AI foundation model labs
The private labs building frontier AI
Foundation model labs compete for the same talent pool, raise at eye-watering valuations, and structure equity very differently. If you're weighing an AI research or engineering offer, these are the cohort to compare.
Members
| # | Company | Valuation | Latest round | HQ |
|---|---|---|---|---|
| 1 | OpenAI | $852B | Series E | San Francisco, CA |
| 2 | Anthropic | $380B | Series E | San Francisco, CA |
| 3 | xAI | $50B | Series C | San Francisco, CA |
| 4 | Mistral AI | $14B | Series B | Paris, France |
| 5 | Perplexity AI | $5B | Series B | San Francisco, CA |
| 6 | Cohere | $7B | Series D | Toronto, Canada |
| 7 | Character AI | $1B | Series A | Palo Alto, CA |
| 8 | Scale AI | $29B | Series F | San Francisco, CA |
Head-to-head comparisons
Every pairwise comparison in this set, side-by-side.
| OpenAI | Anthropic | xAI | Mistral AI | Perplexity AI | Cohere | Character AI | Scale AI | |
|---|---|---|---|---|---|---|---|---|
| OpenAI | — | vs Anthropic | vs xAI | vs Mistral AI | vs Perplexity AI | vs Cohere | vs Character AI | vs Scale AI |
| Anthropic | vs OpenAI | — | vs xAI | vs Mistral AI | vs Perplexity AI | vs Cohere | vs Character AI | vs Scale AI |
| xAI | vs OpenAI | vs Anthropic | — | vs Mistral AI | vs Perplexity AI | vs Cohere | vs Character AI | vs Scale AI |
| Mistral AI | vs OpenAI | vs Anthropic | vs xAI | — | vs Perplexity AI | vs Cohere | vs Character AI | vs Scale AI |
| Perplexity AI | vs OpenAI | vs Anthropic | vs xAI | vs Mistral AI | — | vs Cohere | vs Character AI | vs Scale AI |
| Cohere | vs OpenAI | vs Anthropic | vs xAI | vs Mistral AI | vs Perplexity AI | — | vs Character AI | vs Scale AI |
| Character AI | vs OpenAI | vs Anthropic | vs xAI | vs Mistral AI | vs Perplexity AI | vs Cohere | — | vs Scale AI |
| Scale AI | vs OpenAI | vs Anthropic | vs xAI | vs Mistral AI | vs Perplexity AI | vs Cohere | vs Character AI | — |
Related
Compiled from public sources. Valuations and figures may be outdated. Not investment advice.
Inclusion in a set is editorial — not a recommendation.