Strategy · Business Information · 2026
The incumbents sell public data back to the market at license prices. We take the entire public-derived layer — on free data, with an AI-native entity graph — and leave only the true moat standing.
The thesis
Strip a Bloomberg, FactSet, Capital IQ or Morningstar down to its content and most of it is SEC and regulatory filings, re-keyed and rented back — ownership, insiders, M&A, fundamentals, people, boards, comp, governance, events. That layer isn't a moat. It's extraction. We already run it, and we can take the rest of it for the cost of engineering, not a license.
We mapped 153 companies across the industry and sorted every data domain into four tiers. The point isn't to copy a terminal — it's to see exactly how far free data reaches, and where the real wall is.
The attack surface
Not vaporware — already live
This isn't a pitch deck for data we wish we had. The Tier-0 layer is in production today, entity-resolved on one graph, with an AI chat that answers grounded and cited.
The next wave
Full ranked backlog: 19 Tier-1 datasets, each with its free source and the exact point where the license wall begins.
The honesty that makes it credible
A disruptor that claims everything is noise. We name what we won't chase — and that discipline is the strategy. We take the ~80% of the value that's public-derived at near-zero marginal data cost, and we don't burn capital fighting for the genuinely proprietary last mile.
The disruption
Their model is a data license — recurring cost, per seat, per feed. Ours is extraction + an entity graph + AI agents: build once, near-zero marginal data cost, and every new free source compounds onto the same graph.
Match the public-derived layer that is most of what they sell, at a fraction of the price, and the pricing power on that layer collapses. That's the shock wave.
The filings were always public.
We just made them answerable.