Bessemer Venture Partners on 12 November published its inaugural AI 100, a ranking of the top 100 private AI companies made with Newcomer Media and announced at the Cerebral Valley AI Summit in San Francisco. Anthropic is No. 1, ahead of OpenAI, Databricks, Stripe and Waymo. Bessemer puts the list’s total value at $3.6 trillion.
The value is concentrated. Frontier labs and foundation model builders are 9% of the list and hold 55% of its value, $1,974 billion, according to Bessemer’s write-up. That works out to about $219 billion per lab, roughly 14 times the $15.7 billion average Bessemer gives for the application layer, by our arithmetic.
Business Pill · Power Law
A short explainer on power laws, where a few winners hold most of the value. It teaches the general idea and does not discuss the AI 100.
The key insight: As we read Bessemer’s figures, the list is two populations: nine labs averaging about $219 billion each, and 91 companies averaging about $18 billion each, by our arithmetic.
The AI 100 replaces the Cloud 100, which Bessemer published every year from 2016 to 2025. Companies were scored on market leadership (35%), estimated valuation (30%), operating metrics (20%) and people and culture (15%). The full list is hosted by Newcomer.
The top 10 are Anthropic, OpenAI, Databricks, Stripe, Waymo, Cognition, Anduril, Ramp, Fireworks and ElevenLabs. Bessemer says they average $250 billion each and account for 70% of the list’s value.
Bessemer partner Sameer Dholakia is quoted in the post: “intelligence is going to be bigger as a market than oil”.
Bessemer AI 100, November 2026: nine frontier labs and foundation model builders hold $1,974 billion, 55% of the $3.6 trillion total. The other 91 companies and about $1,626 billion are our arithmetic.
The Numbers Behind the List
Bessemer gives an average valuation of $36 billion and a median of $9.3 billion, against $11.2 billion and $5 billion for the 2025 Cloud 100. The $3.6 trillion total compares with $1.1 trillion for the 2025 Cloud 100.
The average ARR multiple is 57x, above the Cloud 100’s 2021 peak of 34x; the median is 23x. Average ARR growth is 480% and median growth 275%. Bessemer argues that, “on a growth-adjusted basis, the AI 100 is the cheapest cohort in recent history”.
Six companies have each raised more than $10 billion. OpenAI has raised $224 billion and Anthropic $130 billion, according to the post. Excluding those two, the average company raised $3.9 billion and the median just over $1 billion.
Where the Value Sits
Bessemer calls the labs “the most concentrated category in a decade of Bessemer’s benchmarking”. The other 91 companies share about $1,626 billion, by our arithmetic from the $3.6 trillion total.
The application layer is worth $629 billion across roughly 40 companies, Bessemer says. Physical AI has 13 companies worth $332 billion, about $25.5 billion each by our arithmetic. Cybersecurity and biotech have three companies each and together less than 3% of the value. Data and agent infrastructure is the largest category by count, with 16 companies.
As we read the post, the lab share is the binding number: 55% of the value sits with 9% of the companies, so the $36 billion average says more about the labs than about a typical company, whose median Bessemer puts at $9.3 billion.
As we read the capital figures, OpenAI and Anthropic have raised $354 billion between them, about 80% of what the six companies above $10 billion raised combined, according to the post.
Bessemer Venture Partners, The AI 100, 12 November 2026
“the most concentrated category in a decade of Bessemer’s benchmarking”
Three Implications
FOR FOUNDERS OUTSIDE THE LABS Bessemer says the application layer averages $15.7 billion, above the $9.3 billion list median.
FOR INVESTORS The post puts the average ARR multiple at 57x and the median at 23x.
FOR OUR COVERAGE We will report per-company valuations if Bessemer or Newcomer publish them.
The framework says: “That is why the AI economy is not a list of companies racing each other.”
As we read it, Bessemer’s own split points the same way: the list is ranked as one table, but its value sits mostly in one layer, the labs. The framework does not discuss the AI 100.
What Is Not Established
Bessemer did not publish each company’s valuation in the post, and the scoring includes its own estimated valuations. The post does not say how many listed companies Bessemer has backed. Talia Goldberg, one of its six authors, lists Cognition, Ramp, fal AI and Mind Robotics among the companies she is partner to; all four are on the list. Survey figures, such as 87% expecting value to accrue to the application layer, come from more than 300 AI-native applicants and are not market data.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Bessemer’s first AI 100 values the top private AI companies at $3.6 trillion, with Anthropic first. Nine labs hold 55% of that value, about $219 billion each against $15.7 billion for the average application company, roughly 14 times, by our arithmetic.
A note on sourcing. List figures, categories and quotes come from Bessemer Venture Partners’ Atlas post of 12 November 2026, The AI 100: benchmarks for the third supercycle; the list itself is hosted by Newcomer. Per-lab, per-company and remainder figures are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
