WillItSurvive
YC Spring 2026 · United States

Cohesion

AI agents that automate equity research for institutional investors using public data.

Losing its protection
Assessed 27 July 2026 · Evidence as of 27 July 2026
The structural protection is eroding.
On the recordSince 27 July 2026 · Checkable from mid-June 2026 · Every call on the record Checkpoint: YC Demo Day and likely seed round close
The call

Cohesion sells a thinner version of what AlphaSense and Bloomberg already ship inside subscriptions funds already hold.

Cohesion built software that uses artificial intelligence to search through company earnings reports, financial filings, and public conversations to help stock market researchers find information faster. The problem is that AlphaSense, a much larger company with five hundred million dollars in yearly revenue, already does almost exactly this and also owns exclusive interviews with industry experts that Cohesion cannot access. Bloomberg, the financial news giant, also just launched a similar tool, and at least five other startups launched the same idea at the same time. What could change this picture is if Cohesion focuses on one specific type of client or one specific type of research question instead of trying to be useful to everyone, or if major investment firms decide to bet heavily on the company after its June 2026 presentation.

Why
01
AI pressure: AlphaSense's Generative Search (multi-agent AI over earnings, filings, transcripts). AlphaSense already runs the same multi-agent search Cohesion pitches, for funds already paying its seat licenses. Could AI do the core job? Yes, likely.
02
Bigger players: Bloomberg (ASKB agentic interface) and FactSet/PitchBook MCP servers. Bloomberg and FactSet bolted agent layers onto terminals funds already pay for, closing the gap Cohesion is trying to fill. Could a giant copy it? Yes, easily.
03
What protects it. A dozen funds already running it plus the founders' hedge fund and T. Rowe Price contacts, nothing locked in yet.
What would change this call
Cohesion announces a Seed/Series A round after its June 2026 Demo Day. Shows whether buy-side LPs or generic VCs back its vertical bet.
Cohesion narrows to a specific vertical (e.g., biotech, semis) by Q4 2026. Indicates it found a defensible niche AlphaSense won't chase.
AlphaSense, Bloomberg, or FactSet ship new agentic features within 6 months. Confirms incumbents can copy Cohesion's core function quickly.
Key numbers
  • AlphaSense reached ~$500M ARR (Oct 2025)
  • AlphaSense bought Tegus for ~$930M (n.d.)
  • Cohesion claims 10+ funds, $10Bn+ combined AUM (n.d.)
Named risks
  • No proprietary data, integration, or regulatory moat; product replicable by better-funded incumbents. Severity: high
  • Crowded YC batch with five or more near-identical competitors chasing the same institutional buyers. Severity: high
  • No disclosed funding round or investor beyond YC accelerator investment as of July 2026. Severity: medium
  • No disclosed pricing model, leaving revenue path and unit economics unclear. Severity: medium
Rivals, as read in July 2026
RivalDirectionEvidence
AlphaSenselosing toAlphaSense at ~$500M ARR (Oct 2025), Tegus data moat, shipped agentic search Jan 2026
Bloomberglosing toBloomberg's agentic ASKB research interface live by April 2026
Hebbialosing toHebbia claims automating up to 90% of diligence document synthesis, 2026
WithAI, KelAI, Soria, Kimpton AI, Vantedge AIcontestingFive near-identical YC Spring 2026 peers targeting same institutional buyers, Apr to Jul 2026
Questions people ask
What is Cohesion's biggest risk?
The biggest risk is that AlphaSense and Bloomberg already ship the same agentic research tools Cohesion is pitching, inside subscriptions institutional funds already pay for. AlphaSense alone runs at roughly $500M in annual recurring revenue as of October 2025 and claims about 88% penetration of the S&P 100, leaving little room for a newer entrant without a distinct niche. With no proprietary data or technology lock-in, there is almost no technical barrier stopping incumbents from copying its approach.
Is Cohesion in financial trouble?
There is no direct evidence of financial trouble, but Cohesion has not disclosed a seed or Series A round with named outside investors as of the July 2026 assessment. It claims 10+ fund clients with combined AUM over $10 billion, which is a real but modest customer base compared to established players like AlphaSense. Watch for whether it announces funding by the end of 2026 as a signal of financial footing.
Who are Cohesion's biggest competitors?
Its main competitors are AlphaSense and Bloomberg's ASKB, both of which dominate institutional equity research and shipped their own agentic search tools in 2026. Hebbia adds pressure with diligence automation claims of up to 90% document synthesis, and at least five near-identical YC Spring 2026 peers, including WithAI, KelAI, Soria, Kimpton AI, and Vantedge AI, are chasing the same institutional buyers.
Is Cohesion at risk from AI?
Yes, and the risk comes from AI incumbents rather than being replaced by AI itself. AlphaSense's Generative Search already runs the same multi-agent search Cohesion pitches, shipped in 2026, for funds already paying its seat licenses, making it easy for those incumbents to absorb Cohesion's pitch without needing a new vendor.
Does Cohesion have any competitive edge?
Very little. Cohesion has no proprietary data or technology to keep competitors out, and its position is not backed by exclusive contracts. Its main shield is a dozen funds already using it plus the founders' hedge fund and T. Rowe Price contacts, but none of that is contractually locked in yet, and AlphaSense and Bloomberg already offer comparable tools inside subscriptions funds already pay for.
What should I watch to see if Cohesion is gaining traction?
Watch for two things by the end of 2026: whether Cohesion announces a seed or Series A round with named outside investors, and whether it narrows from broad public-equities research into a specific vertical niche by Q4 2026. The YC Demo Day and a likely seed round close around mid-June 2026 is an earlier checkpoint worth tracking.
Timeline
  • 2026-07-14YC lists Cohesion among asset management startups; peer teams sized around 3 employees
  • 2026-07-07Coverage describes Cohesion's agentic equity research product for institutional investors
  • 2026-06-21CEO Devon Krapcho's profile lists Cohesion as YC P26, New York based
  • 2026-06-16YC Spring 2026 (P26) batch holds Demo Day, Cohesion included
  • 2026-04-16Cohesion publicly launches via YC Launch as a public equities agentic teammate
  • 2026-04Bloomberg's agentic ASKB research interface reported live, competing with Cohesion's niche
  • 2026-01AlphaSense launches multi-agent Generative Search, intensifying competitive pressure
  • 2026Cohesion founded, entering crowded institutional equity research AI market
Head to head
Revision history
27 July 2026
First assessed. Losing its protection. The text above is fixed as of this date and is not edited afterwards.
In the same band
Sources