Nominal raises $240M at a $3B valuation to put AI agents on hardware test data

Nominal, which builds software for hardware engineering teams, has raised a $240 million Series C at a $3 billion valuation, Unite.AI reports. Lightspeed Venture Partners led and Insight Partners co-led, and the other named investors include Sequoia, Founders Fund, General Catalyst, Lux Capital and the venture arms of Lockheed Martin, Salesforce and DE Shaw.

What it does

Nominal started by helping hardware teams handle test data: what a part was expected to do, what the sensors actually measured, and why the two differ. Its pitch is that AI is only useful in engineering if it can see that evidence with its context intact, instead of answering a bare prompt. Explaining one odd reading can mean tying it to a hardware configuration, an older experiment and a decision made months back, spread across instruments, simulations, files and teams.

The platform has four layers:

• Connected Control links to physical equipment and test rigs to command them and capture data at the source.
• Data Infrastructure puts sensor signals, files, configs and documents into one shared structure, recording what each measurement means and where it came from.
• Intelligence connects models and agents to that context so they can run analyses and gather evidence engineers can inspect. The first agent, Nominal Analyst, traces its answers back to the underlying data; engineers keep the decisions and approvals.
• Engineering Applications cover development, testing, manufacturing and operations, and Nominal for Developers lets customers build their own apps and agents on top.

It runs in the cloud, on-premises, air-gapped or on edge devices.

Who uses it

Nominal counts over 80 commercial and government customers, including five of the seven biggest defense contractors, across aerospace, defense, robotics, energy and manufacturing. The money goes to new products, new industries and regions, and acquisitions. Next on the list are agents for checking requirements, planning and scheduling, and tracking the health of fleets of machines.

Scoreboard update: Draw. AI agents get access to hardware test data, and engineers keep the final say.

Stay undead.