Quantum Flagship (EU) – “Fertilizer and other quantum computer chemistry” (2023) – Explains the challenge of modeling the enzyme ferredoxin (involved in biological nitrogen fixation) on classical computers and how a quantum computer could help discover a low-energy alternative to the Haber-Bosch process, which currently consumes 2–3% of global CO₂ emissions​qt.eu.

8. IonQ Blog – “Improving Battery Chemistry with Quantum Computing” (2022) – Details how classical methods cannot fully simulate even moderately sized molecules (e.g. lithium dioxide) without approximations​ionq.com. Describes partnerships with automakers (Daimler, Toyota, Hyundai) using quantum algorithms to explore new battery materials, leveraging quantum computers to tackle chemical reaction simulations that are intractable for classical computers​ionq.com.

9. Volkswagen News – “Volkswagen optimizes traffic flow with quantum computers” (Oct 2019) – Reports on a pilot in Lisbon where a D-Wave quantum computer was used to calculate optimal routes for buses in real time, reducing passenger travel times and improving traffic flow​volkswagen-group.com. This was one of the first real-world demonstrations of quantum optimization in transportation.

10. McKinsey & Co. – “Quantum computing use cases are getting real” (Dec 2021) – Identifies high-impact use cases in industries. Notes that in automotive manufacturing, even a 2–5% productivity gain from quantum optimization (for example, optimizing robot path planning in factories) could translate to $10–$25 billion in value per year​mckinsey.com, illustrating the scale of potential gains in logistics and operations.

Web2’s favorite growth hack reconsidered: why the network effect is no longer a durable moat in web3.

Let’s first get the definition of the Network Effect and why it matters in Web2. Per ChatGPT:

• Definition: A network effect occurs when a product or service gains additional value as more people use it. This means that each new user increases the overall value of the product or service for existing users.

• Benefits from Network Effects (NE):

  1. NE Increases Competitive Moat – More users make the product more valuable, blocking competitors.

  2. NE Lowers Customer Acquisition Costs – Existing users attract new users through word-of-mouth, integrations, or ecosystem effects.

  3. NE Creates Higher Switching Costs & Retention – As the network grows, users become more embedded in the ecosystem (e.g., social connections, data, integrations). This makes it costly or inconvenient for them to leave, leading to stronger retention and pricing power.

Expect pushback here, but I’m making a claim that Network Effects are not a lasting moat in crypto. They do not entitle crypto companies to the same staying power and sustainable competitive advantage as they do Web2 companies because of the below idiosyncratic traits in crypto.

Idiosyncrasy 1: Crypto Users Tend to Be More Mercenary

Developers as users: Developers are the users/buyers of blockchains (L1, L2, other “layers”). Blockchains offer developers a similar product: blockspace within an on-chain immutable database that records transaction history. Developers typically have common criteria when selecting where to build:

• Lowest transaction fees

• Fastest transaction processing

• Highest liquidity available

• Most ecosystem/community support, including grants

As evidenced in the chart above from Electric Capital’s Developer Report, Ethereum initially benefited from NE that attracted most developers to build exclusively on Ethereum (“Single Chain Developers”). However, the network effect was futile in saving it from its subpar performance and liquidity fragmentation in the face of performant competitors like Solana and Base, leading to a significant drop in the percentage of “Single Chain Developers” / “Total Monthly Active Developers” starting in 2022. This shift demonstrates the mercenary nature of developers, who flow to where their needs are met rather than staying for loyalty to the network.

Retail as users: As long as DeFi remains the primary use case in crypto, liquidity providers and retail DeFi users will continue to seek:

1. The highest yield on their liquidity

2. The lowest slippage in trades

3. The greatest variety of tokens

4. The most attractive farming rewards

This behavior often comes regardless of user experience or platform preference.

Moreover, wallets function as discrete agents in the crypto ecosystem. This makes switching between platforms—like Uniswap and Hyperliquid—frictionless and easy.

Validators as users: Validators naturally seek the highest notional value of block rewards—whether from their stake (in PoS networks) or from services rendered (as DePIN providers).

The decision to stay with or support an alternative L1, L2, app chain, or DePIN project comes down to a simple cost-benefit calculation. Validators evaluate perceived economic value and the sustainability of block rewards when making this choice.

Idiosyncrasy 2: Crypto is open-source by default, which drastically lowers the barrier to entry for copycats.

Enter the great invention of the “Vampire Attack” – SushiSwap copied the code and exact same UX from Uniswap, then designed a more lucrative token incentive to siphon Uniswap’s liquidity providers and users.

The equivalent attack in Web2 would be much harder to pull off. Someone would have to steal Facebook’s entire codebase, whip out an identical or superior product, then offer all of Facebook’s users money to lure them over to the new platform.

