Last week, a crypto-native voice with a decade of on-chain credibility said something no price chart had managed to say. A tenfold increase in personal net worth, he argued, is now easier through AI equities than through tokens.
There was no exploit behind the statement. No depeg, no unlock cliff, no governance coup. Just a sentence. And it carried more structural information than most of the technical threads I read that week, because it was a statement about where a class of capital intends to deploy next, not about where prices have already been.
I have spent thirteen years watching this industry price things it could not verify. The most expensive of those prices were never denominated in dollars. They were denominated in attention.
Eugene Ng Ah Sio is not an anonymous account. He is a known figure in crypto-native circles with a track record in on-chain trading and market commentary, and a footprint in mainstream financial media. That matters, because the weight of a claim scales with the credibility of the speaker's cost of being wrong.
His position, compressed: over six months he studied AI equities systematically. AI is still early. The equity market presents a greater structural challenge than crypto, yet the tenfold outcome is more reachable there. AI semiconductors have fresh upward momentum. Anthropic's IPO and its subsequent model releases are the catalysts to watch, on a three-year arc toward superintelligence.
Notice the shape. Seven statements, zero specifications. No architecture, no valuation model, no disclosed position, no falsifiable intermediate milestone. It is opinion in its purest form. In a world of noise, code is the only quiet truth. An opinion carries no code, which is precisely why it travels further than an audit.
Place it against the tape. The market is sideways. Ranges compress, narratives thin, and capital sits in money-market equivalents waiting for a direction no single candle will supply. In that environment, opinions from credible operators function less as predictions than as reconnaissance. They show where someone with skin in the game is looking while nothing moves.
Attention has an emission schedule.
Here is the frame I keep returning to, and the one I have not seen applied to this story.
Attention behaves like a monetary base. It has a supply, a distribution schedule, and holders with different cost bases. Its supply expands or contracts according to the participants who control issuance.
Crypto spent a decade issuing attention against a promise of future utility. That issuance ran in three phases — infrastructure, then DeFi, then speculation. Each phase expanded the float and diluted the cohort before it. At no point in the last three years did the schedule converge toward anything a rational holder would call a yield curve.
AI equities run the opposite schedule. The supply of tradable attention is constrained by the number of listed companies with audited financials and legally enforceable claims on earnings. That constraint is not a flaw in the design. It is the design. In a world of noise, code is the only quiet truth — and a 10-K filing is code with a legal executor.
So when a crypto-native says 10x is easier there, he is not comparing technologies. He is comparing issuance schedules. One is fully diluted. The other is still in its low-float era.
Value capture, and why the measuring stick is bent.
Equity is a residual claim on cash flow enforced by courts. A token is a claim on an emission schedule enforced by a contract that the deploying multisig can amend, or by a governance vote most holders never read. These are not one asset class in different clothes. They are two different economic machines.
I have written about this asymmetry before, in a narrower context. Between 2020 and 2022 I ran post-mortems on three collapsed protocols and calculated that their burn rates were mathematically unsustainable within six months. The arithmetic was trivial. The hard part was convincing holders that the yield they were earning was a transfer, not a return. When I advised my network to move 60% into stablecoins, the pushback was not analytical. It was tribal.
That same reflex now runs in reverse. A comparison between the two systems gets read as betrayal rather than as measurement.
And the measuring apparatus inside DeFi is itself unsound. The interest rate models Aave and Compound run on are governance parameters dressed as market discovery — utilization curves set by proposal, not by order flow. When your reference rate is an administrative setting, your risk-adjusted return calculation inherits the error. A comparison between a governed number and a discovered one is not evidence. It is an artifact of the comparison itself.
The IPO is a closure event.
Anthropic matters more than the AI thesis around it, and for a structural reason.
An IPO is a closure event. It converts a private, illiquid, narrative-priced asset into a public, liquid, disclosure-priced one. The primary market's duration risk resolves in a single afternoon. That — not the model releases, which follow a public and predictable cadence — is what makes it a catalyst.
Crypto has almost no closure events. Tokens launch liquid and stay liquid. There is no moment when the illiquidity premium is realized and the market is forced to mark the asset against audited reality. Capital that wants a terminal event keeps choosing venues that provide one.
The reflex is to point at RWA and tokenized treasuries as the answer. I would flag caution. Tokenization routes a regulated asset through a wrapper; it does not import the closure mechanism. If the wrapper is permissioned, the buyer is already inside the legacy system, and the token is a settlement improvement rather than a new asset class.
There is a quieter reason the compliance premium compounds: identity. Soulbound tokens have been a workable concept for roughly three years and have produced almost nothing, for a reason the ecosystem avoids stating plainly. No one wants a permanent, public, transfer-resistant credit record. That refusal is rational and near-universal. Because on-chain identity cannot be built without it, every function depending on verified identity — undercollateralized credit, institutional settlement, regulated securities — defaults to legacy rails. The premium accrues there by elimination.
Distribution beats cryptography.
Where does that leave capital already inside crypto? My read is that it consolidates around distribution, not cryptography. The real difference between OP Stack and ZK Stack was never technical; the architectures converged faster than the marketing admitted. The difference is which team convinces more projects to deploy first. Distribution is the moat. Cryptography is the cost of entry.
AI equities win the attention comparison for the same non-technical reason. Their distribution rail is a brokerage account with an app that loads in under a second. Crypto's distribution rail is a seed phrase, and most of the population correctly assesses that as a tax they will not pay.
This is where the red flag belongs — not in the AI thesis, but in the assumption that the migration is a price event. It is a deposit event. Watch exchange stablecoin reserves, not the candle. If capital is rotating, it shows up in what people are willing to leave on a centralized venue with a known counterparty. Over a two-to-four week window that is observable. The price of Bitcoin over the same window tells you almost nothing.
I will now take the other side of my own argument.
The framing that AI is draining crypto is a convenience for both camps. It gives crypto a reason to be weak that requires admitting nothing structural, and it gives AI an inflow story it has not earned. Capital totals are not fixed. The same allocator can hold both, and most do.
Second, the AI complex has its own issuance schedule, and it is not clean. Semiconductor earnings are priced against a supply chain running through a handful of fabrication facilities and a memory market with physical constraints. When narrative and capacity diverge, the correction is not a chop; it is a repricing. An opinion with no falsifiable timestamp will be validated if the next two quarters are good and quietly archived if they are not. I cannot verify whether the author holds AI equities, and there is no mechanism by which I could. Weight that absence accordingly.
Third, and this is what the crypto side keeps missing: the sentiment is not proof that crypto's opportunity set shrank. It is evidence that the opportunity set was mis-sold. The tenfold claim was always a statement about distribution mechanics, not technology. The protocols that survived 2022 did not survive because their narratives were superior. They survived because their emissions were finite. In a world of noise, code is the only quiet truth, and no balance sheet has ever been audited by conviction.
The next signal will not come from a KOL. It will come from a denominator: the ratio of stablecoin balances held on centralized venues to total stablecoin supply, measured weekly, plotted against margin debt at retail brokerages. If that ratio keeps falling while equity margin rises, the rotation is real. If it does not, the entire conversation was a mood with a timestamp nobody recorded. Verify the number, then form the opinion.
