Figure AI Says Its BotQ Factory Now Builds One Humanoid Per Hour, a 24x Ramp in Under 120 Days

Overview

Figure AI says its BotQ manufacturing facility in California has scaled from producing one Figure 03 humanoid per day to one per hour over the past 120 days, a 24-fold throughput increase that the company describes as the operational backbone for its plan to ship tens of thousands of humanoid robots a year. The announcement, posted on the company’s blog on April 29, 2026, comes alongside an update to Figure’s Helix AI stack that lets the robot navigate stairs and uneven terrain straight from simulation, according to Figure.

In the same period, Figure says it has delivered “over 350 of our third generation humanoid robots,” produced “over 9,000 actuators across more than 10 distinct SKUs,” and shipped “over 500 battery packs” off a line that now reports a 99.3% first-pass yield, per Figure. Interesting Engineering reports the same figures, framing them as evidence that BotQ has cleared the gap between research prototype and high-volume assembly.

What Figure Disclosed

The core metric is throughput. Figure says it cut its production cycle time “from 1 Figure 03 per day to 1 per hour - a 24x throughput improvement in under 120 days,” according to the company’s production update. Interesting Engineering describes the same milestone: “Scaled from one robot/day to 1 robot/hour” in 120 days, the outlet writes.

The facility, BotQ, is what Figure built last year specifically to mass-manufacture its humanoids. The first-generation line is rated for up to 12,000 robots per year, and the company has said its supply chain “can easily scale to 100,000 robots or 3,000,000 actuators in the next four years,” according to Figure’s BotQ announcement. The same document identifies the Figure 03 as “our production robot built for affordability and high-volume manufacturing.”

Quality data accompanies the throughput claim. Figure reports an end-of-line first-pass yield “now over 80% and improving weekly” and a battery production line at 99.3% first-pass yield, per its blog post. Each robot is run through “over 80 functional verification tests,” the company says, and the assembly line operates across “over 150 networked workstations” with “more than 50 in-process inspection points.” Interesting Engineering cross-confirms that the verification suite includes “stress and burn-in exercises such as squats and jogging.”

A New Locomotion Capability, Trained Entirely in Simulation

Alongside the production update, Figure introduced a new capability for Helix’s System 0, the lowest level of its control stack. The company says it has added “perception-conditioned whole-body control” that lets Figure 03 navigate stairs, ramps, and uneven terrain using only its onboard stereo cameras, with no task-specific code, according to Figure.

Notably, Figure says the policy was trained entirely in simulation and transferred to the physical robot “zero-shot - no real-world fine-tuning,” per the same post. “The same network weights that learn to climb procedurally generated staircases in sim now traverse real-world stairs on the robot,” the company writes. Interesting Engineering describes the same upgrade as “visual integration from stereo cameras” feeding the System 0 controller.

For the broader humanoid-robot field, the sim-to-real transfer claim matters because it points at a development loop where new motor skills can be added without manually retuning each behavior on hardware — an approach Figure is betting will scale faster than traditional robotics programming.

Context: A $39 Billion Valuation Behind the Ramp

Figure is spending against a balance sheet it built last year. In September 2025, the company announced it had “exceeded more than $1 billion in committed capital” in a Series C financing round at “a $39 billion post-money valuation,” according to a PR Newswire release. The round was led by Parkway Venture Capital with participation from Brookfield Asset Management, NVIDIA, Macquarie Capital, Intel Capital, Align Ventures, Tamarack Global, LG Technology Ventures, Salesforce, T-Mobile Ventures, and Qualcomm Ventures, per the same release.

Brett Adcock, Figure’s Founder and CEO, framed the raise around exactly the two systems now being scaled: “This milestone is critical to unlocking the next stage of growth for humanoid robots, scaling out our AI platform Helix and BotQ manufacturing,” per PR Newswire.

What We Don’t Know

Figure has not disclosed how many of the 350+ delivered Figure 03 units are with paying customers versus internal pilots, nor has it broken down the monthly production curve in absolute units. The company’s BotQ announcement describes a plan to use “Figure humanoid robots” themselves “in our manufacturing process to build other humanoid robots” within the year, but does not specify how much of current output already comes off lines that include robotic assistance.

The yield numbers — 80% end-of-line, 99.3% battery — are self-reported and have not yet been independently audited. Figure has not published failure-mode data or warranty-return rates, and 80% first-pass yield is well below the >99% benchmarks common in mature consumer-electronics manufacturing, leaving room for further improvement as the line matures.

Figure also has not confirmed how the new sim-trained stair-climbing policy performs on the wide variety of surfaces a factory floor or warehouse actually presents. The company describes successful transfer to “real-world stairs” but has not released benchmarks against alternative locomotion approaches, per the production update.

Why It Matters

The humanoid-robotics industry has spent two years collecting demo videos. What it has lacked is evidence that anyone can actually build the machines at scale. A 24x throughput increase in 120 days, with first-pass yields above 80% and a battery line near 99%, is the kind of operational data that distinguishes a manufacturing program from a research project — and is the headline number Figure’s investors, customers, and competitors will all be benchmarking against for the rest of the year.