Idiosyncrasy 3: Crypto is interoperable by default, which minimizes switching costs for both developers and retail users.

Take the example of USDC, which arguably enjoys the highest degree of network effects in crypto, and compare it to one of its Web2 counterparts—the Visa network. If USDC is not accepted, it takes no time to swap it for USDT, USDe, or PYUSD on a DEX or CEX.

On the other hand, switching a card from the Visa network to Mastercard is much more onerous.

Now going back to my argument on why the Network Effect does NOT entitle crypto companies to the same benefits as their Web2 counterparts:

❌ NE Doesn’t Increase Competitive Moat in Crypto: Because of crypto’s forkable and open-source nature, on top of Bertrand-style competition among undifferentiated products (yield, blockspace, liquidity), network effects don’t necessarily make a first mover with more users more “competitive.”

❌ NE Doesn’t Lower Customer Acquisition Costs in Crypto: Crypto users — whether retail or developers — are much more mercenary in nature than the users of the Magnificent Seven in Web2. Retail optimizes for trading gains and yield. Developers optimize for the highest performance and deepest liquidity. Liquidity stays in an ecosystem for as long as the yield is lucrative for the LP, regardless of whether there’s a network effect.

An argument can even be made that❗️crypto has the opposite of the Network Effect: the more LPs in a pool, the lower the yield; the more users on a chain, the higher the fee and congestion.

❌ NE Doesn’t Create Higher Switching Costs & Retention in Crypto: Switching costs are minimal in crypto due to blockchain’s default composability and interoperability.

There’s no data moat in crypto either. Nothing on-chain can be considered “proprietary data,” which is a key retention lever used by big tech to keep customers in their walled gardens.

To cap this off, let’s examine a case study on Ethereum, widely regarded as the epitome of NE in crypto. Since its launch as the "World Computer," combining blockchain innovation with programmable money, Ethereum has benefited greatly from its early NE:

• Developer Adoption: It attracted the largest blockchain developer community early on, mainly because its EVM became the industry standard for initial blockchain development.

• Liquidity & DeFi Dominance: Ethereum hosts most of crypto’s liquidity through decentralized finance (DeFi) platforms – until being flipped by Solana lately. More liquidity attracts more users → easier and cheaper to trade/lend/borrow → more liquidity.

• Security: Increased usage on Ethereum enhanced its security, attracting more projects and users.

Yet, we all saw the trend break this year—Ethereum misplayed a winning hand by procrastinating on product improvements and over-fragmenting its ecosystem by supporting L2s that cannibalized its own liquidity. Its developer outflow (ETH’s monthly dev count falling 17% in 2024 while Solana’s grew ~83% in new contributors) is as abysmal as its liquidity outflow (dropping from 100% DeFi dominance to 50%, according to DeFiLlama). And Ethereum’s supposed NE is futile in stopping the bleed.

In contrast, Web2 behemoths—namely Meta and Twitter—have also slacked off in their innovation and shipping, yet continue to comfortably dominate their respective markets. Why? Because the Web2 version of NE actually works and has staying power:

• Competitors just can’t fork their code and offer similar products like they can with Ethereum.

• Twitter and Facebook data are truly proprietary and irreplaceable

• They are not interoperable with anyone except within their own ecosystems.

For these reasons, the traditional NE that secures long-term moats for Web2 companies just doesn't work in crypto.

Disclaimer: My firm (obviously) owns STX. This is not financial advice—only an investment thesis.

Most who know me for calling the BTC thesis early know that STX is my highest-conviction pick on the secondary market, and here is a consolidated list of why:

• OG Blue-chip Crypto Project: Stacks was founded in 2013 by Muneeb Ali, a PhD in Computer Science at Princeton University at the time, and Ryan Shea. Stacks is YC class of 2014 with subsequent backers (while private) including Union Square Ventures, Lux Capital, and more.

• The First Compliant Token: A key driver for institutional adoption is to use STX as the de facto BTC staking provider.

  STX was the first SEC-compliant token issuance and, in 2021, released its decentralization framework (also filed with updates to the SEC as a non-security). While many other major tokens will still undergo the battle with the SEC to fight off the "security token" label, STX has achieved sufficient decentralization like Ethereum did in 2017 and will not have to be distracted by legal fights going forward.

• The Most Optimal Return on Risk for Real “Whales” to Generate Yield on BTC with Peace of Mind: Earlier this year, I was preparing a meeting with a sovereign fund with massive BTC holdings. A thought came to mind—if he asked me my recommendation to generate yield on their 9-figure BTC holding, what are even the options out there and who would I be comfortable vouching for? A couple came to mind:

• Options/Structured Products Yield: The risk is the potential to lose some principal.

• Lending Yield: Principal protected, but there’s counterparty risk—and many still have the scar tissue from FTX.

