Skip to content

Moonshots Recap: Figure AI's Brett Adcock on Home Robots, Helix 3, Figure 4 and the Race With China and Tesla

Key takeaways and a timestamp guide from the Moonshots robotics deep-dive with Figure AI founder Brett Adcock (Nov 16, 2026): when humanoid robots might reach homes, his four stages of the humanoid race, Helix 3, a hinted Figure 4, pricing, safety, robots in space and scaling to a million robots a year.

By the GNOUP team Β· Β· 7 min read

The newest full-length video from the Moonshots podcast on Peter H. Diamandis's official YouTube channel, "Robotic Deep-Dive: Beating China, the Race Against Tesla, & New Robotics Timeline With Brett Adcock", was published on November 16, 2026 and runs about two hours and 12 minutes. Unlike the regular numbered news episodes, this is a guest deep-dive recorded at Figure AI's headquarters: Peter Diamandis and Dave Blundin sit down with Brett Adcock, Figure AI's founder and CEO (he previously founded Archer Aviation and Vettery), for what they call their annual catch-up. They cover when humanoid robots could work in homes, how Figure trains its robots, competition from China and Tesla, price, safety, robots in space, and how fast production could grow.

This is an independent summary written by GNOUP. It is not affiliated with, sponsored by or endorsed by Moonshots, Peter Diamandis, Figure AI or anyone who appears in the episode. The views, claims and predictions described below belong to the speakers, not to GNOUP. Figure AI is a private company, and statements about its plans, products and numbers come from its CEO and have not been independently verified. Nothing here is investment, financial or legal advice, and nothing here is a recommendation to buy or sell anything.

Key takeaways

  • More robots than people on campus. Adcock says Figure now has more robots than employees at its headquarters, crossing over about four or five months ago, with well over a thousand robots on campus and some deployed elsewhere. Visitor robots greet guests, use doors, stairs and elevators, and talk with people in real time. His stretch goal is a visit where you never need to talk to a human to get checked in and shown around.
  • Two tracks: work sites and homes. He describes commercial jobs as well-defined tasks with tough speed and performance targets, and homes as the opposite: slower work, but every house, appliance and request is different. He sees both as mainly an intelligence problem, and says Figure will not launch a home robot until it is an "Apple-style" quality product that works without remote human control, not an early device that disappoints.
  • A hoped-for home milestone. Adcock says he hopes, without promising, that within about six months Figure can put a robot in any home and have it carry out long, hours-long tasks from start to finish. Selling to consumers would come later, once safety and reliability are proven.
  • Four chapters of the humanoid race. His framework: first, build excellent, reliable hardware; second, find an AI recipe that runs whole-body tasks end to end; third and fourth, scale up the intelligence and then scale up the number of robots. He argues few groups besides Figure have shown the second chapter, and that many impressive-looking demos rely on table-mounted arms, scripted moves or remote control.
  • Helix 3, a "baby AGI moment." The hosts were shown an unreleased first checkpoint of Figure's next robot AI model, Helix 3. Adcock calls it a "baby AGI moment" for him while saying it was still undertrained and can't do some things well, and says much larger versions are already being evaluated. He says Helix runs fully on board the robot, from planning down to motor control.
  • Figure 4 teaser. Adcock calls today's humanoids the "flip phone era" and says Figure 4 could be the iPhone moment, cheaper and more capable, with much-improved hands. He claims a Figure 4 hand and a human hand look almost identical once both are wearing gloves, and that Figure 4 will make Figure 3 look like "a caveman."

Newsworthy insights

These are claims, predictions and opinions from the episode. GNOUP has not verified them.

  • Price target. Asked whether Figure is still tracking toward roughly a $30,000 robot, Adcock says home robots will need to be affordable, something like a $300 to $500 a month lease. Peter frames that as roughly ten dollars a day and argues robots working at about a dollar an hour would be strongly deflationary.
  • The China question. Adcock says much of the robot dancing seen in videos from China uses motion-capture training or pre-programmed moves, which he says is very different from a robot doing useful, unscripted work. He also says Figure has moved away from parts sourced from China. On Tesla, he says he personally hasn't seen much from Tesla's robot program in two or three years. These are a competitor's views.
  • Compute is the big bill. He says most of Figure's compute spending goes to training, that it spends nine figures a year on compute, and that it signed a deal with Nscale for about 100,000 next-generation Nvidia GPUs for the second half of 2027, which he estimates at well over 100 times its current training compute.
  • Safety and security. Adcock describes layers of protection, from built-in rules stored in the robot's memory to a system-level safety design, and separates physical safety from "semantic" safety, meaning the robot understanding what it should and shouldn't do. He says a sizable in-house cybersecurity team works on keeping robots from being hacked.
  • Humanoids in space. He says Figure has talked with groups about space work and was told it can cost low tens of millions of dollars per hour to have a human astronaut do work in space, which he sees as a natural job for humanoids.
  • A million a year. Adcock says he sees a hard but possible path to producing about a million robots a year from one facility, if Figure cracks the "chapter three" intelligence problems it is working on now. Dave describes factories as robots controlling machines that build more robots, a loop he calls exponential.

Watch the episode

Robotic Deep-Dive: Beating China, the Race Against Tesla, & New Robotics Timeline With Brett Adcock on YouTube (official channel "Peter H. Diamandis," published November 16, 2026, about 2 hours 12 minutes)

Episode guide

These chapters come from the video's own description (titles lightly reworded). Each link opens the video at that point.

About this summary

This recap was prepared from the video's title, description and chapter list on YouTube and its automatic transcript. Automatic transcripts can garble names and numbers, so we kept figures to the ones stated clearly and attributed them to the speaker. We left out sponsor segments and offers mentioned in the episode and its description, unverified remarks about other companies' products, and side banter. GNOUP has no connection to Figure AI or to any product, offer or company mentioned, and does not promote any of them. For the exact words and context, watch the episode. This article is general information only. It is not investment, financial or legal advice, it does not recommend buying or selling any security, and it is not a substitute for advice from a licensed professional.