FieldAI Inc., a developer of artificial intelligence software for robots, today disclosed that it has raised $405 million in funding over two rounds.

The bulk of the capital came in the form of a $315 million investment led by Bezos Expedition, Prysm and Temasek that closed earlier this month. FieldAI also counts the venture capital arms of Nvidia Corp. and Intel Corp. among its backers. The company is now reportedly worth $2 billion, up from $500 million last year.

FieldAI offers a software platform that enables robots to navigate challenging environments such as construction sites. The software is powered by AI models that the company refers to field foundation models, or FFMs for short. FieldAI says that the algorithms are more suitable for use in robots than standard neural networks.

Historically, companies had to use different autonomy software for each type of robot used in an automation project. FieldAI says that its FFMs reduce the need for such customization. The company’s software can run on humanoid robots, autonomous vehicles and a range of other systems out of the box.

FieldAI trains its FFMs using data collected from the customer sites where its software is deployed. The company also uses synthetic data, or files generated by software for AI training purposes. FieldAI generates synthetic data using thousands of robot simulations powered by Nvidia’s open-source Isaac Lab tool.

Customers can deploy multiple FieldAI-powered robots in the same facility and configure them to coordinate their work. According to TechCrunch, the company’s software reduces the risk of navigation errors by estimating “how confident it is” in a given decision. Customers can prevent the robot from making decisions if the risk of errors exceeds a certain threshold.

Before deploying a robot in a new location, developers must usually provide the system with a map of that location. The process of creating a map can take weeks, which increases the cost of automation initiatives. Moreover, maps often aren’t available for fast-changing environments such as construction sites.

FieldAI says its FFMs can manage without maps, GPS access or user-defined robot travel paths. That reduces the amount of effort involved in deploying robots powered by the company’s software.

The company disclosed today that customers have deployed its platforms in hundreds of industrial environments. Real-estate companies can use FieldAI-powered robots to monitor construction sites for blueprint adherence. A factory operator, meanwhile, could deploy robots in its plants to inspect production equipment.

“Rather than attempting to shoehorn large language and vision models into robotics — only to address their hallucinations and limitations as an afterthought — we have designed intrinsically risk-aware architectures from the ground up,” said founder and Chief Executive Officer Ali Agha.

The company will use the proceeds from its two funding rounds to double its headcount by year’s end. The new hires will help speed up its engineering efforts and grow its international presence.

Photo: FieldAI