San Francisco AI lab Mirendil said Thursday it raised a $1 billion Series A led by Kleiner Perkins and Clearlake Capital.
The round arrives just five months after the company’s $200 million seed and places it among the best funded new entrants in frontier AI research.
The round and its backers
Mirendil announced it had raised $1 billion in a Series A, led by Kleiner Perkins and Clearlake Capital, with participation from Battery Ventures, Doerr Capital, Gil & Co, Google, M13, Pear VC, and S32, among other new and existing investors.
The company did not disclose a new valuation in its announcement , though in September, reports said the company was in talks to raise up to $1 billion at a $5 billion valuation with Kleiner Perkins leading.
Mirendil launched in June with a $200 million seed round led by Andreessen Horowitz and Kleiner Perkins at a valuation of around $1 billion, later joined by Nvidia.
The company is led by CEO Behnam Neyshabur, who co-led Google DeepMind’s Blueshift team, and CTO Harsh Mehta, who built the first version of Anthropic’s internal AI research platform.
Kleiner Perkins has backed both rounds, and Google appears on the investor list while also supplying cloud computing capacity under a partnership the company disclosed earlier this year.
What the company is building
Mirendil’s core pitch is that AI agents can take over much of the hands-on engineering work normally done by large research teams.
Mirendil’s pitch is that agents doing the hands-on work let tiny teams do frontier AI research , with a platform called Lor built to do the work of an entire AI lab. Lor runs Anthropic’s and OpenAI’s frontier models alongside open-source models and Mirendil’s own.
The funding follows a research result the company published last month. In an October 21, 2026 research disclosure, Mirendil said one researcher working with AI agents pretrained Mirendil-Base from scratch in 3.5 months using less than $10 million in compute.
That all-in figure covers research and small-scale experiments, the final training run, agents’ token spending and data processing. Agents wrote more than 99% of the code during that pretraining experiment.
That coding-focused result lands at a moment when enterprise spending on AI coding tools has already concentrated around a handful of dominant vendors, a dynamic GCN has previously reported reshaped the market this year.
Mirendil’s agent-built training approach is a bet that the same automation can be pushed upstream into the model-building process itself, not just the tools developers use day to day.
A crowded field of well funded labs
Mirendil now joins a small cluster of richly capitalized startups pursuing frontier AI research outside the largest incumbents.
A $1 billion Series A is rare even in AI, and it puts Mirendil among the best-funded of the new AI research labs, alongside Safe Superintelligence, Thinking Machines Lab and Recursive Superintelligence.
At its June 24, 2026 launch, Mirendil was co-founded by Behnam Neyshabur, Harsh Mehta, Shayan Salehian, and Tara Rezaei, leading a team of 20 researchers and engineers from Anthropic, xAI, Google DeepMind, and OpenAI.
The company has not disclosed revenue, and much of its public case for the new capital rests on the single pretraining result rather than a commercial product track record.
Why the deal matters
The raise underscores how much private capital is still flowing into AI infrastructure and research labs even as compute costs climb and competition intensifies among chipmakers and cloud providers.
That trend is also visible in recent large scale GPU deployments such as Lambda’s buildout with Nvidia’s networking hardware.
For investors, the bet is that a small team paired with autonomous AI agents can match the output of research organizations many times its size, compressing both headcount and cost.
Whether Mirendil’s agent-driven approach scales beyond a single published experiment will likely determine how the rest of its billion dollar round gets spent.