Figure ramps Figure 03 production at BotQ to one robot per hour
Figure AI has achieved a 24x throughput improvement in under 120 days at its California manufacturing facility, delivering over 350 Figure 03 units

Figure AI has increased production of its Figure 03 humanoid robot from one unit per day to one per hour in under 120 days at its BotQ manufacturing facility in California, delivering over 350 units in the process. The ramp was powered by dedicated assembly lines running on custom manufacturing execution software across more than 150 networked workstations, with hundreds of suppliers qualified against incoming inspection criteria and 50 discrete quality checkpoints across the build process; first-pass yields are above 80%, and more than 9,000 actuators have been produced across over ten product variants.
Source: Figure
Figure 03 was designed from the ground up for high-volume manufacturing rather than for research or demonstration purposes. To achieve production-rate economics, Figure switched from CNC machining to tooled processes including injection moulding, die-casting, metal injection moulding, and stamping; parts that previously spent over a week on a CNC machine can now be manufactured in under 20 seconds. The capital cost of tooling is high but the unit economics improve sharply with volume, which is the logic behind building manufacturing capacity before demand is fully confirmed. BotQ's first-generation line is designed for up to 12,000 humanoids per year, with a goal of 100,000 robots over the next four years.
Figure also released a new perception-conditioned whole-body control capability for its AI model at the time of the production update, allowing Figure 03 to navigate stairs, ramps, and uneven terrain using onboard stereo camera perception without task-specific programming or operator intervention. The capability was trained through reinforcement learning in simulation and transferred directly to physical hardware without additional real-world tuning. Figure has described the expanding fleet as a data generation asset; each deployed unit produces training data that feeds the next iteration of Helix, the company's proprietary vision-language-action model.
Figure AI was founded in 2022 by Brett Adcock, who previously co-founded Vettery, an AI recruiting platform acquired by Adecco for approximately $100 million, and Archer Aviation, an electric vertical takeoff and landing company that listed on the NYSE at a $2.7 billion valuation. Before starting Figure, Adcock visited the bipedal robotics lab of Aaron Ames at Caltech, a researcher in walking mechanics who was himself a student of Marc Raibert, the founder of Boston Dynamics. Adcock assembled a founding team from Boston Dynamics, Tesla, and Apple and seeded the company with $100 million of his own capital.
Figure raised $70 million in Series A funding in May 2023 at a $500 million valuation, then $675 million in Series B in February 2024 at a $2.6 billion valuation, with Microsoft, OpenAI Startup Fund, NVIDIA, Amazon's Industrial Innovation Fund, Jeff Bezos, Parkway Venture Capital, and Intel Capital participating. The Series C, led by Parkway Venture Capital in September 2025, exceeded $1 billion in commitments at a $39 billion post-money valuation, a 15x increase from the Series B mark in 18 months. Total funding across all rounds is approximately $1.9 billion.
The OpenAI partnership, announced alongside the Series B, ended in early 2025. Adcock said the collaboration produced no meaningful value; Figure's internal team moved faster than OpenAI's engineers, who were rarely present for real-world testing. The decision to end the relationship accelerated when Adcock learned OpenAI was restarting its own robotics programme, at which point Figure stopped sharing any information with the OpenAI team and developed its Helix VLA model entirely in-house. The breakup was a vertical integration decision as much as a competitive one; owning the full stack from hardware to AI model is now the central design principle at Figure.
Figure 02 ran at BMW's Spartanburg, South Carolina facility for eleven months in daily ten-hour shifts, loading over 90,000 parts across more than 1,250 runtime hours and contributing to over 30,000 X3 vehicles. Reliability findings from the forearm and wrist electronics in that deployment were carried directly into Figure 03's redesign. The company became revenue-generating in January 2025 with its first commercial Figure 02 deployments, operating under a Robot-as-a-Service model priced at approximately $1,000 per robot per month.
The BotQ ramp is a manufacturing milestone; it demonstrates that Figure can produce Figure 03 at increasing volume with rising yield. Adcock has also stated a goal of full supply chain decoupling from China by summer 2026. No commercial customer deployments for Figure 03 at scale have been disclosed; the BMW relationship covered Figure 02, and the current Figure 03 fleet is primarily in internal operation and testing.
Have a robotics update Korthos should review? Send news, deployments, product releases, funding rounds, research, or media to tips@korthos.xyz or reach out on X at @agkorthos.