IA40

Methodology

Research-driven, industry-sourced list derived from investors who have put several billion into the sector.

The methodology for the IA40 is based on Madrona's experience investing in founders and engaging with researchers involved in AI and Machine Learning for over a decade, and inspired by a research and survey methodology developed by Wing.vc. This list was compiled by and voted on by experts in the space: investors who are funding the next generation of intelligent application companies, technology companies, and a leading investment bank.

  1. Phase 1

    Building the voting list

    In 2025, we received over 340 private company nominations from over 70 venture investors at 54 top-tier venture capital and corporate investment firms. Each judge had the opportunity to nominate companies for consideration in each category. The Madrona team vetted the companies to ensure they met the definition of intelligent applications.

  2. Phase 2

    Voting

    After we collected the nominations, we asked the same venture capital investors to vote on their top 5 companies in each category (early, mid, late, and enabler). Each venture capital investor was limited to voting for only 2 of their portfolio companies.

  3. Phase 3

    Tabulation

    We tabulated the votes from investors by category and leveraged PitchBook's "Venture Exit Predictor Tool." Their model considers and weighs many data points, including the average deal size of each round, number of employees, and number of investors. We weighted the venture community votes and PitchBook's Venture Exit Predictor Opportunity Score to produce a final composite score that resulted in the winners and emerging enablers.

Defining Intelligent Applications

We define application intelligence as the process of using machine learning models embedded in applications that use both historical and real-time data to build a continuous learning system. These learning systems solve a business problem in a contextually relevant way — better than before, and typically deliver rich information and insights that are either applied automatically or leveraged by end users to make superior decisions.

  • Extracting and cleaning data from more than one source, and aggregating it in some form of a data lake/warehouse.

  • Preparing the data and then using it to train and test machine/deep learning models.

  • Refining those models and deploying them in applications to generate contextually relevant predictions, recommendations, inferences or insights.

  • Combining the above "dataware" and cloud-native software, built on a micro-services architecture, to deliver results in real time.

  • Applications enabled to continually monitor and improve the underlying models and predictions.

  • From an end user (or machine) perspective, leverage modern APIs and user experiences.

The Judges

Over 70 investors from 54 top-tier venture capital and corporate investment firms.

  • .406 Ventures
  • a16z
  • Acrew Capital
  • Addition
  • AI2 Incubator
  • Album VC
  • Altimeter
  • Amex Ventures
  • Anthos Capital
  • B Capital
  • Bain Capital Ventures
  • Basis Set Ventures
  • Battery Ventures
  • Bessemer Venture Partners
  • Canaan
  • Coatue
  • Cowboy Ventures
  • CRV
  • Dimension
  • Emergence Capital
  • Essence VC
  • First Round Capital
  • FirstMark
  • Fuse VC
  • General Catalyst
  • Geodesic
  • Goodwater Capital
  • Gradient
  • Greycroft
  • Greylock
  • GV
  • Iconiq
  • Insight Partners
  • IVP
  • Kleiner Perkins
  • M12
  • Madrona Venture Group
  • Mango Capital
  • March Capital
  • Mayfield Fund
  • Menlo Ventures
  • Meritech Capital
  • NEA
  • Notable Capital
  • Playground Global
  • Pruven Capital
  • Radical Ventures
  • Redpoint Ventures
  • Sapphire Ventures
  • Scale Venture Partners
  • Theory Ventures
  • Thrive Capital
  • VMG Partners
  • WndrCo