Tesla Optimus, Figure 02, and Agility Digit are all in production deployments. We compare capabilities and costs.
The humanoid robot race has moved from demo videos to real factory deployments. In Q1 2026, three companies—Tesla, Figure AI, and Agility Robotics—have robots performing productive work in commercial settings, marking a turning point for the industry.
Tesla's Optimus Gen 3 is the most widely deployed, with over 5,000 units operating across Tesla's own manufacturing facilities. The robots handle repetitive assembly tasks—inserting wiring harnesses, placing battery cells, and conducting visual inspections—at roughly 60% of human speed but with 99.2% accuracy and zero fatigue-related errors.
Figure AI's Figure 02 has taken a different approach, targeting logistics and warehousing. BMW's Spartanburg plant has 200 Figure 02 units performing bin picking, parts transport, and quality checks. The robots use a multimodal foundation model that lets them understand natural-language instructions from floor supervisors, reducing the need for pre-programmed routines.
Agility Robotics' Digit is finding its niche in environments designed for humans but unsuitable for traditional industrial robots—think narrow aisles, stairs, and mixed human-robot workspaces. Amazon has deployed 750 Digit units across three fulfilment centres, where they handle tote movement and shelf restocking.
The economics are compelling but not yet transformative. A humanoid robot costs $50,000–$150,000 and can replace roughly 0.5–0.8 FTEs depending on the task. At current prices, the payback period is 18–24 months for three-shift operations. As costs fall and capabilities improve, the economic case will strengthen rapidly.
For teams evaluating robotics AI models, Vincony's Deep Research can synthesise benchmark data across all major robotics foundation models—from RT-3 to GR00T—in a single session.