Some deals would leave both sides better off, and still they never happen. To prove that what you have is worth paying for, you usually have to show it. And the moment you've shown it, the other side no longer needs to buy it.
This is one of the oldest unsolved problems in business, and it's enormous. Most of the data companies collect is never used. The large majority of patents sit idle. On top of that is a mountain of research and spare capacity nobody can put to work.
The Two Problems
You can't safely look. To ask the world "who has what I need?", you have to describe what you're working on, and that description is often the exact thing you can't afford to reveal. Your question is your position, and giving it up gives up your leverage. A biotech publicly hunting for one specific patent is announcing where its program is stuck. So the smart move is not to look.
You can't safely reveal. Even when two sides find each other, proving your side means showing it, and showing it removes the reason to pay.
This splits the world into two kinds of company. Those willing to announce what they're looking for (a directional search), and those holding assets but announcing nothing (the double-blind). The double-blind aren't searching for anything in particular. They're just sitting on it.
The Ideal A2A Infrastructure
No one has legitimately solved this yet, and the category is broad. Assume that a lot of companies in the future will run company brains: an agent, or a consortium of agents, that manages context across every dimension of the business. Granting that, it's reasonable to want a network where these agents can talk to each other. Thinking about it a little longer reveals a stack of secondary problems:
- How do you control the IP that goes into the agent? (solvable)
- How do you control the IP that comes out of the agent? (solvable?)
- How do you protect against prompt injection and the like?
- How do you make the agent act faithfully in what it does?
Two Kinds of Search
Public search: making the world legible
We read the public trail every business already leaves (what it builds, hires for, files, ships, publishes) and turn it into a clear map: who likely holds what idle asset, and who's likely stuck on what. Because it's inference from public information, this map covers companies that never signed up for anything. Think of it as an outward-looking agent watching the entire market on your behalf.
Secret search: acting on what only you know
Once the public map is good, a user takes a position they'd never announce and matches it privately against that map, using our tooling, on their own side. The tool points them at the counterparties their private position should target. Then they reach out directly.
The thing being protected here is the user's own question, and we protect it the only way anyone will actually accept: by never holding it.
Why Every Existing Fix Falls Short
- The human broker had to be told everything, which makes them the leak, and could only juggle a few relationships at once.
- Data marketplaces (Snowflake, Databricks, AWS Data Exchange) only work if the owner lists what they have, and listing is the leak. Useless for anything sensitive.
- Data clean rooms let two parties compute over data without exposing raw records, but only after they've already found each other and agreed to work together.
- Patent and IP marketplaces are public-listing models where the good stuff never gets posted, for the same disclosure reason.
- Expert networks (GLG, Tegus) match knowledge to need using people, which is bandwidth-limited, and the network sees everything that passes through.
- The "trust our secret-protecting box" model relocates the problem onto a vendor nobody has a reason to trust with their most sensitive position.
The real incumbent is the boutique merchant bank, which does exactly this with humans who must be told the secret and who scale one hire at a time.
The Target
Stuck on purpose is not our market. Sometimes one side deliberately keeps something hidden because the other side not knowing is the whole advantage. There's no honest way to unlock those; unlocking them destroys the value.
Stuck for a fixable reason is. Both sides would win, and the only thing in the way is that finding and proving the deal would force an exposure neither can afford. It's defined by the absence of any reason to keep the thing secret.
Where the stuck things live: data (companies sit on huge piles they never touch; surveys put dark, never-used company data around 55–68%), patents and inventions (an estimated 75–95% sit idle; only about 5–7% ever become a product), research and know-how (labs quietly re-derive what other labs already have), and spare capacity (idle compute, unused factory lines, money on a balance sheet).
These take three shapes, and they line up neatly with the two searches:
- A dead asset meets a live need, and an urgent need hunts for an asset it can't name. In both, at least one side leaves a public shadow. Public search plus the user's private action closes these.
- Two half-holders who don't know each other exist, where the deciding fact is hidden on both sides and leaves no public trace.
The Build
- The public-search engine. A model of every business built from its public exhaust, turned into a live, legible map of idle holdings and latent needs: an inferred graph of the whole market.
- The private overlay. Tooling that lets a user match their own confidential position against that map, on their side, and act.
- Judgment about which matches are worth acting on.
The Economics
S = V − c − E
do the deal iff S > 0
- S
- surplus created
- V
- the asset's value to buyer B
- c
- seller A's cost to transfer it
- E
- exclusivity: what A loses by no longer being the only holder (0 for a dead asset; huge for leverage assets, which pushes S negative)
Why it doesn't happen today: discovering that S > 0 means exposing yourself first. A party searches only if p × (their share of S) − L > 0, where L is the leak cost of revealing your position and p is the odds a match exists. L is big, so nobody looks. We kill L by searching over public inference instead of deposited secrets.
Money: take a fee f out of the surplus, with f < S. Surface a match only when S > f > 0.
Beachhead: dormant assets, where E ≈ 0, so S ≈ V − c. Not every surfaced match is a good deal. A deal is only worth doing when what the buyer gains beats two things combined: what it costs the seller to hand it over, plus what the seller loses by no longer being the only one who has it. If being the only holder is the seller's whole edge, that last number is huge and the deal should stay dead. Our tooling helps estimate this; the user, holding their private information, makes the call.
How we make money: a slice of the value we create, taken when a deal that otherwise wouldn't have happened actually closes.
The Steps
- Do it by hand. Personally broker deals where we already know a mismatch exists.
- Start with young, fast startups. Easy to reach, friendly to us, little reason to keep it secret. Their idle data is a clean first supply; the buyers are other startups with urgent, specific needs.
- Seed with dead assets. Things where what the seller loses by sharing is basically zero, making them the safest to surface and the easiest to price.
- Chase a few marquee deals for proof.
The Risks
- Is public inference actually good enough? The whole thing rests on our ability to read the public trail and form specific, correct guesses about idle holdings and hidden needs.
- The truly trace-less deal stays out of reach. We do unlock double-blind search, but only as far as public search can generate a hunch about who to check.
- "See who has what" invites comparison to ordinary sales-intelligence tools. Our defense: we surface idle assets and latent needs those tools can't see, and the money is tied to closed deals, not to access.
- We could misread how this market matures and build the wrong layer for it. We guard against that by staying hands-on and learning inductively until the real shape of demand is unmistakable.
Who We Are
Ryaan. Cornell (computer science, economics, philosophy). Built several software businesses to real revenue in school; cofounded a slide-deck automation company and Cornell's entrepreneurship lab, with genuine acquisition offers. Dropped out for the residency in San Francisco.
Advikar. Harvard (computer science, statistics). Published research in optimization theory at a top machine-learning conference. Founding engineer and first employee at a venture-backed startup; shipped production systems used by major global institutions.
Vinay Rao. Early advisor, from Anthropic.
Our research bench, a group of international olympiad medalists, is pointed at the two hard problems: reading the open world well enough to make it truly legible, and building the sealed comparison that unlocks the double-blind case without anyone handing over a secret. Making the world legible is the wedge. Letting two parties confirm a deal that neither will reveal, and that no one else can even see, is the moat.
The figures above come from public research and industry surveys (IBM, Splunk, Seagate, WIPO, and academic work on patents and dark data). The underlying ideas trace to Kenneth Arrow on the information paradox, Alvin Roth on what makes a market work, and the broader economics of information. A fuller, sourced version lives in our technical memo.