Figure AI vs Agility Robotics

AI Architecture: Foundation Models vs. Domain-Specific Control

Figure AI's Helix 02 system represents the most ambitious AI architecture in humanoid robotics. The dual-system design — pairing a vision-language model for scene understanding with a vision-language-action model for motor control — enables capabilities like multi-step task reasoning, autonomous error correction, and real-time speech interaction. The recent Helix 02 upgrade extended neural network control to the entire body, enabling continuous walking, manipulation, and balancing as one integrated system. Figure demonstrated this with an end-to-end dishwasher loading task requiring four minutes of uninterrupted autonomous operation.

Agility's AI approach is more targeted but no less sophisticated within its domain. Their whole-body control foundation model, trained through sim-to-real transfer, enables Digit to coordinate locomotion and manipulation simultaneously — walking while carrying objects and adjusting gait for different payloads. Where Figure is optimizing for generality, Agility is optimizing for reliability in a constrained environment, and the 100,000-tote milestone at GXO proves that approach works in production.

Commercial Traction: Revenue vs. Ambition

Agility holds a meaningful lead in commercial deployment breadth. With paying customers including GXO, Toyota Canada, Spanx, and Mercado Libre, Agility has the most diverse roster of commercial humanoid robot customers in the industry. The Toyota deal — deploying seven Digit units on the RAV4 production line — is particularly significant because it validates Digit outside pure warehouse logistics and into automotive manufacturing, traditionally Figure's claimed territory.

Figure's BMW deployment is its flagship commercial reference, and the company is expanding into additional automotive and logistics facilities. However, Figure's strategy has always been to prioritize capability over near-term revenue. The $39 billion valuation reflects a bet on future market capture rather than current commercial traction. The question is whether Figure's AI-first approach will compound faster than Agility's customer-first approach.

Manufacturing and Scale Strategy

Figure's BotQ facility represents the most aggressive production scaling plan in the humanoid industry: 12,000 units per year initially, with a target of 100,000 robots over four years. This capacity dwarfs Agility's RoboFab, which produces thousands of Digits annually. If Figure can fill that production capacity with paying customers, the unit economics advantage could be decisive.

Agility's manufacturing advantage is maturity. RoboFab has been operational longer, and the company has iterated through production challenges that Figure is only beginning to encounter at scale. Agility's Robots-as-a-Service (RaaS) model also reduces the capital barrier for customers, making it easier to scale deployments incrementally — a significant advantage in risk-averse logistics operations.

Fleet Management and Integration

Agility Arc, the company's fleet management platform, is a genuine differentiator that often gets overlooked. Arc doesn't just manage Digit robots — it integrates with autonomous mobile robots (AMRs) from MiR and Zebra Robotics, allowing warehouse operators to orchestrate mixed fleets of humanoids and traditional automation from a single platform. This positions Agility not just as a robot company but as a warehouse automation orchestration layer.

Figure's fleet software focuses on Helix model updates and teleoperation capabilities for data collection, reflecting its AI-first priorities. As Figure scales deployments, the gap in enterprise fleet management tooling could become a friction point with operations-focused customers who need robust monitoring, scheduling, and integration with existing warehouse management systems.

Safety and Human Collaboration

Agility is on track to deliver the first cooperatively safe humanoid robot in 2026 — meaning Digit would be certified to work in direct proximity to human workers without safety cages or exclusion zones. This is a massive commercial unlock. Most humanoid deployments today require restricted work zones, which limits the tasks robots can perform and the value they deliver. Cooperative safety certification would let Digit work alongside human warehouse workers seamlessly.

Figure has not yet announced a comparable safety certification timeline. Given that Figure 03 is a larger, more powerful platform designed for general-purpose tasks, achieving cooperative safety certification may take longer and require more conservative operating parameters. For customers where human-robot collaboration is essential, this could tip deployment decisions toward Agility in the near term.

Long-Term Strategic Positioning

The fundamental strategic question is whether the humanoid robot market rewards generality or specialization first. Figure is betting that general-purpose AI capability is advancing fast enough that a sufficiently intelligent robot will outperform domain-specific competitors across all tasks — including warehousing. Agility is betting that real-world deployment experience, customer relationships, and domain optimization create a moat that general-purpose competitors cannot easily cross.

History offers precedents for both outcomes. In the autonomous vehicle space, Waymo's general-purpose approach eventually outpaced domain-specific competitors. But in industrial automation, purpose-built robots continue to outperform general-purpose systems at specific tasks decades after humanoids became technically feasible. The humanoid robot market may be large enough for both strategies to succeed in parallel, but the next 18 months of commercial deployment data will reveal which approach scales faster.