• Protocol Yield: Perhaps the highest yield options with a wide spectrum of risks, mostly protocol-specific. Now let’s double-click to see the options:

  • STX: By far the longest-running protocol—founded in 2014 and token launched in 2018. It’s the earliest—and one of its kind protocol until the Bitcoin renaissance in 2023—to offer native Bitcoin yield natively for its stakers. Given the recent sBTC upgrade, the security of staking via STX is further graded by removing the need for a third-party custodian and now inherits 100% of Bitcoin security.

  • Arch Network, Babylon: Front-running solutions for protocol yields in the private market. Both enable native BTC yield generation with relatively low risks, but both are likely too new for conservative institutions’ risk appetite on their 9-figure holdings.

  • EVM Bitcoin L2s: Non-native and have major trust assumptions on the bridging solution, which are mostly via multisig. I’d be very hesitant to recommend protocol yield from EVM BTC L2s to someone like a sovereign fund.

• The Undisputed Forerunner in Bitcoin L2 with 10 Years of First-Mover Advantage: Stacks has 10+ years of first-mover advantage over other new BTC L2s arising mostly between 2022-2023. STX is one of three BTC L2s that I can say with 100% certainty is not going anywhere.

• A Rare Token Standard Moat: as detailed in this X thread, one of the only true moats of any ecosystem (L1 or L2) is its own token standard. And it’s harder than most think to build an ecosystem of startups using your token standard. This is why STX’s first-mover advantage and their own SIP 10 token standard will pay long-term dividends—this moat is very hard and takes very long to build, but once it’s there, liquidity and projects stay in your ecosystem. In the duopoly of EVM and SPL token standards as the two dominant threads, STX still managed to have 155 monthly active developers and 26 projects with live token built, and likely 60+ pre-launch startups building on top—that is no small feat

• STX is Already an Ecosystem: Building an ecosystem—aka persuading founders to adopt your token standard (vs. using EVM or SPL)—takes Herculean efforts to achieve. In fact, this is why most new L1 and L2s are heavily subsidizing incubator/grant programs to attract builders.

• Benchmarking & Comparables: As the dominant and largest Bitcoin L2 on the liquid market, STX is at ~$3.7B FDV (Fully Diluted Valuation) compared to more recent Ethereum L2s/app chains: Celestia $9B, Optimism $11B FDV, Arbitrum $11B FDV. Don’t forget, BTC market cap is more than 5x that of ETH. I wrote about my—relatively conservative—prediction and rationale for STX to be a $50B project here.

• Token Maturity: STX’s circulating market cap/FDV ratio is 100% based on CoinGecko, fully vested, meaning there will be very limited sell pressure from unlocking/vesting. Personally, I’ve always stayed cautious of new tokens for long-term holding because of the sell pressure. 100% vested means that STX tokenomics has reached maturity—Solana is also about 81%, contrasted with other newer chains such as Arbitrum (~41%), Optimism (~29%), and Celestia (~41%).

• Upcoming ecosystem catalysts:

  • Officially becoming a Bitcoin L2 via the Nakamoto upgrade (complete): STX now inherit 100% of BTC finality.

  • sBTC (mainnet on 12/16)

  • WASM ClarityVM, opening the doors to other languages like Rust, Solidity, etc., directly at the Stacks L2 level.

  • L1 integration: bring BTC to other L1s like Aptos network via sBTC

[Crypto & Capital]: Research & Thinkings

Finding the Right Quantum Crypto Play is Harder Than You Think (Published under Portal Account)

[Noggin Noms]: Annotative Reading

A Linkable Notebook of My Annotative Reads & Listens

2024

October: A month of eclectic readings (book list to come)

November: A month of studying Charlie Munger

2023

[Candid Corner] Reflections & Actualizations

Organizing this X thread into an article form here.

What to expect in this thread:

1. An analysis of the DePIN business model & why it's challenging.

2. Areas where DePIN founders can create “moats” to stand out among competition

-

Context

I got into DePIN in 2022 before it was a word by researching the Decentralized Data & Storage Space during my MBA & produced this 34-page report. Subsequently, I joined Filecoin/Protocol Labs, the OG & DePIN category creator. There, I looked at 100+ DePIN projects across 20+ accelerators before getting back to investing at Portal Ventures

Why winning in DePIN is hard

• Reason 1. The curse of the "marketplace" business model

  All DePIN projects are marketplaces connecting latent supply and excess demand for goods & services. The web2 comparison is Marriott (centralized) vs. Airbnb (DePIN). While DePIN adds a "verification" layer, the value prop remains similar across categories.

  DePIN goods and services are generally commoditized: GPU, storage, bandwidth, energy, etc. Customers choose DePIN over incumbents mainly because:

  a. They're cheaper.

  b. They're more available (sometimes... e.g., decentralized vs. Nvidia GPUs).

  c. They are better? Let’s be honest—not really. DePIN often lags behind centralized competitors in service quality, support, or integration.

