What he is (one line): founder and CEO of Figure AI (humanoid robots) and Hark (a personal AI assistant that uses computers for you); before that he founded Archer Aviation (eVTOL) and Vettery (a hiring marketplace, sold 2017). On Moonshots with Peter Diamandis he is a "friend of the pod", the guest of the annual Figure site visit, and the opener booked for the Abundance Summit.
Why it is in this archive
Gate: named in 28 held transcripts (regex count for "Adcock", 2026-11-17), and the guest of Moonshots #unnumbered. Promoted 2026-11-17. Company facts live on Figure AI and Hark; this page holds his own claims and how they moved.
His model, step by step
Built from his own words on Moonshots #unnumbered. Steps are his view, not checked claims.
- Hardware first. "you have to build like really good functional hardware… you can't build like AGI on a robot without really good hardware" (
0:11:18). Assumes the capital to build your own robot. - Then an AI recipe that runs the whole body. "you need to understand what the knee and hip and feet are doing as it relates to fingers and hands… The same exact time stamp" (
0:16:07). Helix 1 (upper body) to Helix 2 (full body) took "a year" (0:15:15). - Then scale data and compute. "our input is data and compute and our output is getting robots to do more useful work" (
0:19:12). He calls chapters 3 and 4 "scaling intelligence and scaling robots". - Build the parts yourself when the market can't. The first actuator they bought was "literally three times too big and three times overweight"; it needs to cost "tens of dollars at the most" (
0:38:30). - Ship only when it works. "we don't want to be teleyoperating the home" (
0:04:59); "if it comes out and doesn't work really well, you're kind of dead" (0:10:02). - Robots, not people, should take orders from ASI. "the humanoid should be doing this work and not humans" (
1:18:06).
What an operator can do with this
- Supply the parts he says do not exist at the size and price: small, cheap actuators, harnessing, thermals. The FCC's 28 Jul 2026 rule on foreign-made robots pushes US makers to onshore parts (Fortune, 23 Oct 2026). Caveat: Figure says it does all of this in-house.
- Don't plan around home humanoids before 2028. His own sale date is "27 28" (
0:06:34).
How his claims moved
| when | what he said | source |
|---|---|---|
| Jan 2026 (EP229, as read back by Dave on 09-19) | "By the end of 2026, we will have humanoid robots performing unsupervised multi-day tasks in homes they've never seen before." | Moonshots #292 1:34:26 |
| Nov 2026 | "I hope to see in the next sub six months perhaps even sooner we can put a robot into any home and do fairly large like long horizon tasks… over like hours and hours"; "I'd be very shocked if we exit 2027" without "this whole picture"; consumer sale "27 28" | Moonshots #unnumbered 0:08:38, 0:24:03, 0:06:34 |
| Jan 2026 (EP229) | OpenAI split: "my our team just ran circles around them" | backfill EP229 0:12:17 |
| Nov 2026 | "we fired them like two years ago"; "they never showed up at the office"; "10 to one 100 to one progress" | Moonshots #unnumbered 1:59:29–2:00:34 |
| Nov 2026 | Tesla: "I personally haven't seen it much out of Tesla in two or three years"; closest competitor "certainly China today" | Moonshots #unnumbered 1:28:10, 1:29:13 |
| Nov 2026 | Hark vs Figure: "I bet Hark is bigger than Figure in sub six months… Maybe sub 90 days" | Moonshots #unnumbered 1:52:02 |
Claims that fail or are unverified
- 🔴 "low tens of millions of dollars per hour to put a human in space to do work" (attributed to "one group",
0:56:10). NASA's published commercial crew-time price is $130,000 an hour (NASA). - ❓ "September 2026, the world ran out of compute" (
1:45:25): his experience, no public record. - ❓ Hark "already bigger than figure from revenue standpoint" (
1:52:12): neither company discloses revenue.
Related
Figure AI · Hark · OpenAI · Tesla · Unitree · Peter Diamandis · Dave Blundin · Alexander Wissner-Gross