Iván Hernández Dalas: FieldAI raises $405M to scale ‘physics first’ foundation models for robots

August 20, 2025

line of robot equipped with FIeldAI FFM brains.

FieldAI is expanding the capabilities of autonomous robots through the use of Field Foundation Models. | Credit: FieldAI

FieldAI today announced that it has raised $405 million in two consecutive rounds. The Mission Viejo, Calif.-based company said it will use the capital infusion to accelerate its global growth. Field AI also said it plans to double in size by the end of the year as it continues product development across locomotion and manipulation.

The latest round was oversubscribed, following rapid customer adoption and multiple expansion contracts for FieldAI’s “general-purpose robotics intelligence,” which it has tested and deployed across hundreds of complex real-world industrial environments.

The market for general-purpose robotics is exploding. With form factors like humanoids grabbing headlines each week. FieldAI said it is at the forefront of this revolution, developing a single software “brain” that is powering a variety of robots in diverse environments.

In addition to humanoids, the company said it has also found success with quadruped applications in construction, energy, manufacturing, urban delivery, and inspection.

Field Foundation Models push robotics limits

At the core of FieldAI’s innovations are Field Foundation Models (FFMs), a class of models developed for embodied intelligence. These AI models are designed to handle uncertainty, risk, and physical constraints, explained the company.

In contrast to standard vision or language models adapted for robotics, FFMs are built to manage physical-world challenges. This design allows for safe, reliable robot behaviors in situations they haven’t been trained for, letting them navigate dynamic, unstructured environments without needing prior maps, GPS, or set paths.

“Our team has spent years in the field, driving major breakthroughs in ‘field robotics’ and safety-critical robotic AI in complex environments,” stated Ali Agha, founder and CEO of FieldAI. “With a deep understanding of the resilience and robustness required to deploy robotic AI in complex real-world conditions, we have taken a fundamentally different approach. 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.”

“With Field Foundation Models, we are enabling robotic operations to scale seamlessly across diverse environments with varying risk profiles, moving beyond the constraints of traditional solutions,” he added.

FieldAI claimed that its architecture is a breakthrough in the robotics space. With an FFM, robots no longer require manual programming. Equipped with an FFM, a robot can safely and dynamically adapt to new and unexpected conditions while executing complex tasks reliably in unstructured environments, it said.

Investors look to long-term benefits of FFMs

Investors in FieldAI included Bezos Expeditions, BHP Ventures, Canaan Partners, Emerson Collective, Intel Capital, Khosla Ventures, NVentures (NVIDIA‘s venture capital arm), Prysm, Temasek, and others. Previous investors include Gates Frontier and Samsung.

“Enabling autonomy solutions at scale is an extremely difficult problem, but the deep expertise of the FieldAI team and their unique approach to embodied intelligence reflects a pragmatic path forward,” said Vinod Khosla, founder of Khosla Ventures. “FieldAI is at the forefront of the general-purpose robotics revolution, and its ability to rapidly deploy will unlock long-term economic and societal value.”

“We are excited and privileged to be partnering with the FieldAI team on this next phase of their journey,” said Jay Park, co-founder and managing partner of Prysm Capital. “Their new class of foundation models offers the reliability and adaptability required for autonomous robotics deployment at scale across numerous sectors.”

“FieldAI’s revolutionary models not only greatly broaden possible use cases but also enable risk-aware deployment, a critical element for scaling AI that has the potential to reshape how robots interact with the physical world,” said Park.

Learn about FieldAI development roots

FieldAI was founded in 2023, but its core development team has nearly a decade of experience developing robotic systems and participating in multiple DARPA challenges. The company said its technology enables robots to operate autonomously in real time, with decisions made directly by the models at the edge, integrating into real customer workflows.

The robots have logged many real-world data and operational hours, demonstrated a transformative pace of model evolution, and delivered value and cost-effective autonomy at scale, claimed FieldAI. As industries turn to automation to address labor shortages, safety risks, and efficiency goals, the company said demand for its platform continues to accelerate.

Learn more about field robotics at the RoboBusiness 2025, which will be at the Santa Clara Convention Center on Oct. 15 and 16. FieldAI will be among the more than 100 exhibitors at the event, which will include the Physical AI Forum, a new session track. Conference attendees will have an opportunity to interact with FFM-equipped robots and speak with experts from the company. Registration is now open.