  See the pro vs. con comparison between Decentralized vs. Centralized Storage from my report

  

  Since the primary drivers for DePIN adoption are lower prices + greater availability, DePIN projects effectively enter into a Bertrand competition with others in the same category.

  Bertrand Competition = undifferentiated products, compete on operational efficiency to lower costs (price war) and availability of products on demand

  Brutal.

• Reason 2: both the demand side (customers) and the supply side (service providers/"nodes") are mercenary.

  It's very difficult to create network effects or sustain first-mover advantage in marketplace business model. Exhibit A: OpenSea (not DePIN ofc but an example how first mover advantage doesn't matter in marketplace biz model)

  As a customer, I go where I can buy the cheapest computing power; as a supplier, I go where I get paid the most for my GPU. There’s little loyalty to any single DePIN platform. Sure one can argue that Amazon/AWS is a marketplace and managed to sustain a strong network effect. But the difference is:

  • Amazon has spent 30 years relentlessly building demand-side relationships (known for customer satisfaction).

  • It owns a big portion of the supply (Amazon brand) vs. suppliers to DePIN platforms are generally individual actors

  • Amazon invested to develop major differentiators like logistics, Prime, etc. - we will get into how DePIN projects can also build moats like this

Now, let's be more constructive: how can a DePIN project stand out and win in its category?

• Moat 1: Winning at either or both ends of the barbell

  Own the demand side or the supply side first, then expand - discussed more in here. ExaBITS is a great example of this. It is now generating high 7-figure ARR by crushing the supply side and becoming the mothership supplier of most de-GPU providers

• Moat 2: Consistency and reliability

  Often, cheaper prices is not enough to move the needle, especially for institutional customers. No one gets fired for choosing Amazon/Nvidia, and there's a switching cost.

  Don't over-focus on costs and availability: recognize the importance of reliability and predictability in pricing and service quality. Attention to SLAs and enterprise-grade support will increasingly become a differentiator.

• Moat 3: Integration with ancillary services

  People stay with AWS over De-storage like Filecoin because of its entire product suite. System and platform integration is a $100B+ industry in web2 for a reason—people like a one-stop experience.

  Most DePIN projects are standalone offerings for one specific product, but integrating across multiple partners to provide a one-click experience is currently a gap AND opportunity

• Moat 4: [Short-mid term] Tokenomics design for price subsidy

  Most DePIN projects’ unit economics lose the advantage of economies of scale (enjoyed by their centralized counterparts), but can still offer cheaper services because

  a. they have lower OPEX costs to cover, saving of which can be passed down to customers

  b. latent supply generally are cheaper

  c. token subsidies. Service providers (like GPU suppliers) earn additional token subsidies/block rewards to offset the prices they demand from customers. Careful moderation of emissions and buy-back mechanisms can help maintain token performance & sustain competitive pricing

• Moat 5: Hardware

  Create (proxy) platform loyalty by forcing switching costs & off-boarding friction. Projects like helium or Filecoin require suppliers to acquire specialized hardware equipment to join the network. While such hardware requirements raise suppliers' barrier to enter, they also create a cost for offboarding from Helium/Filecoin, thereby fostering a form of proxy loyalty to the platform/marketplace.

• Moat 6: Regulatory differentiator

  For certain DePIN (or DePAT) projects like SkyTradeNetwork, platform-level regulatory complexity can become a true moat to help to fend off competition & sustain the first mover advantage + network effect

• Moat 7: Tech differentiator

  Depending on the type of project, the DePIN layer can also serve as a verification and trust layer between customers and service providers, especially in hardware-related DePIN (storage, GPU, bandwidth, etc.). For example, Filecoin has a major tech differentiator by generating ZK to ensure data integrity & tamper-proofness

/End

This might be my first travel journal. Before Bhutan, I can't recall a journey that stirred within me the desire to note down my thoughts, to integrate and synthesize them into something to reflect upon later. This time is different.

I’ve always taken an interest in exploring the depths of consciousness, free will, and the extent of control we hold over the "matrix" of our existence. Along this journey, I encountered diverse schools of thought at the intersection of physics, Buddhist philosophy, metaphysics, meditations, etc.

While each theory stands alone in its complexity, they all seem to converge toward three principles

• Intention: Our mind is a wild horse. Tame it, and it can take us anywhere we wish to go. It is the primary force through which we shape our reality. Sadly, many struggle with manifestation because they lack control over their mind, unable to crystallize their desires.

• Energy: Everything is energy. What we attract and are attracted to are dictated by our energetic vibrations. To become something new, we must recalibrate our energy.

• Intuition: Seek from within, or, as they say, "trust your gut." We each have a higher self, brimming with the wisdom and power to create our desired life. Intuition is the gentle nudge through which it communicates with us.

