Mirendil, the San Francisco startup founded by two former Anthropic researchers, has raised $1 billion in a Series A led by Kleiner Perkins and Clearlake Capital, the company said in a release on PR Newswire on 12 November at 12:30 ET. Battery Ventures, Doerr Capital, Gil & Co, Google, M13, Pear VC and S32 also took part, “among other new and existing investors”.
The Series A is five times the $200 million seed round Mirendil announced on 24 June, as TNW reported, and it came 141 days later, by our count. The two rounds add up to $1.2 billion. The release does not state a valuation; TNW reported the seed at a $1 billion valuation.
Business Pill · Power Law Explained
An evergreen Business Engineer explainer on power laws, where a few outcomes take most of the returns, the logic behind venture rounds of this size. The video teaches the general idea only and does not discuss Mirendil.
The key insight: As we read the release, Mirendil is raising on its tooling rather than on a model it sells: the $1 billion Series A is five times its seed, and the release names a private-beta agent product, Lor, with no price, customers or revenue.
The company says it is “building tools to democratize frontier AI R&D and help put it within reach of every scientist and engineer”. It says it shared its first results over the past three weeks.
Per the release, “One researcher, working with agents, pretrained Mirendil-Base, a competitive coding model, from scratch in 3.5 months with less than $10M in compute”. Mirendil says it then post-trained that model into Fern, which “holds up to 10M tokens of context”.
Its third product, Lor, is “now in private beta” and runs “hundreds of agents writing code and running experiments for days at a time”, the release says, adding that “Mirendil runs on Lor every day”. The company calls itself “still a small team with a singular focus”; the release gives no headcount, customers or revenue.
Mirendil’s $1 billion Series A (company release, 12 November 2026) against its $200 million seed announced on 24 June (as reported by TNW). The 5x ratio and 141 days are our arithmetic; the release states no valuation.
The Founders’ Anthropic Record
The founders are Behnam Neyshabur, the chief executive, and Harsh Mehta, TNW reported in June. The pair met at Google in 2019, moved to Anthropic in late 2024 and left in December 2025, according to TNW. Neyshabur had spent more than five years at Alphabet co-leading reasoning research for Gemini, and Mehta built the first version of Anthropic’s internal AI-research platform, TNW reported.
At the seed stage, TNW reported, Mirendil had about 20 researchers and engineers drawn from Anthropic, xAI, Google DeepMind and OpenAI, and the seed was co-led by Andreessen Horowitz and Kleiner Perkins, with Nvidia joining. Kleiner Perkins is the one firm named as a lead in both rounds, by our reading of the two sources.
The Numbers, by Our Arithmetic
The $1 billion Series A is at least 100 times the under-$10 million of compute Mirendil says it spent pretraining Mirendil-Base. From the founders’ December 2025 exit from Anthropic to this round is about 11 months, and the $1.2 billion raised across two rounds came within that span, by our count.
As we read the release, the under-$10 million pretraining claim is the company’s own figure and is offered as evidence that agents shrink the team a frontier projectneeds; the release gives no independent benchmark for Mirendil-Base.
As we read the two sources together, Kleiner Perkins leading both rounds within 141 days, by our count, is the clearest continuity between the seed and the Series A.
Mirendil release via PR Newswire, 12 November 2026
“when agents do the hands-on work, the size of your team stops limiting what you can build.”
Three Implications
FOR OTHER LAB SPINOUTS Mirendil raised $1.2 billion across two rounds within about 11 months of its founders leaving Anthropic, by our count, without a stated valuation in this release.
FOR INVESTORS Kleiner Perkins and Clearlake Capital led the Series A, the release says; Google, Battery Ventures and five other firms are named as participants.
FOR AI RESEARCH TEAMS Lor is in private beta, the release says, running hundreds of agents for days at a time; no launch date or price is given.
The framework argues: “The surface, not the model, governs autonomy.”
As we read it, Lor, an agent product that runs hundreds of agents for days, is a harness bet of that kind; the framework predates the round and does not discuss Mirendil.
Mirendil’s $1.2 billion across two rounds (our sum of the $200 million seed, as reported by TNW, and the $1 billion Series A) against the less than $10 million of compute the company says it used to pretrain Mirendil-Base.
What Is Not Established
The release does not give a valuation, the split between new and existing investors, the amount each lead invested, or a launch date or price for Lor. It does not name Andreessen Horowitz or Nvidia among this round’s investors; whether they took part is not established from the release. The founders’ backgrounds come from TNW’s June report, not from the release. Mirendil was not contacted for this piece.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Mirendil, founded by two researchers who left Anthropic in December 2025, says it has raised a $1 billion Series A led by Kleiner Perkins and Clearlake Capital. That is five times the $200 million seed reported in June, 141 days earlier, and at least 100 times the under-$10 million of compute the company says it used to pretrain its first model, by our arithmetic.
A note on sourcing. The round, its investors and the product claims come from Mirendil’s release on PR Newswire of 12 November 2026, read in full. The seed round, its valuation and the founders’ backgrounds come from TNW’s report of 29 June 2026. Ratios, day counts and sums are our arithmetic. Mirendil was not contacted for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
