FieldAI Reportedly Raising $700M at a $10B Valuation

FieldAI Inc., a startup that builds AI models to control robots, is reportedly in the middle of a large fundraise. According to a source cited by Business Insider, the company is looking to raise $700 million in a round that could value it at $10 billion.

That figure would be five times FieldAI's valuation from last August. The report says the company has already signed a term sheet. A term sheet is a preliminary, nonbinding document that sets out the main conditions of an investment before the final agreement is signed.

Who might be writing the checks

The report did not name the investors expected to join the round. Large growth-stage rounds often include money from a startup's existing backers. FieldAI's current investors include Bezos Expeditions, Intel Capital and NVentures, the venture fund of Nvidia Corp.

The problem FieldAI is trying to solve

Most robots need a map before they can work. Engineers usually have to give a machine a detailed chart of its operating environment, and building those charts takes time, especially across large sites.

Static maps also go out of date. A construction site changes from week to week, and a map drawn at the start of a project may no longer match what is actually on the ground. That mismatch can create real accuracy problems for a robot that relies on it.

FieldAI says its models avoid this step. According to the company, they can navigate without a map, without GPS, without an internet connection and without travel paths defined by a user. In practice, this means less setup work for a robot's navigation stack. Less setup work lowers the cost of an automation project.

The models are also designed to react to risk. FieldAI's example: if a robot enters part of a factory where the lighting is faulty, the model may slow it down. Adjustments like that are meant to reduce collisions and navigation errors.

One model, many machines

FieldAI says its software can run on a wide range of hardware, from autonomous vehicles to humanoid robots. In March, the company partnered with Boston Dynamics Inc. to support Spot, Boston Dynamics' four-legged robot. The integration is aimed at industrial customers who want to automate certain equipment inspection tasks.

Digital twins from sensor data

Industrial robots typically carry more than cameras. Many also use lidar, radar and other sensors. FieldAI takes the data from those sensors and builds a digital twin of the site where the robots are working.

A digital twin works much like a simulation, with one key difference: it is updated continuously with data from the real environment it represents. That ongoing feed is what keeps its output accurate.

Customers can use these twins to test site changes virtually before making them. A manufacturer, for example, could model how a new factory floor layout would play out before moving any equipment. FieldAI says the twins can also be used to train AI models.

Revenue and customers

FieldAI reportedly passed "$100 million in revenue and in customer contracts" in June. According to the latest report, that combined figure has since risen above $135 million. The company counts more than 30 customers in sectors including construction, energy and the public sector.

Our Take

The numbers here are worth reading carefully. The revenue figure combines revenue with customer contracts, so it does not show how much money FieldAI has actually booked. And the round itself is still only reported, based on a single source and a nonbinding term sheet. Until the deal closes, the $10 billion valuation should be treated as provisional.

Still, the direction fits a wider pattern. Investors are putting serious money into AI that has to deal with the physical world, from world model startups like General Intuition to AMD's acquisition of World Labs. FieldAI's pitch is more practical than most: cut the setup cost of deploying robots on messy, changing industrial sites.

It is worth watching whether the round closes at the reported terms, which investors take part, and whether the Boston Dynamics integration turns into visible customer deployments.