Anthropic is now worth $965B, OpenAI $852B, and xAI $230B — the top 10 private AI companies are collectively worth roughly $2.37 trillion. That's the short answer. The longer answer is more interesting.
Eighteen months ago, a $30B valuation put a company in the AI top five. Now it barely cracks the top ten. The gap between the two frontier labs and everyone else has widened into a canyon, and the money behind it is moving faster than any prior tech cycle — including the dot-com boom and the 2021 SaaS peak combined.
The Highest-Valued AI Companies in 2026, Ranked
The highest-valued AI companies in 2026 are led by Anthropic at $965B and OpenAI at $852B, followed by xAI at $230B and Databricks at $134B. Only four private AI companies have crossed the $100B mark, while the rest of the top 10 — Anduril, Figure AI, Safe Superintelligence, Scale AI, Perplexity, and Physical Intelligence — range from $11B to $61B.
| # | Company | Valuation | Round / Date | Prior Mark |
|---|---|---|---|---|
| 1 | Anthropic | $965B | Series H, May 28 2026 | $380B (Feb 2026) |
| 2 | OpenAI | $852B | $122B raise, Mar 31 2026 | ~$500B (Oct 2025) |
| 3 | xAI | $230B | Series E, Jan 2026 | ~$200B (2025) |
| 4 | Databricks | $134B | Series L, early 2026 | ~$100B (prior round) |
| 5 | Anduril | $61B | Series H, May 13 2026 | $30.5B (mid-2025) |
| 6 | Figure AI | $39B | Series C, early 2026 | $2.6B (Feb 2024) |
| 7 | Safe Superintelligence | $32B | $2B raised, 2025–26 | $5B (Sept 2024) |
| 8 | Scale AI | $29B | Meta stake, June 2025 | $13.8B (pre-Meta) |
| 9 | Perplexity AI | $21B | $200M raised, Sept 2025 | $18B (July 2025) |
| 10 | Physical Intelligence | $11B | In talks, Mar 27 2026 | $5.6B (~4 months prior) |
Figures are the latest confirmed or reported valuations as of June 2026, blended from TechCrunch, CNBC, Bloomberg, and company announcements. Databricks and Physical Intelligence figures marked "in talks" in original reporting were not yet closed as of publication; Mistral AI (~$12.7B) and Suno ($5.4B) rank below the top 10 shown here.
Top 10 Private AI Companies by Valuation ($B)
Anthropic and OpenAI alone account for 77% of the combined top 10 valuation.
TechCrunch, CNBC, Bloomberg — latest confirmed valuations as of June 2026.
Why the Highest-Valued AI Companies Keep Getting More Valuable
Anthropic's jump from $380B to $965B took roughly three months. OpenAI added over $350B in value in the same window. That kind of repricing doesn't happen because usage grew 150% — it happens because the pool of capital chasing a handful of deals has gotten enormous, and because crossover investors (hedge funds, sovereign wealth funds, mutual fund complexes) are now writing venture-sized checks to avoid missing the round entirely.
The math: AI venture funding hit $255.5B in Q1 2026 alone, per PitchBook — nearly matching the full $258B raised across all of 2025. Just three deals (OpenAI, Anthropic, xAI) soaked up 67% of that quarter's AI capital, or roughly $172B. That concentration is the story. It isn't that AI investing broadened; it's that a tiny number of frontier labs are absorbing a historically unprecedented share of all venture dollars, private and crossover combined. Track how this compares to public company multiples on our AI Valuations dashboard.
OpenAI vs Anthropic Valuation Growth, 2023–2026 ($B)
Anthropic closed a $585B valuation gap with OpenAI in under two years.
Company funding announcements, TechCrunch, CNBC — valuation at each confirmed funding event.
The Highest-Valued AI Companies Outside the Foundation Model Race
Not every name in the top 10 is a chatbot company. Anduril ($61B) builds autonomous weapons systems and defense hardware, closing a $5B Series H on May 13, 2026 led by Thrive and Andreessen Horowitz — up from $30.5B roughly a year earlier. Figure AI ($39B) and Physical Intelligence ($11B, in talks) are both humanoid robotics plays betting that embodied AI is the next platform shift after chatbots. Databricks ($134B) is the outlier: an enterprise data and AI infrastructure company with real, disclosed revenue rather than a research lab burning cash on compute.
Scale AI sits in its own category after Meta took a 49% non-voting stake for $14.8B in June 2025 — the company remains formally independent and privately held at roughly $29B, but the relationship with its largest customer-turned-investor is unusual enough that any list like this should flag it. Compare how these later-stage private marks stack up against listed comparables on the Unicorns dashboard.
What Happens When One of These Valuations Doesn't Hold
Not every mark in this market goes up. Thinking Machines Lab, the Mira Murati-founded lab that raised a $2B seed at a $12B valuation in July 2025, was reportedly in talks for a $50–60B follow-on round in late 2025. That round collapsed by January 2026, and the company's valuation is still sitting at $12B as of mid-2026 — a reminder that even well-pedigreed AI labs can't assume the next markup is guaranteed. Groq is a starker case: after Nvidia struck a roughly $20B non-exclusive IP licensing deal in December 2025 and hired away founder Jonathan Ross, Groq raised a fresh $650M in June 2026 without disclosing a valuation at all, down from its prior $6–6.9B mark.
The IPO exit is also starting to happen, just not for the biggest names yet. Cerebras listed on Nasdaq (CBRS) on May 14, 2026, raising $5.5B at $185/share — a roughly $23B IPO valuation that popped toward $56B in market cap on debut. Neither OpenAI nor Anthropic has filed, and most operators in this top 10 have said nothing concrete about timing. For LPs and crossover funds marking these positions quarterly, that's the real risk: paper gains this large, this fast, with no public market yet to validate them.
The Bottom Line
Ten companies are now worth a combined $2.37 trillion, and two of them — Anthropic and Openai — account for 77% of it. That concentration is the single most important fact about AI investing in 2026. It means the fate of an enormous share of global venture capital now rides on the execution of two labs, priced at multiples that assume years of continued hypergrowth with no serious competitive erosion. History says some of these marks will hold and some won't. Anduril and Databricks look the most durable because they have defensible revenue and customers outside the AI hype cycle. Thinking Machines Lab and Groq are the cautionary tales already playing out in real time. If you're underwriting exposure to this list — directly, through a fund, or through a crossover vehicle — model the downside case as carefully as the upside, because at these valuations, being early and being right are no longer the same thing.
