Standard Intelligence lands $75M to scale computer-use AI models

Standard Intelligence, a San Francisco-based AI startup building foundation models for computer-use tasks, has raised $75 million in new funding. The round was led by Sequoia Capital and Spark Capital, with participation from angel investors including Andrej Karpathy, Stanley Druckenmiller, and Milan Kovac. Public filings and secondary data providers indicate the financing was structured across several Series A-related tranches, with documentation referencing Series A-2, A-3, and A-4 share classes that together sum to roughly the announced $75 million. Multiple reports characterize this as the company’s first major institutional round, backing a six-person team focused on training models that operate software interfaces through video-based learning rather than only text prompts.

The capital will be used primarily to scale compute and further develop Standard Intelligence’s FDM-1 foundation model, which is optimized for general computer-use tasks such as navigating software, filling forms, and executing multi-step workflows across applications. Company materials state that early FDM-1 results have shifted the bottleneck from data to compute, and the new funding is intended to unlock several orders of magnitude more training capacity and support alignment research for agentic, video-pretrained models. According to a startup funding database and one industry report, the round implies a post-money valuation of approximately $500 million, more than 16 times the company’s late-2024 seed valuation. This raise positions Standard Intelligence among the better-capitalized players working on AI agents that can use computers like humans, a capability seen as strategically important for automating knowledge work and complex digital operations.