The Equation Series
Menu

Artificial Intelligence · Profile from AI Research Companion

Hebbia

The New York finance-research AI company, founded in 2020 by Stanford PhD dropout George Sivulka, whose Matrix interface — a spreadsheet in which every row is a document, every column a question and every cell a model-generated answer linked to its source — became one of the best-known AI tools on Wall Street. Hebbia sells to asset managers, private-equity and credit funds, investment banks and law firms; it said in June 2025 that 40% of the largest asset managers by assets under management used its agents, naming BlackRock, Carlyle and Centerview (company). It raised a $130m Series B led by Andreessen Horowitz at a reported ~$700m valuation in July 2024, when it was reported to be profitable on ~$13m of revenue (TechCrunch), and had raised more than $160m in total (Bloomberg, as summarised). Since then it has not announced a confirmed priced round, while the frontier labs it builds on (OpenAI, and Anthropic through Claude for Financial Services) moved into finance themselves. Its 2025–26 response — FlashDocs, a new CTO, the Max agent, Matrix 2.0 and a network of data partners — is the story of a vertical AI company trying to become a platform before its suppliers make it a feature.

At a glance

Founded
August 2020, by George Sivulka (CEO), who left a Stanford PhD programme to start it
Headquarters
New York; San Francisco office from October 2025 (575 Market Street)
Ownership / structure
Private. Investors include Peter Thiel, Floodgate, Index Ventures, Andreessen Horowitz (a16z) and GV; individual investors reported to include Eric Schmidt and Jerry Yang
Leadership (2026)
George Sivulka (founder and CEO); Aabhas Sharma (CTO from October 2025; ex-Found, Uber, Postmates); Paul Boland (partnerships, per company blog; verify title)
AI products
Semantic document search (2020); Matrix (2022); multi-agent Matrix on OpenAI reasoning models (2025); FlashDocs deck generation (2025); Max agent (July 2026); Matrix 2.0 (August 2026); Headless Hebbia API and MCP server (October 2026); memory, version history and Word plug-in (October 2026)
Scale
~$13m revenue and profitable (mid-2024, TechCrunch, reported); ~a third of the 50 largest asset managers as customers (2024, reported, verify) → 40% of the largest asset managers by AUM (company, June 2025); >250bn tokens processed a month (company, July 2026); no later revenue disclosed
Capital raised
Early round from Thiel and Floodgate (2020); $30m Series A led by Index Ventures (2022); $130m Series B led by a16z at ~$700m (July 2024); >$160m in total (Bloomberg, as summarised on Wikipedia); 2026 round claims on aggregator sites unconfirmed (verify)
Key partnerships
OpenAI (model supplier and customer showcase, 2025); Daloopa (September 2026); Octus, Datasite and Databricks (October 2026)
Signature moves (incl. setbacks)
Founding and semantic search (2020); Matrix and the Series A (2022); the $130m Series B at ~$700m (2024); multi-model agents on OpenAI (2025); the FlashDocs acquisition (2025); the frontier labs enter finance (2025, setback); the CTO hire and San Francisco office (2025); Max, Matrix 2.0 and the data-partner network (2026)

8 strategic moves analysed

Each one gets a full strategy file in the companion: context, the decision, the financing, what happened, and the lesson.

  1. 2020

    Founding and the Semantic Search Start

    Founding

  2. 2022

    Matrix Launch and the Index Series A

    Launch / funding round

  3. 2024

    130m Series B at 700m

    Funding round

  4. 2025

    CTO Hire and San Francisco Expansion

    Key hire / office expansion

  5. 2025

    FlashDocs Acquisition

    Acquisition (acqui-hire with product)

  6. 2025

    Frontier Labs Enter Finance

    Competitive threat (move made by others) / setback

  7. 2025

    Multi Model Agents on OpenAI

    Launch / partnership (model supplier)

  8. 2026

    Max Matrix 2 and the Data Partner Network

    Launch / partnerships