People meditate for different reasons. For me, active meditation is the crucible where these principles merge, creating a flywheel of intention, energy, and intuition. In each session, I quiet my mind, set an intention, recalibrate my vibration, and then listen to the quiet whispers of my “gut feeling”. I call this "SiFi"—set intention and follow intuition.

And Bhutan felt like the universe’s affirmation of this approach.

At the start of the year, I set my intention to deepen my connection with the universe. By May, I had heard about Bhutan for the first time at a family office event. A seasoned traveler described it as his favorite destination, out of all the countries he’d visited. It intrigued me, but I didn’t think much of it.

Fast forward three months, a new friend invited me to the Bhutan Innovation Forum. Although it seemed random, there was a strange pull toward the country - as though I couldn’t find a valid excuse not to go. Already being in Asia for Token2049 and SuperReturn, I leaned into it.

Later I learned, that you don’t just go to Bhutan. Bhutan calls you in.

I got there three days ahead of schedule because my intended flight was sold out—a twist that I later realized was all part of the “plan”. Had it not been for those unexpected extra days, none of the serendipitous encounters ensued would have unfolded.

Tiger’s Nest Monastery, the most sacred pilgrimage site in Bhutan, was my first destination. It had been raining all week, yet during the hours of my up-climb, the skies cleared. My guide happened to be a childhood friend of one of the two lamas stationed at the monastery. These lamas, often in years-long meditation retreats, are rarely available. Yet, on that very day, my guide’s friend had just emerged from a three-year, three-month, and three-day meditation training and was willing to see me at the monastery.

Moments after I arrived, a torrential downpour began, lasting about an hour. Because of this, I found myself alone in the monastery’s temples, meditating in absolute peace with only the sacred energy and lama’s low, resonant hum. Afterward, I was invited to the mysterious Royal Family Guest House for a private conversation with the lama. We talked lots about meditation: techniques, best practices, and visualizations, etc.

As I descended from the monastery, the rain ceased, leaving behind a delicate mist that hung like a veil over the landscape. It felt like there was only me, the mountains, and the spirit of the monastery—it was truly heaven on earth.

The list goes on for the next 5 days. The entire journey felt serendipitous yet deliberate. From the moment I landed, everything unfolded with seamless precision, as if each moment had been meticulously orchestrated to lead me to the next. “SiFi” once again, worked.

And some Bhutan notes for future visitors

Bhutan is, without a doubt, the country with the purest energy I’ve ever encountered. I was enveloped by a soul-cleansing aura of peace and simplicity for my entire stay.

• It is a place where safety is not a concern and ill will seems non-existent

• The majority of Bhutanese people speak fluent English, as it is the primary language of instruction in schools, starting from kindergarten

• Make sure to bring cash—ATMs have daily withdrawal limits, and online payment systems are not fully developed. Cash is essential for buying local foods, artifacts, and most importantly, tipping - you are gonna want to leave extra cash with everyone here for their kindness

• Hiking boots are a must, though you can purchase them locally if needed

• The national dish, chili cheese, is as simple as its name suggests. Bhutanese cuisine, along with Thai and Indian influences, dominates the local food scene

For those with a philosophical lean towards Buddhism, Bhutan would be your paradise. Exquisite temples and pilgrimage sites are everywhere with stunning architectural and spiritual impressions.

-

Before Bhutan, my favorite part of solo travel was wandering through cities with music, podcasts, or audiobooks filling my ears—a feast for both my eyes and brain. But in Bhutan, my AirPods never left the hotel. Here, I tuned in to something new & better: the voice within.

This marks the end of a major chapter and the start of a new one.

I'm excited to announce that I recently graduated from Wharton MBA and have joined Portal Ventures as an Investment Partner.

In the interest of brevity, I will summarize what led me here and the learnings from each step, in reverse chronological order.

Why did I join Portal Ventures?

There are crypto-native VCs and there are traditional VCs that invest in crypto. That said, very few truly combine the best of both worlds. Evan is a rare exception. Both I and the founders he works with have been impressed by his ability to combine his experiences from Insight Partners and Goldman Sachs with first-principles business model thinking for the crypto economy. We share the vision of accelerating the protocol economy through deep research and business best practices.

Protocol Labs gave me a strong dose of founder empathy. Part of my job is to "make it work" for Protocol Labs' startups in fundraising, BD introductions, strategy formation, etc. Having been in the shoes of founders, I have noticed that VC feedback to startups often focuses on "why it won't work" instead of "what will make it work."

Since my time at Protocol Labs, I have made a promise to myself that when I return to venture capital, I will be the most constructive and supportive VC possible, even when I don’t invest. This means being more judicious when taking up founders' airtime, so that I can take the time to produce thoughtful feedback per call. Founders do the heavy lifting, sometime the most impactful thing an investor can do is just believe.

