At 2:14 on a Tuesday morning in Dublin, a due-diligence report landed in my inbox. Nine sections. Forty-one tables. A Howey test broken into its four elements, each with its own row. And in almost every cell that mattered — technical architecture, token supply, unlock schedule, team identity, regulatory posture, risk matrix — the same three characters, repeated with the patience of a metronome: N/A.
The project had closed a $100 million round eleven days earlier.
I have spent fifteen years reading crypto documents, and the loudest thing on the page is almost never the number. It is the blank. In the chaos of a bull market, we found our winter soul: an industry that has learned to dress emptiness in the typography of rigor — bold headers, clean tables, professional disclaimers, all of it scaffolding erected carefully around an absence.
I want to be precise about what that report was, because it would be easy to mock it and miss the point. It was not lazy. It was the output of a two-stage analysis pipeline: first, a parser extracts discrete, citable information points from source material; second, an analyst layer reasons over those points across nine dimensions — technology, tokenomics, market structure, ecosystem position, regulation, team and governance, risk, narrative, and supply-chain transmission. When the first stage returns nothing, the second stage has exactly two honest options: print N/A in every cell, or invent.
This report chose N/A. Almost none of the ones I read in 2026 do.
That choice is the story. Not the $100 million. Not the eleven days. The fact that a document arrived on my desk containing zero verifiable claims — no contract address, no audit link, no unlock table, no sequencer specification, no cross-chain verification configuration — and that this document was the entire evidentiary basis for a nine-figure allocation. The parser did not fail because it was badly built. It failed because the source material made no commitments. A document that cannot be parsed into claims is a document that has made no promises.
The industrialization of research is what got us here, and I want to name the mechanism plainly. The cost of producing a plausible report has fallen to near zero. A landing page, an X thread, a sponsored "research note," a language model prompted toward optimism, and you have something that looks like diligence from three feet away. The cost of producing a true report has not fallen at all. It still requires reading the contract, emailing the team, checking the unlock vesting against the on-chain distribution, and sitting with the discomfort of not knowing. When one cost collapses and the other does not, the market does not split the difference. It buys the cheap one and calls it coverage.
So here is the first thing the empty template taught me, and it is the thing I keep failing to teach others: the N/A is not a verdict on the analyst. It is a measurement of the subject. Twelve missing fields is not twelve neutral facts. It is a fingerprint.
Now the part that should worry anyone holding a bag this cycle. In any scoring rubric, N/A must eventually be mapped to a number — and the default mapping is zero. Zero risk. Zero weight. A project with twelve unknowns ends up scoring identically to a project with twelve verified, benign findings. Worse than identically: the project that discloses twelve genuine weaknesses gets flagged twelve times, while the project that discloses nothing gets a clean sheet. Transparency is penalized and opacity is rewarded, structurally, by the arithmetic of due diligence itself.
Watch how this plays out in practice, because it is not abstract. Take two rollups with the same architecture — same centralized sequencer, same upgrade multisig, same blob-posting cadence. The first publishes its sequencer address, its multisig signer set, and its batch submission history. An honest analysis reads that and writes "centralization risk: high." Correct. The second publishes a whitepaper and a vision statement. The same analysis reads nothing and writes "centralization risk: N/A." The second rollup, functionally identical and materially more opaque, walks out of the room with a cleaner file. The difference between them was never safety. It was disclosure. And the market has been quietly pricing disclosure as if it were danger.
This is where the blank hides in 2026, and I can point to the exact seats. In oracle design, a protocol that refuses to publish its heartbeat and deviation thresholds cannot be assessed for feed latency — which means the single most common failure mode in lending markets is filed under "insufficient information" and then forgotten. In rollup economics, post-Dencun fee models were built on cheap blob space; the saturation curve is a computable, disclosed fact, and yet almost nobody puts a blob-cost sensitivity table in their documentation. The number that will double their users' gas fees is available. It is simply not written down, and unwritten numbers do not get audited. In cross-chain messaging, when a team declines to publish its verifier set — which oracle, which relayer, which threshold, who can change them and under what quorum — the claim of "decentralized interoperability" becomes unfalsifiable. Unfalsifiable claims are the native habitat of N/A. They cannot be wrong because they cannot be tested.
I learned this the hard way at twenty-two, auditing a decentralized exchange called EtherSwap during the 2017 frenzy. Six weeks of reading. My peers were chasing allocations; I was tracing vote weights. What I found was not a hidden exploit — it was an undocumented one. The voting mechanism allowed whale wallets to route around consensus, and the reason nobody had caught it was that nobody had written the rule down. The flaw was never secret. It was merely unwritten. I published four thousand words instead of buying the token, and the lesson calcified: the vulnerability is rarely hidden; it is usually just unrecorded. Six years on from DeFi Summer, I keep returning to a line I wrote then — in the chaos of summer, we found our winter soul — because it describes exactly this: the moment a market's enthusiasm outruns its documentation, the documentation is what winter will audit.
By 2025 the blanks had learned to vote. At GovernAI, automated agents began steering proposal outcomes under the banner of efficiency, and when I went looking for the metric that justified them, there was none. The efficiency claim was itself unaudited — an N/A dressed as a number. Fifteen of us fought for a human-in-the-loop charter and won, and what we actually won was the right to write the rule down. The year before, at CivicChain, we ran a quadratic voting pilot across ten thousand participants and lifted non-whale participation forty percent. That worked for one reason: we published the weights. Everyone could see how voice was converted into influence. Disclosure was not a feature of the design. It was the design.
So who pays for the N/A? Nobody. Research in this industry is a loss leader for allocation, and in a bull market the buyer of research is a person who wants a reason to buy. The market clears at the optimistic report because that is the report with a customer. This is not corruption, and I resist framing it that way; corruption is easier to fix. This is incentive, which is load-bearing.
And yet I will not let the honest blank off the hook either, because I have watched it become an alibi. "Insufficient information" is the only conclusion that can never be wrong, and there is a version of rigor that is really just a refusal to stick your neck out. A report that says nothing is not automatically brave. The pragmatism test is simple: an N/A must be converted into falsifiable questions with owners and deadlines. Three of them would have settled the $100 million file in a day. Publish the contract address and name the holder of the admin key. Publish the unlock table with cliff dates and the wallets that will receive them. Publish the sequencer operator and the cross-chain verifier set, and say who can change them. Each is answerable within twenty-four hours by anyone who genuinely possesses the code. If a team cannot answer in twenty-four hours, that is not a gap in the analysis. That is the analysis. Silence in the bear market is where truth compiles; in a bull market, silence is a financing strategy.
Governance is not a vote, it is a vigil — and so is research. Both are practices of sustained attention performed when nobody is watching and no reward is immediate. The next cycle will not be decided by who ships the best deck. It will be decided by whose claims survive a parser, whose numbers are written down before they are needed, and whose blank spaces were treated as warnings rather than as clean sheets. Code is law, but conscience is the compiler, and the compiler is where we decide what counts as evidence. We do not build walls, we weave nets of trust — and a net with holes in it is not a net, it is a story we tell about a net.
When the next nine-figure round lands on your desk with a blank where the risk section should be, you will have a choice. You can read the round. Or you can read the blank. Only one of them is telling you the truth.