May 1, 2026 · 1h 12m · sourcery
Figure's First Full HQ Tour: From the Lab to the Factory Floor · Sourcery with Molly O'Shea
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of Sourcery, host Molly O'Shea takes an exclusive tour of Figure's San Jose headquarters and BotQ manufacturing facility with Founder and CEO Brett Adcock. The walkthrough showcases Figure's rapid evolution from early prototypes to mass humanoid production, powered by their proprietary Helix AI vision-language-action neural network.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Molly holds 17% of the talking time here. How this is scored →
Molly as informed peer 1.4 Guest teaching 2.8 Guest disagreement 0.1 Molly pushing back 0.3
speaking balance: gold is Molly, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 30:33 Brett refutes teleoperation rumors
Brett firmly dismisses skepticism around Figure's living room demonstration, emphatically denying any secret room or remote human teleoperation.
Hardest push from Molly ▶ 30:31 Molly raises teleoperation rumorsMolly directly challenges the demonstration's legitimacy by raising persistent industry rumors that humanoid demos are secretly teleoperated.
Biggest teaching moment ▶ 5:32 Combinatorial math of 40 motor positionsBrett educates Molly on the mathematics of humanoid control, showing that 40 motors with 360-degree rotation generate more discrete body states than atoms in the universe.
Molly holds their own ▶ 52:14 Molly cites peer robotics visitsMolly demonstrates her industry background by referencing first-hand site visits to advanced hardware peers including Applied Intuition, Skydio, Archer, and Anduril.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Molly as informed peer | Guest teaching | Guest disagreement | Molly pushing back | Why |
|---|---|---|---|---|---|---|
| Arrival at Figure HQ and Reception Desk Discussion | 1 | 2 | 0 | 0 | Molly opens the tour with basic exploratory questions about Figure's facilities and mission. Brett outlines the humanoid vision, fleet wireless docking capabilities, and their recent White House appearance. | |
| System Integration Lab Tour and Helix Neural Network Intro | 1 | 4 | 0 | 0 | Brett explains the scale of degrees of freedom across 40 actuators, demonstrating that 360^40 body states exceed the number of atoms in the universe and necessitate neural network control rather than written code. | |
| Helix AI Training Methods and Robot Workwear Aesthetics | 2 | 3 | 0 | 0 | Brett walks through the System Integration Lab's 'trust but verify' philosophy, explaining hardware durability, burpee recovery, and fabric workwear. Molly asks pragmatic questions about failure modes and joint softness. | |
| Live Physical Push Testing with Reinforcement Learning Control | 1 | 3 | 0 | 0 | Controls lead Moritz introduces the reinforcement learning balance stack before Molly physically shoves a robot. Molly asks about real-world defense mechanisms against vandalism. | |
| Sim-to-Real Transfer and General-Purpose Humanoid Philosophy | 2 | 4 | 0 | 0 | Brett articulates Figure's philosophy on general-purpose hardware and explains why Figure abandoned hundreds of thousands of lines of C++ inverted-pendulum code in favor of onboard vision-language-action neural networks. | |
| Sponsor Segment: Brex Financial Platform | 1 | 3 | 0 | 1 | Brett demonstrates Project Vulcan, showing how an RL controller handles an in-flight knee joint failure without collapsing. Molly playfully remarks that she anticipated a more dramatic physical collapse. | |
| Figure Organizational Teams and Engineering Disciplines | 1 | 2 | 0 | 0 | Brett details the roughly twelve engineering and operations divisions required to manufacture humanoids, drawing parallels to his prior experience building eVTOL aircraft at Archer. | |
| Autonomous Housekeeping Demonstration with Helix II | 2 | 3 | 1 | 3 | During an autonomous living room cleaning demo, Molly directly confronts Brett about industry rumors alleging teleoperation. Brett flatly refutes this, confirming the Helix policy runs onboard. | |
| Training Data Scale, Privacy, and Consumer Pricing Model | 2 | 3 | 0 | 1 | Molly presses on consumer data privacy and European regulatory hurdles. Brett outlines their data flywheels, anonymization protocols, and the intended car-lease style subscription pricing model. | |
