Autoscience, a San Mateo, California, applied research lab that is building what it calls an automated AI research lab, announced a $14m seed round on 18 March 2026. General Catalyst led the round. Toyota Ventures, Perplexity Fund, MaC Ventures and S32 also took part, according to Pulse 2.0 and SiliconANGLE.
The company said it will use the money to expand its platform with a select group of Fortune 500 and large private companies and to grow its engineering team. No valuation was disclosed, and neither report stated whether Autoscience had raised money before.
“We've reached a point where human intuition is no longer enough to navigate the complexity of algorithmic discovery”Eliot Cowan, CEO, Autoscience
What Autoscience does
Autoscience says it staffs a virtual laboratory with autonomous AI scientists and engineers that invent, validate and deploy machine learning models. Pulse 2.0 described the planned product as a managed service that runs hundreds of automated AI researchers at once to keep improving a customer's models. The company's argument is that thousands of AI and machine learning papers now appear on arXiv every week, more than human research teams can follow.
Early deployments are in financial services, manufacturing and fraud detection, according to Pulse 2.0. It did not name any customers.
Chief executive Eliot Cowan said: "We've reached a point where human intuition is no longer enough to navigate the complexity of algorithmic discovery." SiliconANGLE also quoted him saying the company aims to "compress a decade of machine learning research into months".
The track record
Autoscience points to two milestones. Its earlier agent, called Carl, produced a paper that was accepted at a workshop of the ICLR 2025 conference. SiliconANGLE reported that human edits were limited to citations and formatting, and gave the paper's title as "Investigating Alignment Signals in Initial Token Representations". The company describes this as the first academically peer-reviewed research produced by an AI system.
SiliconANGLE noted that Tokyo-based Sakana AI had built its own AI scientist and also submitted a paper to an ICLR 2025 workshop that passed peer review, so the claim of being first is the company's own.
Pulse 2.0 reported that Autoscience's autonomous system won a silver medal in the Kaggle Santa 2025 competition, which drew more than 3,300 teams, and that the company described this as a first for a fully autonomous system in a live Kaggle event. Autoscience has also introduced Mira, a machine learning research agent aimed at automating research workflows for frontier and applied AI labs.
Who invested
General Catalyst led the round. Yuri Sagalov, a managing director at the firm, said in the announcement that Autoscience is tackling an increasingly important challenge in machine learning. The other investors are Toyota Ventures, Perplexity Fund, MaC Ventures and S32.
The founding team, the year of founding and the number of employees were not given in the coverage reviewed for this article.
What the money is for
The seed money is meant to widen the platform to more large enterprises in sectors where model quality matters, such as finance, manufacturing and fraud detection, and to speed up what the company calls autonomous research capabilities. Hiring is focused on engineering. Autoscience did not say how many people it plans to add or when the managed service will be generally available.
Company website: autoscience.ai