Mirendil announced a US$1 billion Series A on November 12, 2026, led by Kleiner Perkins and Clearlake Capital. The company says the financing will advance its tools for AI research and development. Its research-agent platform, Lor, remains in private beta; the funding announcement does not announce a public release. Mirendil’s announcement sets out the round and its intended direction.
Who is backing Mirendil’s Series A?
Mirendil named Battery Ventures, Doerr Capital, Gil & Co, Google, M13, Pear VC and S32 among the participating new and existing investors. The US$1 billion figure is the financing amount.
Transaction counsel provides a second account of the investment. In a November 13 notice, Sidley Austin said it represented Clearlake in its co-lead investment. The firm said the funding would support tools intended to make frontier AI research and development more accessible to scientists and engineers.
What Lor and Fern already do, according to Mirendil
The financing follows the company’s earlier product announcements. In its October 28 introduction of Lor, Mirendil described agents that write code, conduct experiments and work on results over multiple days. It says agents can request computing nodes, perform work and release those resources afterward. Access is by invitation during the private beta.
Mirendil also says users can follow agents’ progress and stop them, and that agents operate only on machines users have added. Those are company descriptions of the platform’s controls and capabilities.
Mirendil describes Fern as a model built from Mirendil-Base through post-training, with a context window of up to 10 million tokens. The company positions it for tasks including reading training logs, examining agent traces and working on AI research code. That capacity is a company-reported specification, rather than a finding about a university’s own research workload.
Funding, access and capability are separate signals
The announcements describe three different developments. For research offices assessing the news, the distinction is:
| Signal | What the primary sources establish | Editorial implication for a research team |
|---|---|---|
| Financing | A US$1 billion Series A, announced November 12 and corroborated by transaction counsel. | Treat the round as funding for development, rather than evidence that a particular research task succeeds. |
| Platform access | Mirendil describes Lor as an invitation-only private beta. | Separate interest in a pilot from an assumption of immediate access. |
| Model capability | Mirendil reports that Fern supports up to 10 million context tokens. | Distinguish context capacity from accuracy on the team’s experiments. |
What this means for university research
Our editorial assessment is that the institutional question is how to supervise delegated experimental work. Because Mirendil describes agents operating on selected machines and allocating compute, a university considering a pilot could examine which machines and data it would expose, who could stop a run, and what records it would retain.
For researchers, code review, a task-specific model evaluation suite and records supporting computational reproducibility offer concrete ways to assess a trial. Those are evaluation priorities suggested by the proposed workflow, not outcomes established by the financing. The new round backs Mirendil’s development ambitions; a research team’s own results would determine the tools’ usefulness for its work.
Sources
- Mirendil: Mirendil Raises $1B to Democratize Frontier AI R&D. Checked 2026-11-15.
- Sidley Austin: Sidley Represents Clearlake Capital Group in Mirendil’s US$1 Billion Financing. Checked 2026-11-15.
- Mirendil: Meet Lor. Checked 2026-11-15.
- Mirendil: Introducing Fern. Checked 2026-11-15.