| Campus Walk to BotQ and BMW Assembly Line Milestone | 1 | 2 | 0 | 0 | While walking between facilities, Brett discusses Figure's deployment at BMW assembling X3 vehicles, continuous 24/7 testing at The Grid, and internal physical security against IP theft. | |
| BotQ Factory Tour: Head Assembly and Sensor Calibration | 2 | 3 | 0 | 0 | Inside BotQ, Brett shows the sensor-laden head assembly and custom battery pack manufacturing line. Molly tests the battery's physical weight and inquires about single-pack versus distributed architecture. | |
| Assembly Stations and Real-World Embodiment Path to AGI | 3 | 3 | 0 | 0 | Brett explains why physical world interaction data could be the key catalyst for reaching AGI before digital systems. Molly brings outside context from her visits to defense and robotics factories like Skydio and Anduril. | |
| Multi-Sponsor Segment: VCX, Public.com, Merge, and Deel | 0 | 1 | 0 | 0 | Mid-roll sponsor reads followed by conversational walking banter regarding the campus facility choice and local employee commute patterns. | |
| Secret Design Studio: Evolution from Figure 1 to Figure 3 | 2 | 3 | 0 | 0 | In the private design studio, Brett reveals Figure 01 through 03, explaining how they achieved a 90% cost reduction by moving away from CNC machining and optimizing for serial manufacturing. | |
| Evolution of Hand Prototype Generations and Mechanics | 2 | 4 | 0 | 0 | Brett unveils Figure's unreleased generation-one tendon-driven hand and explains the Frankenstein forearm adaptation using foot motors, before walking through toe-articulation foot mechanics. | |
| Time Magazine Cover and Deadmau5 Concert Collaboration | 0 | 1 | 0 | 0 | Brett wraps up the studio tour by recounting Figure's Time Magazine cover feature and live robot dance performance alongside Deadmau5 at Red Rocks. |
Statements from this episode (34)
Adcock: Figure humanoid robots charge wirelessly at 2 kW with 5-hour runtime
“The robots can charge through their feet at two kilowatts. So basically, like the battery lasts about four to five hours, then we can charge basically for an hour and go back and do work again.”
Adcock: Figure placed the first humanoid robots in White House history
“We got a call asking to be basically the first humanoid group basically ever to put robots in the White House, and so last week we basically had the, you know, first robots in, humanoid robots in history there doing stuff, talking, greeting folks we were basic…”
Adcock: Figure has 500 total employees, 250-300 at HQ
“There's about 253 hundred people here, and we have 500 in the company.”
Adcock: Figure robots run entirely on Helix vision-language-action model
“You can't write code to make this work. So all of our robots here run on a neural network we call Helix. It's a vision language action model we designed here internally to tell the robot what to go do, to stay balanced, like how to move its joints basically fr…”
Adcock: Figure runs onboard neural network inference hundreds of times per second
“And we have to do that like a few hundred times a second from camera images and on a neural network that runs on board the robot.”
Adcock: Figure's humanoid has over 40 moving joints and in-house supply chain
“Like the robot has 40 plus moving joints. It's like a walking cell phone, self-driving car. All the supply chain is basically new. We've designed almost all of it.”
Adcock: Figure robots dock at 10-15% battery and swap within 30 seconds
“We dock it, like, yeah, around 10 to 15%, they'll go to dock, and then if we need another robot in, we'll sub them in off the dock, and you'll see here later, you know, like, logistics and other use cases that need, like, constant, like, 24 seven attention, th…”
Adcock: Figure's robot hardware is robust; most failures are software and AI
“We're at a point where the hardware's gotten, like, pretty, like, almost, like, very robust. Like, the, we can run them all day, every day, like we still, we see, we still see, like, hardware failures, but It's very few and far between. Most of our problems no…”
Adcock: Trying to code humanoid robot behaviors is a complete dead end
“You basically can code your way out of this, and I think that's a full dead end, or you can run an AI-first strategy in the market, and that's what we do here at Figure.”
Adcock: Figure robots run Helix AI inference onboard without network connectivity
“We run Helix on board. So they're actually loaded into GPUs on board and memory, and we run inference on board. Meaning if we if we lose internet connection, we can still do housework and logistics.”
