Mirendil raises $200M to automate AI R&D
San Francisco-based AI startup Mirendil has raised $200 million in seed funding at a $1 billion valuation co-led by Andreessen Horowitz and Kleiner Perkins. Founded by former Anthropic researchers Behnam Neyshabur and Harsh Mehta, the company aims to build autonomous AI systems that automate experimental design, model evaluation, and iterative training.
Hot Take: Automating the AI research cycle is the ultimate force multiplier, and if Mirendil succeeds, it will disrupt the talent moat currently held by Google, Anthropic, and OpenAI.
* Orchestration over models: Instead of training another foundational model, Mirendil targets the R&D process itself, offering a platform to automate complex, months-long research tasks in days.
* Top-tier talent density: The startup has recruited a team of around 20 researchers and engineers from leading AI companies including DeepMind, OpenAI, xAI, and Anthropic.
* Democratization of frontier AI: Packaging automated R&D tools enables domain-specific researchers, like university biologists, to train custom models without needing elite ML staff.
* Valuation and risk: At a $1 billion valuation for a pre-product seed startup, this is a massive high-risk bet on whether R&D automation is viable today or still years away.
DISCOVERED
2h ago
2026-06-29
PUBLISHED
2h ago
2026-06-29
RELEVANCE
AUTHOR
AI Revolution
