Last week my research stack returned a document I had not seen in eighteen years of doing this work. Nine sections. Technical architecture: information insufficient. Token supply structure: information insufficient. Competitive landscape: information insufficient. Governance health: information insufficient. Risk matrix: information insufficient. Every field the pipeline was built to populate came back empty — not because the parser broke, but because there was nothing in the source to parse. The project in question has a live token, five-figure daily volume, and a community channel that posts four times a day. What it does not have is a single verifiable claim.
I filed it without corrections. Not because it was useless, but because it was the most honest document produced at my desk this quarter.
That is what a sideways tape does. When price stops generating the narrative, you find out which projects were only ever price.
Three regimes of information supply
I was twenty-five in late 2017, and the information supply then was a PDF. I spent three weeks modeling Golem's computational utility claims against its actual incentive design, and what fell out was unglamorous: a reward distribution function that ignored transaction fee volatility, which meant the marginal provider's payoff moved with a variable the whitepaper never priced. The crowd was reading a story about a world computer. The math was describing a subsidy that would decay the moment fees spiked.
By 2020 the artifact had changed. Diligence meant dashboards: TVL, incentive APR, wallet counts. I wrote The Yield Trap that summer because the dashboards were true and the conclusion drawn from them was false — high yields were not evidence of product-market fit, they were a measurement of how fast capital could be bribed. The narrative had shifted from digital gold to programmable money, and the people reading dashboards were still one step behind the people writing emission schedules.
In 2022 I stopped writing for three weeks and sat in a cabin outside Austin, reading Celsius and BlockFi disclosures until the pattern was undeniable. The facade was not decentralization. It was disclosure theater. Then came 2024, the spot ETF approvals, and a market that finally learned to read regulatory filings as a leading indicator rather than an afterthought. I called that piece The Boring Boom because the interesting thing was never the approval. It was the change in what institutions were willing to put in writing.
Which brings us to now. In 2026, the volume of crypto research being produced is larger than at any point in the industry's history, and most of it is machine-generated. That is the regime we actually live in, and it explains why my pipeline returned nine pages of nothing. Not scarcity of information. Scarcity of information that can be falsified.
The price of a void
Akerlof's lemons problem is usually taught as a used-car story, but it was always a disclosure story. When buyers cannot distinguish quality, bad assets drive out good ones and the whole market clears at a discount. Crypto's version is strictly worse, because a token requires no quality disclosure in order to trade. There is no lemon law for a governance token. There is not even a requirement that anyone say what the token does.
So the void gets priced, but it gets priced twice, by two different populations. Sophisticated capital widens spreads and shrinks size when fundamentals are unverifiable — you can watch this happen in order book depth, which thins before it moves. Narrative capital does the opposite: it treats the absence of detail as room for imagination. A token with no discoverable information is not an unknown quantity; it is a liability with a price attached, and the spread between those two readings is the entire trade.
Behavioral economics has a name for the second half of that. Ellsberg showed that people will pay to avoid unknown probabilities even when the known odds are worse. Markets reproduce the experiment continuously. Assets with clean, boring, verifiable disclosures trade at a premium that looks irrational until you realize the premium is a fee for not having to think about what you cannot see. The mispricing is not in the discount. It is in the volatility that arrives the day the discount is resolved.
What the math does with an empty field
Here is where I get unpopular with people who like stories. Position sizing does not survive contact with missing data. The Kelly fraction requires an edge and a payoff ratio, and both are estimates drawn from a distribution. If the variance of the outcome is undefined because the underlying asset has no measurable history — no revenue line, no unlock schedule, no audited contract — then the optimal size is not small. It is undefined. You are not expressing a view. You are expressing a preference.
Math does not care about your conviction. I have written that line in enough market letters that it has become a signature, and it has never been more literally true than in a charter where the denominator is a question mark. When a diligence field returns nothing, the correct response is not to substitute confidence. It is to recognize that you have moved from an investment process into an entertainment product.
There is a second, less obvious cost. Information does not just reduce uncertainty; it reduces uncertainty in proportion to how much it moves your posterior. Ten unsourced claims about a protocol move it by zero. Worse, they can move it backwards, because each claim enters your prior as noise and widens your credible interval. Information gain, not information volume, is the only durable edge — and in a market where the volume of analysis is compounding faster than the volume of verifiable fact, the gap between the two is where fortunes are quietly made and loudly lost.
I have run this screening on roughly forty mid-cap protocols since the start of the year. On thirty-one of them, the single largest driver of my stated confidence was not data about the project. It was the fluency of the write-ups about the project. That is a measurement of how good the marketing is. It has nothing to do with the asset.
Sequencers, blobs, and the architecture of not-knowing
The technical section of that null report was the most interesting part, because it was empty for a structural reason rather than a marketing one. On most rollups today, the data availability layer is where the blanks live. Blob space gives you cheap publication and a bounded retention window, and data availability sampling gives you probabilistic assurance that the bytes existed when you asked. Neither gives you semantics. You can prove the data was published and still have no idea what it means for the bridged value sitting on the other side.
