Key Notes
- Mirendil announced a $1 billion Series A led by Kleiner Perkins and Clearlake Capital, with Google among the participants.
- Its Lor platform coordinates agents that write code and run experiments, while researchers choose what to investigate and trust.
- Lor remains in an invitation-only private beta, and the funding announcement did not disclose a valuation.
Mirendil has raised $1 billion to build systems that let AI agents take on more of the work of developing AI, giving a young research lab substantial backing for one of the industry’s most consequential bets.
The company announced the Series A on November 12, led by Kleiner Perkins and Clearlake Capital. Participants include Google, Battery Ventures, Doerr Capital, Gil & Co, M13, Pear VC and S32. The announcement did not disclose a valuation.
Mirendil’s aim is to make the engineering and experimentation behind frontier models accessible to smaller research teams. Its pitch is that scientists should be able to direct ambitious work without first assembling all the people and infrastructure of a major AI laboratory.
Agents doing the hands-on research work
The company’s Lor platform coordinates agents that write code, run experiments and work through results over extended periods. Mirendil describes a system in which agents can divide a project among many parallel workers and request computing resources when needed.
Researchers can assign work such as running a set of experiments overnight or reproducing a paper’s results. They can follow the agents’ progress and stop them. The company’s description keeps people responsible for deciding what to investigate and which results to trust.
Mirendil says one researcher, assisted by agents, pretrained its Mirendil-Base coding model in three and a half months using less than $10 million in compute. That figure describes computing expenditure, rather than the full cost of operating the company or developing its products.
A platform that works across model providers
Lor supports models from other AI developers alongside Mirendil’s own. Its Fern model has a claimed context window of up to 10 million tokens, intended to hold the material generated during long stretches of agent work.
The company describes self-accelerating AI research as its broader goal. Automating engineering tasks is one part of that effort; demonstrating that a system can reliably improve research itself is a more demanding test.
Major funding before broad access
Lor remains in an invitation-only private beta. The funding announcement therefore marks an expansion of resources behind the project, rather than a general release that every researcher can use immediately.
The investment arrives as established labs are also examining how AI changes their own development cycles. Anthropic has warned that future models could accelerate research faster than it can validate safeguards, while saying it has not designated a current model as crossing its automated-research threshold.
For Mirendil, the practical challenge is turning the work of many agents into results that researchers can reproduce, evaluate and build on. The new capital gives it room to pursue that approach at a much larger scale.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
