Autonomy Software Firm Helm.ai Signs $70 Million in Contracts Over 12 Months and Says It Is Nearing Breakeven

While most autonomy headlines still revolve around funding rounds and robotaxi geographies, Redwood City's Helm.ai is making a quieter argument: that self-driving software can be run like a software company, not a capital furnace. The company announced Wednesday that it has signed $70 million in commercial contracts over a 12-month period, spanning global automotive OEMs, Tier 1 suppliers and industrial automation companies — and that it is on a path to operating breakeven.

The contracts cover the full stack of the company's physical-AI platform: perception, full-stack autonomous driving, automated data labeling and generative simulation. Helm.ai says its production-bound vehicle programs are now in deep integration with next-generation vehicle architectures ahead of start of production — the phase where autonomy vendors either become supply-chain fixtures or quietly disappear. Beyond the road, the company has deployed production-track perception for heavy industrial equipment in open-pit mining, and says the same model lineage is expanding into robotics.

The technical pitch is "Deep Teaching," Helm.ai's unsupervised methodology for training foundation models on the structure of the physical world rather than on narrowly labeled driving data. Founder and CEO Vladislav Voroninski argues this separates understanding an environment from acting in it, letting the same core models transfer across passenger vehicles, industrial machines and robot platforms with a fraction of the data and compute that end-to-end rivals consume. "Capital efficiency isn't a constraint we manage, but rather a property of the technology," he said in the announcement.

That positioning is aimed directly at the structural economics of the autonomy industry, where the dominant approaches — fleet-heavy data collection, monolithic end-to-end models, and premium in-vehicle silicon — compound costs with scale. Helm.ai's claim is that converting autonomy from a capital problem into a software licensing decision is the only route to surviving the stretch between pilot programs and mass production. Founded in 2016, the company has deliberately flown under the mega-round radar while its competitors burned billions.

The $70 million figure is a contract total rather than recognized revenue, and the company — true to form — disclosed no valuation or new funding alongside it. But the breakeven claim, if it holds through the coming production launches, would make Helm.ai one of the few autonomy software firms to reach the milestone on licensing economics alone. The company says it will share additional announcements later this year; watch for named OEM program wins, which would be the first hard confirmation that the model-transfer story survives contact with a real production line.