Skip to content

Forbes AI 50

An annual, unranked list of 50 privately held AI companies that Forbes, with Sequoia Capital and Meritech Capital, judges most promising. It is a closely watched signal of which AI startups top-tier investors and editors consider category leaders.

Reliability
Grade C · 7/12
Publisher
Forbes
Ranks
Startups, Artificial intelligence, privately held companies (North America 2019 to 2022, global from 2023)
Editions
8 since 2019
Method
Mixed
How items get in
Self-nominated
Status
Active

Editions

EditionPublishedTop of the listOn record
2026April 16, 2026Not recorded50
2025April 10, 2025Not recorded21 of 50
2024April 11, 2024Not recorded1 of 50
2023 Method changedApril 11, 2023Not recorded11 of 50
2022May 6, 2022Not recorded13 of 50
2021April 26, 2021Not recorded50
2020July 3, 2020Not recorded50
2019September 17, 2019Not recorded50

How the ranking works

Companies self-nominate and submit qualitative information (business model, technical talent, AI use) plus metrics such as valuation and funding. An algorithm designed by Sequoia partner Konstantine Buhler scores nominees using submitted data and public signals (PitchBook, Crunchbase, LinkedIn, employee satisfaction, customer traction).

Top scorers go to expert panels: AI experts from academia and industry and, since 2023, a VC panel; judges recuse from companies they have invested in. Forbes editors pick the final 50, presented alphabetically. The shortlist size has varied (top 100 in 2020 to 2023, top 70 in 2026).

Unranked list: Inclusion is the honor. No positions are published. Companies submit through an open nomination form (no application fee per Forbes). Data partner Sequoia Capital designs the screening algorithm and Sequoia and other VC judges have portfolio stakes in some honorees; Forbes says conflicted judges recuse. Mayfield was the 2026 sponsoring partner.

Read the publisher’s methodology

How far to trust it

Superscout grades every ranking on six criteria, each scored 0 to 2. Forbes AI 50 scores 7 of 12, grade C.

Transparency 1/2
Process described each year, but algorithm weights are not published.
Independence 1/2
VC partners with portfolio stakes help design and judge, with recusals.
Evidence 1/2
Mix of data screen and panel judgment.
Stability 1/2
Geographic scope and panel structure changed in 2022 and 2023.
Coverage 1/2
Top slice of applicants.
Track record 2/2
8 editions.

History

Launched September 2019 as 'America's Most Promising Artificial Intelligence Companies' (U.S. only). It moved to spring publication from 2021, covered North America in 2022, and opened to companies worldwide in 2023, when submissions nearly doubled to 796. In 2026 Forbes added the AI 50 Brink List for early-stage startups and named Mayfield as sponsor.

How it is received

Honorees publicize inclusion widely, and Sequoia publishes a companion trend essay each year. The involvement of venture firms as data partners and judges is disclosed by Forbes, which lists judges' investments (for example Bain Capital Ventures' stakes in several 2026 honorees) and says conflicted judges recuse. The list has concentrated on well-funded companies: in 2026, OpenAI and Anthropic accounted for about 80% of the honorees' $305.6 billion in funding.

Strengths

  • Multi-stage process with an algorithmic screen and expert judges.
  • Conflicts of interest are disclosed and judges recuse.
  • Global since 2023 and free to enter.
  • Annual editions since 2019 give a record of which AI startups were favored.

Weaknesses

  • Unranked, so it cannot separate leaders from the rest.
  • VC data partners and judges have stakes in some honorees.
  • Self-nomination means non-applicants may be missed.
  • Scope changed from U.S. to North America to global, limiting year-over-year comparisons.

Open questions

Facts not verified yet. They stay visible so they are not mistaken for settled facts.