Agility Robotics has become the first pure-play humanoid robotics company to go public on the stock market through a reverse merger deal with a Special Purpose Acquisition Company (SPAC). This move aims to raise approximately $465 million in new capital, supported by a mix of individual investors and private funding, at a time when startups in the field of physical artificial intelligence are facing increasing pressure to deliver real financial returns.
Field Trials and Company Revenue
Unlike competitors who focus on showcasing dancing robots or acrobatic movements, Agility has focused since its founding as a research spin-off from Oregon State University in 2015 on repetitive manual tasks that humans prefer to avoid. The company has deployed its Digit robots with nine major clients, including Toyota, Mercado Libre, and GXO, where they have been used to move more than 100,000 heavy boxes in supply facilities.
Although Digit robots generated modest revenue of only $174,000 in the first half of 2026 amid financial losses of $362 million, the company is banking on the new generation, Digit 5. The company has received planned contracts worth $300 million from 30 clients who intend to deploy the new version, which has the ability to work safely alongside humans instead of in isolated cells.
What is Changing Practically?
Agility's strategy centers on providing a Robot-as-a-Service (RaaS) model that saves companies approximately $100,000 annually per robot that replaces humans in lifting heavy, repetitive loads that cause workplace injuries. The company relies on accelerating the training of its systems using more than 65,000 hours of real operational data collected from actual client facilities, in addition to data simulation to compensate for the shortage in physical internet for robot movements.
Challenges and Open Questions
The company faces fundamental questions regarding the speed of finding new clients and its ability to expand the capabilities of the Digit 5 robot, whose prototypes will be showcased during CES 2027. Potential clients must pay $500,000 to conduct a field proof-of-concept before full-scale expansion. Despite the robot's 98% accuracy in picking and placing loads, it still lacks the complex manipulation tasks that generative AI technologies are attempting to address through transformer models.