FieldAI Inc., a startup that develops artificial intelligence models for robots, is reportedly seeking $700 million in funding.
A source told Business Insider today that the round could value the company at $10 billion. That’s five times what FieldAI was worth last August. It’s believed the company has already signed a term sheet, a nonbinding preliminary version of a fundraising agreement.
The report didn’t name the investors who are expected to participate in the deal. Large, growth-stage funding rounds often draw contributions from a startup’s existing backers. FieldAI’s investor roster includes Bezos Expeditions, Nvidia Corp.’s NVentures fund and Intel Capital.
Engineers must often provide their robots with a map of the environment where they’re expected to operate. Preparing such charts is a time-consuming task, particularly when they need to cover a large area. Furthermore, static maps don’t account for the fact that environments such as construction sites change on a regular basis. That can lead to potentially significant accuracy gaps.
FieldAI says that its AI models don’t require a map to navigate. They can also make do without GPS access, an internet connection or user-defined travel paths. That significantly reduces the amount of work involved in setting up a robot’s navigation stack, which lowers the cost of automation initiatives.
FieldAI’s models can automatically adjust their behavior in response to risks. For example, they may slow down a robot when it enters a section of a factory where the lights are malfunctioning. Such changes make it easier to avoid collisions and navigation mistakes.
FieldAI says that its models can run on a variety of systems ranging from autonomous vehicles to humanoid robots. In March, it partnered with Boston Dynamics Inc. to add support for the latter company’s four-legged Spot robot. The integration will enable industrial customers to automate certain equipment inspection tasks.
Industrial robots often include not only cameras but also lidar equipment, radars and other sensors. FieldAI turns data from those sensors into a digital twin of the environment where the robots are deployed. A digital twin is similar to a simulation in many respects. The difference is that it’s continuously updated with data from the real-world environment that it emulates, which boosts output accuracy.
FieldAI customers can use the digital twins it generates to virtually test site changes. For example, a manufacturer might simulate the impact of factory floor modifications before implementing them. FieldAI says that its digital twins can also be used for AI training.
The company is said to have surpassed “$100 million in revenue and in customer contracts” this past June. According to today’s report, that number has since topped $135 million. FieldAI has more than 30 customers across verticals such as construction, energy and the public sector.