Mirendil, Co-Founded by 23-Year-Old Ex-OpenAI Intern, Raises $1.2B to Automate AI Research
快鲤鱼 · wechat · 2026-11-16
Mirendil announced a $1B Series A led by Kleiner Perkins and Clearlake with Google participating, bringing total funding to $1.2B within a year at a reported $5B valuation. Its four co-founders include 23-year-old MIT grad and ex-OpenAI intern Tara Rezaei, plus veterans from Google/Anthropic (CEO Behnam Neyshabur, CTO Harsh Mehta) and xAI.
- Thesis: let AI agents handle model training, experiment design, and code development; researchers only set goals and supervise
- Outputs: Mirendil-Base (claimed trained from scratch in 3.5 months for under $10M in compute, 99% of code written by agents), the 10M-token-context Fern, and Lor — a multi-agent R&D platform orchestrating hundreds of agents for days at a time, in closed beta, supporting Claude/GPT/Gemini/DeepSeek/Kimi/GLM
- Compute: a reported $100M+ multi-year Google Cloud deal for TPUs and Nvidia GPUs
- Open questions: no disclosed customers, revenue, margins, or pricing; self-reported numbers lack independent evaluation
The piece also notes Chinese peers: Chaoyan Intelligence (near ¥400M raised, self-evolving foundation models) and Wujie Evolution (¥100M+ Pre-A, virtual cells and recursive scientific discovery).
More from coding & agent
- Agents Now Reply to Slack Messages on Their Own, Resolving a Prod Incident Ghost-to-Ghost — iamrobotbear · 2026-11-16
- Engineer's Interview Prep Method: Use Claude to Find Which Answers Undersell Your AI Work — brandon_galang · 2026-11-16
- Generic API-catalog MCP server vs one tool per job: a context-bloat experiment — MandeepKaur9407 · 2026-11-16
- Berkeley paper: LLMs answer with stale values even when the update is in context — ombre-ai · 2026-11-16
- Effective HTML: 4.2k-star agent skills turn wireframes into self-contained interactive prototypes — tom_doerr · 2026-11-16
- skills-gate: one gate skill fronts 148 agent skills at just 389 tokens — Distinct-Pie2389 · 2026-11-16