Agility Robotics CEO Details Digit 5 Safety Model, $300M Order

Agility Robotics has begun detailing the operational reality behind Digit 5, its latest humanoid robot designed for warehouse deployment, including how the system will handle close encounters with human workers and what convinced a customer to place a $300 million order.

CEO Peggy Johnson told Automation Watch that Digit 5 will begin early deployments in the first half of 2025, with general availability by year-end. The timeline reflects a pragmatic approach to safety certification that prioritizes insurance requirements and third-party validation over speed to market.

Cooperative safety, not collaboration

Digit 5 operates under what Johnson calls "cooperative safety" — a model that allows humans and robots to share space but not directly interact. The robot uses 360-degree sensors to detect nearby people. If a human approaches too closely, Digit 5 brings itself to the ground in a controlled descent, powers down its limbs, and waits for the person to pass before resuming work.

This approach addresses payload concerns — the robot may be carrying expensive items like chip wafers — and meets the certification bar set by insurance companies. Johnson noted that Agility worked with nationally recognized test labs for OSHA compliance and will pursue the same third-party certification for Digit 5 that it obtained for Digit 4.

The sitting protocol is intentional. "We bring everything down in a controlled manner and allow the human to pass and then bring it back up again," Johnson said. The next tier, "collaborative safety," would involve direct human-robot interaction, but that remains a future milestone for the industry.

Why it matters

Agility's disclosure of its safety model and customer commitments offers rare transparency in the humanoid robotics sector. The company is preparing to go public via SPAC before year-end, making it the first pure-play American humanoid firm to do so. With over 65,000 operational hours logged and a major customer willing to commit $300 million based on Digit 4 performance, Agility is setting deployment benchmarks that competitors will need to meet. The cooperative safety framework also clarifies realistic near-term expectations: humanoids will work near people, not with them, for the foreseeable future.

Built for three-shift operations

Digit 5 carries up to 50 pounds — a 40% increase over its predecessor — and features a 10-to-1 run-to-charge ratio: 90 minutes of runtime, nine minutes of charging. That enables 20 hours of operation per day across three shifts, a capability customers specifically requested.

The robot now includes an ISO standard flange for interchangeable end effectors, allowing customers to use industry-standard grippers alongside Agility's own tools. Current and planned capabilities span tote moving, de-palletization, machine tending, kitting, quality control, and palletization.

Johnson emphasized that Agility trains its physical AI using a mix of operational data, simulation, and motion-capture demonstration — not relying on any single method. Skills are added based on customer workflow requests, with AI accelerating the training process.

The $300 million customer

The company's recent SPAC filing revealed a single customer accounts for the majority of its $300 million in commitments. Johnson described this customer as the first to complete Agility's acceleration program, which validates workflows with Digit 4 before scaling deployment with Digit 5.

Agility projects adoption of approximately 800 units in 2027, 7,000 in 2030, and up to 25,000 in 2035, with an $8,500 monthly robotics-as-a-service licensing model. Johnson cited 400,000 unfilled U.S. manufacturing jobs in 2025, projected to reach 1.9 million by 2033, as the core market driver.

The company positions itself as "enterprise hardened" after three years of deployment experience, contrasting its approach with what it views as less pragmatic competitors. "We try not to fall into any hype cycles," Johnson said. "We just focus on the need."

Details were first reported by Automation Watch.