Hivemind Capital was the best foray into venture that I could have asked for. I am eternally grateful for the platform it provided me to build my initial crypto network and to speak with and learn from over 500 crypto founders during my six months there.

Concurrent to the work experience above, Wharton MBA taught me invaluable lessons beyond the classroom.

• Organizing the Penn Blockchain Conference of 2023 was one of the best decisions I made during my MBA. Yes it was a lot of time and commitment, but in the process I got to work with & befriend some of the most brilliant undergrads, such as Sofia Wawrzyniak and Chunda McCain; strengthen the community ties of the "Penn crypto mafia" alumni; and build/deepen relationships with 110+ top industry leaders who joined us as panelists

• Being a launch manager of the Cypher Corporate Accelerator taught me the ins and outs of designing a crypto curriculum, which later became instrumental when working with Professor Werbach in designing a crypto MBA course for Wharton

• Researching at Venture Foragers taught me about collaborative research and was a critical stepping stone that led to the subsequent independent research pieces like the Metaverse Report, the Decentralized Storage Report, and other research blogposts

• Being the Social Chair of the Wharton MBA PE/VC Club & Founders’ Club in 2021-2022 was the highlight of my Wharton experience, during which I made some of my most enduring friendships.

Above all, I am filled with gratitude towards those who have helped, supported, or believed in me throughout this journey.

1. The State of BTC: Market Observations and Commentary

2. How Our Thinking Has Evolved

   • Beliefs That We Still Firmly Hold

   • Areas Where Our Opinion Has Shifted

3. Chapter Two of BTC: What Comes Next

   • Predictions

   • Whitespaces: Sectors and Verticals That Excite Us

In the interest of brevity, we will present our observations, opinions, skepticism, and predictions in bullet-point format.

The State of BTC: Observations and Commentary on the Sector

1. Fierce competition among EVM L2s/side chains with indistinguishable value propositions: The competition shifted from "who's the most trustless" to who can 1. execute the most sophisticated airdrop game, and 2. access the deepest liquidity via BTC "whales." This shift also explains the geographic concentration of L2s emerging from APAC.

2. Intensified fragmentation across

   1. Liquidity & scaling solutions: Over 80 sidechains/rollups, along with 5+ metaprotocols.

   2. Token standards: A duopoly exists between BRC20 and RUNE, followed by a long tail of ARC, CRBC, RGB, among other metaprotocols.

   3. Indexer: Each token standard necessitates its own indexer.

3. Side Chains & Rollups' Aspirations for BitVM to Trustlessly Peg Them Back to L1: I've become quite skeptical of “L2s” that initially position as a side-chain and claim to become "trustless" upon integrating with BitVM. My skepticism is two-fold:

   1. Technical Feasibility: The on-chain implementation of "optimistic rollup" style verification involves developing millions, if not billions, of logic gates. This process is not only costly because it operates on the base layer, but it's also slow due to the constraints imposed by the BTC block time. Moreover, it could take a significant amount of time to complete. To my knowledge (please correct me if I'm mistaken), BitVM has become a community project, facing the usual challenges of decentralization: no single entity is accountable for its development timeline, milestones, quality, and overall success.

   2. GTM timing: According to various sources, BitVM will not be ready for another 18-24 months. Even if everything goes perfectly for BitVM and they deliver on their promises, this means these L2s will remain centralized via multisig for the foreseeable future. What will they compete on until then?

4. A wide array of trust assumptions across L2 (side chains + rollups), and metaprotocols It might be worth clarifying the taxonomy of what constitutes a side chain vs. a rollup. Below is a table detailing the difference based on conversations we’ve had over the past few months - feedback is welcome.

• Trust assumption of side chains

  The peg between the BTC base layer and Layer 2 solutions is primarily managed centrally through a multisig controlled by the core team.

  The state and transaction finality are not verified by the BTC base layer, but instead by the project team.

• Trust assumption of ZK Rollups

  Currently, there's no way to perform zk verification on Bitcoin. Sequencers need to be centralized (similar to ETH L2s), and there's a trust assumption that the decentralized verifier network will correctly verify the transactions checked by provers.

  Rollups are limited in the number of transactions they can process due to the requirement of posting data back to Bitcoin.

• Trust assumption of BitVM (Optimistic Rollup)

  BitVM’s main use case is the promise of a trustless bridge. The high level steps are that BitVM codes can be deconstructed into logic gates for fraud provers to do binary search and find the point where execution is not accurate.

  While anyone can provide fault proof to confiscate an operator's collateral if the operator behaves maliciously, BitVM presents an economic challenge. The operator needs to match the amount of liquidity being bridged over as collateral. For instance, if I'm bridging over 10BTC via BitVM, the operator of BitVM needs to place 10BTC as collateral for this single transaction, which can be economically difficult to scale.

