Mirendil raises $200M in one of AI’s biggest seed rounds

A new AI lab staffed by veterans of Anthropic, OpenAI, Google DeepMind, and xAI has raised one of the largest seed rounds in artificial intelligence history, highlighting continued investor demand for startups pursuing frontier AI research.

Mirendil, co-founded by former Anthropic researcher Behnam Neyshabur, raised $200 million in seed funding at a valuation of roughly $1 billion. Andreessen Horowitz (a16z) and Kleiner Perkins co-led the round, while NVIDIA participated alongside other investors.

The financing places Mirendil among a growing group of AI startups securing unusually large amounts of capital before launching commercial products. At a reported $1 billion post-money valuation, the $200 million raise implies that incoming investors acquired roughly 20% ownership in the company, illustrating how venture firms are increasingly willing to fund AI research organizations at a scale traditionally associated with later-stage startups.

A lab built to automate AI R&D

Mirendil looks to innovate frontier AI benchmarks from world-class experts in engineering and scientific research and provide those capabilities to third parties.

“When you think about what does a scientist or an engineer do, the main skill they have is get very deep in a domain and build a very sharp expertise that accumulates over time,” Neyshabur stated in an X video shared by a16z.

The company’s vision aligns with a more general trend across the industry to develop autonomous AI and agentic AI systems to reason, write code, utilize different types of tools, and accomplish more complicated tasks with less human involvement. Mirendil has a specific aim of enhancing the capability of AI research and development by using AI solutions to aid in the creation of new AI systems.

According to a16z, the objective of the company is to expand AI frontier capabilities, which have traditionally remained with a limited number of research laboratories globally. Kleiner Perkins’ Mamoon Hamid explained that the company’s goals are like the goals of a frontier laboratory based exclusively around developing frontier AI research and development and that the team was achieving remarkable results in a short amount of time.

Twenty researchers, four co-founders

Mirendil’s founding team reportedly consists of 20 engineers and researchers, many of whom have come from notable companies such as Anthropic, xAI, Google DeepMind, and OpenAI. Co-founders of Mirendil include Neyshabur, Harsh Mehta, Shayan Salehian, and Tara Rezaei, who have extensive experience working on large-scale machine learning and optimization research.

Research shows that founders who leave traditional AI companies often do so to launch independent labs focused on research into foundation models or new types of AI development methods. Neyshabur has, therefore, created Mirendil as part of this trend of ex-frontier lab researchers creating start-ups like Mirendil that focus on developing next-generation AI technologies and applications.

According to Matt Bornstein of a16z, the team is working on one of the major hyperscale problems in AI, but he did not go into details.

Mirendil’s seed round gives insight on the AI funding marketplace

Mirendil’s investment is part of a trend of investors increasing the amount they spend on top-tier talent in AI, and investing significantly into long-term research goals versus looking at how quickly they will generate revenue or how much customer growth they have.

Mirendil’s round comes after a series of very large investments in newly formed AI labs. One such lab is Safe Superintelligence (SSI) started by former Chief Scientist at OpenAI, Ilya Sutskever, who has raised approximately $1 billion in 2024 with a valuation close to $5 billion. Another prominent newly opened lab is the Thinking Machines Lab, which was led by former OpenAI chief technology officer, Mira Murati, who has also received a $2-billion commitment from investors despite not presenting much information regarding the development of any product.

Mirendil’s newest round is smaller than the amounts that have been raised by others; however, Mirendil is still among the most sizable new funding rounds for companies operating in the AI field. Venture capital firms are now increasingly looking at and viewing frontier AI research as valuable strategic assets that could provide outsized value to the firm if they can create innovative new models.

No revenue reports, product launches or technical details for Mirendil’s technology have been released to date.

NVIDIA’s strategic interest

NVIDIA is increasingly getting involved in the AI startup space as AI development becomes more compute-intensive. NVIDIA has been investing in a number of AI startups, and these investments may help to increase the amount of demand for their GPUs that will be used to train and run advanced AI models. As a result, as NVIDIA has financed startups and partnered with many of them, it is building up its presence in the AI ecosystem and increasing the visibility of future AI trends in development that will ultimately result in sales for NVIDIA.

Autonomous AI research gains momentum

Mirendil’s emphasis on conducting AI-driven research establishes it as a key player in one of the fastest-growing trends within the AI sector. Top-tier AI companies (i.e. Anthropic, OpenAI, and Google DeepMind) are increasingly shifting their R&D focus away from traditional human-assisted AI technology and toward building agent-like systems that can support performing complex tasks with minimal human assistance.

The essential idea behind agent-like systems is that once created, they will then be able to assist with creating future agent-like systems and help advance scientific discovery, improve software development platforms, and enhance the capabilities of existing computer models. Thus, more and more, investors are willing to invest in these types of technologies, even before there are many established commercial applications for them.

In addition to demonstrating how the distinctions between startups and research institutions are becoming less clear, the Mirendil deal also demonstrates how the frontier AI sector is moving away from traditional evaluation criteria related to near-term revenue-generating potential to the scientific research talent and technical aspirations and computing resource requirements of the AI companies.

The implication for investors in this case is that the $200 million raised by Mirendil represents a different type of bet by investors on whether this type of autonomous AI-driven research could represent a significant area of growth for the next generation of AI advancements. Furthermore, the deal emphasizes how competitive the frontier AI laboratory segment has become as financing continues to be funneled into teams attempting to advance the field of artificial intelligence.

The smartest crypto minds already read our newsletter. Want in? Join them.