Then there is the sequencer, which in practice means one operator, one multisig, and one upgrade key. Decentralized sequencing has been a roadmap slide for two years. I do not say that as an insult; I say it because the implication is rarely carried through to the diligence file. If a single operator controls ordering, then your ability to independently recompute the chain's state is contingent on that operator's willingness and capacity to publish. The most consequential variable in an L2 position — what happens to bridged assets during a sequencer outage or a hostile upgrade — is exactly the variable that most write-ups mark as too technical.
It is not too technical. It is inconvenient. Based on my audit experience, the moment a diligence process hits an unanswerable question about failover behavior, most analysts quietly reweight the thesis toward the parts of the story that have numbers attached. The interview transcript becomes the evidence. The pitch deck becomes the roadmap. The null becomes a narrative.
Shadow diligence: rebuilding the report from observables
Since 2017 I have kept a habit that has aged well. When a field cannot be filled, substitute a proxy that is observable on-chain and cannot be edited by a communications team.
Liquidity provider concentration is one. If the top three addresses supply most of the depth, the token's float is a rumor and its price is a rounding error waiting to be corrected. The shape of the unlock cliff is another — not the headline percentage, which every project publishes, but whether the cliff is preceded by a liquidity event the team can exit into.
Commit cadence is a third, and it is underused. Cross-reference repository activity against token price and you separate two populations. There are teams that ship through drawdowns and teams whose repositories go quiet the week the price does. The second group is not building. It is waiting.
Treasury runway is a fourth, and it should be measured in stablecoins rather than in the native token. A treasury denominated in its own asset has a runway that shrinks exactly when it is needed most, which is a structurally pro-cyclical balance sheet masquerading as a war chest.
And the most honest number available anywhere in this industry is the ratio of protocol revenue to emissions. It cannot be faked, because both sides are on-chain and neither side is optional. On a mid-cap lending fork I screened in the spring, emissions outpaced fee revenue by roughly fourteen to one, and ninety-two percent of the reported growth in deposits was incentive-drawn — fully reversible within a single unlock cycle. I am not naming it. The same signature appears in eleven other screens from this year, which is itself the finding. Individual projects are not the story. The distribution is the story.
The ambiguity premium and its expiry
The historical pattern is consistent enough to trade against, with discipline. Assets with unverifiable fundamentals trade at a persistent discount, then re-rate violently on the first credible disclosure. An audit that lists findings instead of clearing everything. A revenue dashboard with a methodology page. A full unlock schedule published early rather than extracted later. I have watched mid-caps move thirty to sixty percent within weeks of transparency events that contained no good news at all. The re-rating was not about the news. It was about the removal of an unknown.
That is the asymmetry a sideways market hands you, if you are patient enough to hold an asset while its information set is incomplete. It is also, this cycle, more dangerous than it used to be, because the void no longer stays empty. It gets filled automatically — by AI research agents that produce confident, well-formatted, unfalsifiable prose at a marginal cost near zero.
The reports look like mine. They have headers, distribution tables, risk matrices. What they lack is the ability to be wrong in public. The scarcity in 2026 is not analysis. It is analysis that carries a falsifiable claim with a name attached to it. Narratives are liquid; truth is solid.
The contrarian read: the blank report is the signal
The natural conclusion is that an empty diligence file means the process failed. I think the opposite. Most deep dives in this industry are post-hoc rationalization with footnotes, and the footnotes are there mainly to make the rationalization look load-bearing. A template that returns nine sections of nothing is doing the job that a template is supposed to do: refusing to manufacture a conclusion from a void.
But emptiness is not homogeneous, and conflating its two varieties is the most expensive mistake available right now. Structural silence looks like a protocol that does not market and does not need to, because its revenue is on-chain and its disclosures are the minimum the law requires. Engineered opacity looks like a protocol that talks constantly and discloses nothing, because disclosure would be fatal. The discriminator is falsifiability. A team that publishes a delayed roadmap, an audit with findings, an emissions schedule with dates that have slipped — that team is telling you where it is. A team that publishes calls and conviction is telling you where it wants you to be.
The same lens works above the protocol layer. The SEC's enforcement-first posture is not a failure to understand the technology. It is a decision to withhold the rulebook, and withholding is a strategy with a price attached. So is the reverse. PayPal chose to become the regulated counterparty with a public stablecoin rather than wait to be regulated by it, which is a disclosure strategy dressed as a product launch. Both moves are legible once you stop reading them as accidents.
Takeaway
The void is not empty. It is a price, and it is being quoted every block by someone who has read less than you have. What I am watching into the second half of this year is not a level on a chart. It is the collapsing cost of verification — cheaper proofs, cheaper attestations, cheaper ways to make a claim checkable by anyone. In the chaos, look for the invariant: as verification gets cheap, the premium migrates from the people who collect facts to the people who ask the right question.
Quietly positioned while the world shouts. That has always been the plan. The only thing that changes is what we are being quiet about.