• Trust assumption of Meta-protocols: Indexers

  BRC20, RUNE, PIPE, Trac, etc., all require their own indexers to translate the "state" for account-based models from or to BTC's native UTXO model. Ethereum has internalized indexing because its VM can calculate state, whereas BTC's indexer is similar to Ethereum's GETH. To understand the concept of an indexer, imagine all UTXO transactions as raw Excel data where a thousand addresses have transactions with each other.

  To comprehend who owns what and the final balance (account state), indexers serve a similar purpose as a Pivot Table. They compute the additions and subtractions and determine the final balance by address. Presently, indexers like BestInSlots, GeniiData, and ALEX Labs Oracle provide APIs for developers to directly pull the balance or "account state" of meta-protocols like BRC20.

• Discrete Log Contracts (DLCs): dependency on external oracles DLC in Bitcoin relies on external oracles to determine the outcome of a contract. In a DLC, the oracle's role is to sign a message indicating the event's outcome. This signature is then used by the contracting parties to construct and broadcast the transaction that settles the contract on the Bitcoin blockchain. The security and trustworthiness of a DLC heavily depend on the reliability and honesty of the oracle, as its data directly influences the contract's resolution.

5. Ongoing skepticism among VCs in the Western Hemisphere: I've had several discussions with fellow investors about the value of BTC L2s. While we agree that most Bitcoin "L2s" are viewed as "fake" because they don't necessarily inherit BTC security and many will likely disappear in the coming months, we believe there’s still value in this vertical because\

   1. Flawed, nascent, & inevitable: Over the past 10 years, the segment of BTC holders who want to put BTC in a hard wallet and bury it in their backyard has saturated. While the layer two solutions of BTC are sub-par compared to their Ethereum counterparts in terms of security and trustlessness, this past year has only been the inception of scaling & programmability solutions on Bitcoin. The genie is out of the bottle for Bitcoin’s metamorphosis into a more general-purpose, programmable chain, in addition to its existing identity as a store of value. This trend is evidenced by the public enthusiasm for staking, trading, and the parabolic on-chain fees on Bitcoin since 2023 (as much as $40/tx in Dec). With the advent of OP_CAT, various rollups, and side chain solutions, use cases that used to be exclusive to Ethereum and Solana are now being explored on Bitcoin.

   2. Enable BTC capital efficiency: BTC remains one of the most highly valued assets among institutions.There is a need for more financial instruments built on or derived from Bitcoin. However, the design constraints of the base layer make such products difficult to implement. This highlights the necessity for second-layer programmable solutions.

   3. This trend solves BTC’s existential crisis: More programmability generates demand for Bitcoin’s block space, which leads to more fees as net-new incentives for miners to secure the network. This is especially crucial with each halving. This was discussed more in-depth in the previous article.

6. Increasingly inventive ways to utilize Bitcoin's block space

• As a method to inscribe data in both the witness block and transaction block

• As permanent on-chain data storage (which is an excellent product-market fit for luxury/collectible use cases)

Reflecting on the thesis from a year ago: how our thinking has evolved or persisted

Where we got right: The overall trend, timing, and demand relate to the issuance of digital assets on Bitcoin (Metaprotocols), programmability solutions (Layer 2s, Rollups), and efforts to increase the capital efficiency of BTC (Babylon, Lorenzo).

Where the jury is still out: The future of BTC could be either native vs. xVM. Although this is not a consensus view, we maintain that BTC should develop its own native ecosystem instead of adopting the seemingly "convenient" Ethereum Virtual Machine (EVM) approach.

While we fully acknowledge the benefits of the EVM approach (such as interoperability with existing defi ecosystem, ease to onboard solidity developers, & solidity being the most battle-tested language, etc.) , it seems contradictory to have a BTC Layer 2 token in ERC-20 format, much like Arbitrum launching its token as a Solana SPL.

Where our stance has shifted: In the original article "Panda Renaissance", two paths were proposed: making BTC more "programmable" like a general-purpose L1, or making it more "capital efficient" to extend its functionality and become a more productive financial asset. At the time, I leaned toward the latter. BTC's inherent constraints from the Script language made it unsuitable for complex programmability.

However, my perspective has shifted over the past year. The market has shown a demand for Bitcoin's block space, transforming it from a dormant "digital gold" chain to a programmable, general-purpose chain. Ethereum has now effectively become a “sandbox”: the Bitcoin community is learning from Ethereum's DeFi development over the past 7 years to innovate on Bitcoin.

Chapter Two of BTC Eco: Prediction and Whitespace

Predictions

1. There will be 1-2 winners per approach. This suggests that there would be 1-2 winners on EVM (Botanix), Metaprotocol (BRC20 or Rune), bridge-less base layer approach (Arch), ZK rollup, and STX (already a first mover). This totals 5-6 major players in the next cycle, each reaching a valuation of over $50B at their peak.