Adcock: Project Vulcan prevents Figure robots from falling during leg joint failures
“We've been working on a project we call Vulcan here internally. That basically allows us to lose a single or even multiple joints in the legs and still not fall.”
Adcock: Figure designs over 100 custom PCBs for its robot
“We designed almost there's over a hundred PCBs that we design here on the robot that we do on that team.”
Adcock: Figure Helix AI trained on under one million hours of data
“We've probably had in Helix, like the, like, in, like, kind of base, pre, and mid-training Helix models that we're working on now have maybe, like, maybe, like, a little under a million hours of total data. And then we also do from mid-training and pre-trainin…”
Adcock: Figure demonstration robots run purely autonomously with onboard Helix AI
“This is not tele-operated.
For sure not tele-operated.
No secret room.
these robots are running purely autonomously from an onboard AI policy called Helix running onboard the robot in the torso.”
Adcock: Data is Figure's biggest blocker to large-scale humanoid deployment
“The biggest blocker for us now of going from where we're at today to, like large-scale deployment is data. We need, like, an enormous amount of data.”
Adcock: Figure's data collection will generalize across global environments
“Our hypothesis now is that we will see enormous amount of positive transfer from the data collection efforts we're doing now into pre-training across, like, basically any environment in the world.”
Adcock: Figure plans to lease home humanoids for $400-$600 monthly
“The in-home, like, we have, we're not selling right now to the home. We want to sell here like, in the near term, and we want to sell the robot, like, for, like, hundreds of dollars a month, and as, like, somewhere like a car lease maybe, like, four or five, 6…”
Adcock: Figure Deployed Robots for Six Months on BMW Body Shop Line
“We work at BMW, and last year we deployed for six months robots on the basically body shop factory line to build cars.”
Adcock: Figure-Assisted BMW X3 Is First Car Built by Humanoid Robot
“This is an X-III that we helped basically like we basically helped, the robot helped assemble it, and this was the first car in the world built by a humanoid robot that we're aware of.”
Adcock: Figure humanoid heads include 5G, thermal sensors, and IMUs
“Our heads have a, basically Bluetooth, like Wi-Fi, five G. They have camera systems on, on board. We have lights we have thermal systems we have an IMU.”
Adcock: Figure's robot uses a 2.25 kWh torso battery pack
“And then at the end, pops out basically a pretty heavy 2.25 kilowatt hour battery pack. ... So this so this battery will basically go right into the torso. It's one of the heaviest components we have.”
Adcock: Figure got its first robot walking within a year
“We have our figure one robot. It's kind of gnarly. It's got wires everywhere. It's our first generation. We had that walking before we were a year old, and we think it's probably one of the fastest times in human history.”
Adcock: Figure will ship humanoids onto its production line this year
“We're, we will be shipping robots, human, our humanoid robots into the production lines here this year”
Adcock: AGI might be achieved first in humanoid robotic embodiments
“So in some way I think we'll get to at or even beyond human level intelligence in these systems. It actually might be the case that we get to artificial general intelligence first in these embodiments.”
Adcock: Real-world interaction data is the missing piece to true AGI
“Because we're able to, like, this interaction data of, like, touching the world and seeing what happens through trial and error is, like it basically most of human intelligence is built this way, and I think this is, like, the last missing piece to get the tru…”
Adcock: Figure will reach full lights-out manufacturing where robots build robots
“At some point here, it'll just be like it'll just be full lights out manufacturing. We'll have like robots only building robots and sending out to the world. Robots will be getting into boxes themselves. Other robots will be boxing them up and we'll be shippin…”
Adcock: Figure's current humanoid robots cost well under $100,000 each
“So this was like hundreds of thousands of dollars and the robots we have now are like you know, well under a 100,000 dollars each.”
Adcock: Figure reduced robot costs by 90% between Figure 2 and 3
“We reduced the cost by about 90% between these two.”
Adcock: Figure in-house designed all motors and actuators for Gen-1 hand
“So our first-generation hand, you can see here, is a tendon-driven hand.
We designed all the motors and actuators here ourselves even the gearboxes.”
Adcock: Tendon-driven robotic hand was the wrong engineering choice for Figure
“And it ended up becoming ended up becoming like the wrong engineering choice.
And we ended up pivoting away from it really early.”