Reports / Frontier AI Labs talent structure: OpenAI, Anthropic, xAI and 13 labs

FRONTIER AI LABS

Frontier AI Labs talent structure: OpenAI, Anthropic, xAI and 13 labs

These maps compare who works at frontier labs, where they came from, and which roles concentrate. Thirteen labs contribute 19,057 current profiles and an 11,914-person technical pool. A separate US cut of OpenAI, Anthropic, and xAI holds 9,421 profiles, with engineering between 40.3% and 44.7%.

Visible public-profile samples. Not official headcount.

Ask for the lab bench behind these maps

13 labs, 11,914 technical profiles, and a 9,421-profile US comparison of OpenAI, Anthropic, and xAI.

Reports in this map

Questions this hub answers

  1. What share of OpenAI technical staff hold a PhD? 14.4% of 3,041
  2. Where does OpenAI hire engineers from? Google 702 (23%)
  3. What is the engineering share at OpenAI vs Anthropic vs xAI? Engineering 40.3–44.7%
  4. What share of xAI is Human Data & Evaluation? 28.4% · 425 / 1,496
  5. How many visible moves from OpenAI to Anthropic? 47 vs 10
  6. What share of the 13-lab technical pool is ethnic-Chinese? 4,119 · 34.6%

FAQ

How many people are in the 13-lab technical pool?

11,914, inside 19,057 current profiles at 13 labs.

Do OpenAI, Anthropic and xAI have the same engineering mix?

Engineering runs from 40.3% to 44.7% in the 9,421-profile US sample.

What share of OpenAI technical staff hold a PhD?

14.4% of 3,041.

How many visible OpenAI-to-Anthropic moves?

47, against 10 the other way.

What share of the 13-lab technical pool is ethnic-Chinese?

4,119 people, 34.6%.

How to cite

Metix AI. (2026). Frontier AI Labs talent structure: OpenAI, Anthropic, xAI and 13 labs. Metix AI Talent Intelligence. https://metix.ai/reports/topics/frontier-ai-labs

How we build these maps

Ask Metix AI to map a lab role mix

Open a hiring workspace on the same frontier-lab cuts.

Start a Mira trial

中文

中文