2. Flight to safety: Currently, scaling solutions are competing for attention and access to liquidity. However, over time and especially after a security-related incident, liquidity will gravitate towards the safest L2/rollups with the least trust assumption.

3. The majority of other Layer 2 solutions, sidechains, and rollups will not survive. Until that happens, it will remain a fierce competition to see who has the best airdrop program, access to high-signal investors, and connections to bitcoin whales.

4. Developers will gradually awaken from the BitVM dream upon realizing it's too late, too slow, and too costly to meet builders' need for trust minimization.

5. BTC DeFi will surpass ETH's: I've long maintained that Ethereum serves as BTC's testing ground. Everything since 2017 has been based on the mistaken belief that “BTC CAN'T”. However, that's no longer the case. If I'm taking risk with my token (staking, farming, leveraging), would I prefer my yield in low-value coins or BTC? I fully expect BTC's future to feature an equal, if not stronger, DeFi and infrastructure ecosystem compared to Ethereum's.

6. Bitcoin's native "ERC-20" token standard will emerge: Bitcoin needs its own token standard like ERC20, which can operate across various L2s and the base layer. One way to achieve this is through a collaboration between Arch (a bridge-less base layer programmability solution) and Auran (a native Bitcoin TSS-based interoperability solution). Such a standard will be a major unlock for winners to finally emerge in stablecoin, institutional adoption, on-chain yield products, and more.

Whitespace verticals that we are excited about

1. The bridge-less experience (for use cases better suited for base layer): complete trustlessness is improbable on any scaling/programmable layer - and frankly, it's not what most users care about, barring any hacks. Holding trust assumptions constant across most rollups/L2s, there will be a slew of low-velocity use cases (such as exchanges of high-value collectibles, lending, native yield, staking, etc.) where users & developers prefer to achieve the same level of programmability on the base layer (via solution like Arch Network) without needing to bridge to another L2.  The tradeoff here, of course, is speed, as transactions will be constrained by BTC block time. However, not all operations need to be exceptionally fast on-chain. Bridge-less solutions will be vital in unlocking a new segment of “slow & steady” use cases on the base layer, such as stablecoin, lending, prediction market, and so on.

2. Unifying infrastructure through state orchestration across various L2s (for use cases better suited on Layer 2): provide a "one-stop" developer/consumer experience to address the rampant liquidity and functional fragmentation in Bitcoin ecosystem today. Auran Network is a pioneer in this sector that we are excited about. “Formula for startup success: Find large highly fragmented industry w low NPS; vertically integrate a solution to simplify value product” - Keith Rabois, Founders Fund

3. Exports of BTC Liquidity into other L1s: Numerous high-profile new/existing alt L1s are exploring ways to leverage BTC's liquidity for their ecosystem growth. Solutions like the Zeus Network, which creates a messaging layer between Solana and BTC, will likely see traction.

4. Winner in stablecoin: There have been over 10 stablecoin projects. Forming deep collaborations with prevailing programmability and interoperability solutions will be the key to market dominance. We are excited to back Bima, the first omnichain Bitcoin stablecoin

5. OP CAT: OP_CAT (BIP-347) is a Bitcoin Improvement Proposal aiming to simplify and expand Bitcoin’s functionalities to enable logical loops and conditionals. It will allow the creation of rules or conditions on how Bitcoin can be spent, opening the door to many development possibilities, including Layer 2s, Smart Contracts, and more. The timeline is looking like it's 12+ months out.

6. Native Ordinals trading platform: If history repeats itself as seen with SOL and ETH, the majority of the trading volume will be dominated by a native professional trading platform like Blur or Tensor. So far, most of the trading activities for Ordinal have taken place on Magic Eden and OKX. As the NFT winter recedes, we anticipate a native Ordinal trading platform, like Ordinal Hive, to flourish.

7. Alternatives to BitVM to enable trustless, or at least trust minimize L2 <> base layer bridge

8. On-chain BTC-on-BTC yield products for institutions: STX, given its BTC-yield bearing, and regulatory-compliant traits, could be a winner here.

9. Liquid staking on bitcoin: skeuomorphic to ETH but there’s merit to LST-on-BTC projects like Lorenzo Protocol to optimize for liquidity in BTC Defi

And more - Reach out to me on X @DotCuriousCat if you have ideas that I haven't mentioned, or even better, haven't heard of. Let's build together!

Thank you Amine & Matt of Arch Network, Eril of BestinSlots, Jay of Bison Labs, and Vishal of Multicoin for your review & input.

Disclosure: Portal Ventures is an investor in Arch Network, Bison Labs, BBL Stablecoin, Auran Network, Lorenzo Protocol, Ordinal Hive, Zeus Network