Flow Engineering raises $50M Series B at a $750M valuation

Flow Engineering, a San Francisco startup that puts AI agents to work on hardware design data, has raised a $50 million Series B at a $750 million valuation. The round was announced in a press release dated September 30, 2026 and reported by several outlets on October 1 and 2. The valuation puts a software-style price on a company selling to carmakers, defense firms and rocket builders. Its customer numbers explain why investors paid it.

Who invested

Antonio Gracias, founder of Valor Equity Partners, and Gavin Baker, managing partner at Atreides Management, co-led the round. Sequoia Capital, which led Flow's Series A, took part again, along with Human Capital, Evantic, SV Angel, Odyssey and EQT. Individual investors include Hugging Face co-founder Thomas Wolf, Mercedes-Benz chief information officer Jonas von Malottki and Formula 1 champion Nico Rosberg.

Roelof Botha, the former Sequoia managing partner who led the Series A, sits on Flow's board as an independent director and invested personally in this round, TechTimes reports. "My conviction in Flow Engineering has only grown since I first led the Series A," Botha said, according to American Bazaar.

Round	Date	Amount	Lead
Series A	October 2025	$23 million	Sequoia Capital
Series B	September 2026	$50 million	Antonio Gracias (Valor) and Gavin Baker (Atreides)

TechTimes puts total funding at $73 million.

What Flow's software does

Hardware teams keep their work in many places: requirements documents, CAD files, simulation outputs, code repositories and test results. Flow pulls all of them into one live graph, TechTimes reports. Its AI agents then watch for changes, trace what each change affects downstream and check whether the design still meets its requirements across those tools.

The pitch is speed. TechTimes says the agents compress manual validation work "from weeks to seconds." Robotics and Automation News says the goal is to cut iteration cycles from months to days. American Bazaar describes the agents aligning CAD drawings with requirements, simulation results and other testing automatically.

"AI has transformed software engineering. Hardware engineering is next," founder and chief executive Pari Singh said, as quoted by American Bazaar. In the press release he put it as a deadline: "Every hardware company now has to decide how fast it will adopt AI."

The Rivian numbers

The customer figure every report repeats is Rivian's. The electric-vehicle maker went from 40 Flow users to 1,500 in seven months, according to the press release and TechTimes. It now makes millions of API calls to the platform each week. "We evaluated 30 tools and nothing came close to Flow," said Scott Mackenzie, Rivian's director of product development, in a quote TechTimes carries.

Flow says 96% of its customers came to it inbound. Its named customers span cars, defense, aviation and space: Rivian, General Motors PPU, RV Tech, Anduril, Joby Aviation, Astranis, Radiant Industries, Stoke Space, Intuitive Machines and Pacific Fusion.

The company moved from London to San Francisco within the past year, TechTimes reports. Gavin Baker summarized the investment case in one line: "Flow is building the OS on which physical products are specified and verified."

What this means for developers

For engineers at hardware companies, the practical question is whether your requirements and test data are structured enough for an agent to read. Flow's model depends on ingesting requirements, CAD, simulation, code and test results together. Teams that keep requirements in slide decks and test results in spreadsheets will get less from any tool like this until that changes.

For software developers, Flow's growth is a signal about where AI agents are finding paying customers. The Rivian figures show adoption spreading inside one company more than thirtyfold in seven months. The 96% inbound figure suggests demand is pulling the product in rather than a sales team pushing it.

For buyers, watch two things. The investor list is heavy with people who back physical-world companies, which says the bet is on hardware adoption rather than on another coding assistant. And the customer list is concentrated in vehicles, defense and space. A team outside those fields should ask for a reference in its own industry before committing.