> ## Content Index
> Fetch the complete content index at: https://longtalkshort.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# A Race Is Not a Delivery Date: Brett Adcock on Humanoids, China, and the Home Robot
- URL: https://longtalkshort.com/a-race-is-not-a-delivery-date-brett-adcock-humanoids/
- Published: 2026-11-19T04:13:45.000Z
- Updated: 2026-11-19T04:13:45.000Z
- Description: Figure CEO Brett Adcock tells Peter Diamandis and Dave Blundin the humanoid race is real, especially with China, and that a capable home robot should be visible as a picture by the end of 2027 — a forecast, not a ship date. Interests disclosed. Not investment advice.
- Author: Truman
- Tags: Economics, Brett Adcock, Figure AI, humanoid robots, Peter Diamandis

If you have been waiting for a date when a humanoid robot will simply walk into the house and get on with the day, Brett Adcock will not give you a clean one. On Peter Diamandis and Dave Blundin’s Moonshots conversation, the founder and CEO of Figure AI does give you a race, a four-stage map, and a set of launch conditions he says his company will not relax. The race, as the hosts frame it, is with China and with Tesla. Adcock’s answer is sharper on the first than on the second. He says he has not “seen it much out of Tesla in two or three years.” China, he says, has been “dancing.” “We just got to go beat them. It’s a real race here.” Those are a competitor’s impressions, not a scoreboard.

*This is the Moonshots episode with Peter Diamandis and Dave Blundin on Figure’s robotics timeline, at* [*youtube.com/watch?v=E864U582FVY*](https://www.youtube.com/watch?v=E864U582FVY&ref=longtalkshort.com)*. Adcock is founder and CEO of Figure, and he also talks about Hark, another of his companies. Blundin is founder and general partner of Link Ventures, which is also named as a sponsor, alongside Diamandis’s Abundance Summit. Diamandis points to his own ventures, including a report and Fountain Life. The show’s footer already says the views are personal opinions and not financial, medical, or legal advice. This article follows that. Auto-generated captions were used for wording; obvious names have been corrected from the official title and description.*

**Disclosure:** Adcock is the CEO of the company that benefits if his timeline is believed. Blundin runs Link Ventures, a named sponsor of the episode, and Diamandis promotes his own ventures. We have not established whether Link Ventures or the hosts hold a stake in Figure. Treat the forecasts as a founder’s claims.

*In short: Adcock maps a four-stage humanoid race — hardware, an AI recipe, then scaling intelligence and scaling robots — and treats China more than Tesla as the rival. He says Figure will not ship a home robot until it can do real work without teleoperation, and that he would be “very shocked” if the capability picture for the home were not visible by the end of 2027\. That is a forecast, not a ship date. This is not investment advice.*

Adcock’s map of the “humanoid race” has four stages, or “chapters.” First, “really good functional hardware.” Second, “an AI recipe.” Third and fourth, “scaling intelligence and scaling robots.” He is, unsurprisingly, more bullish on where Figure sits than on the highlight reels coming out of China. The dancing robots you see, he says, are often “deep learning RL controllers” trained from people in mocap suits, then replayed “open loop.” “The tech stack underneath it is totally contrived,” in his words: “not a real neural net that’s learning on the fly.” That is a pointed competitive claim. It may be fair about some demos. It is not a complete audit of an entire country’s robotics programs, and it should be read as his brief, not as a field report.

![A four-chapter path — hardware, an AI recipe, scaling intelligence, scaling robots — with a fork at the end labeled ‘commercial tasks’ and ‘the messy home.’](https://storage.ghost.io/c/d7/93/d793216c-cf13-441f-bdae-4f3aef998c0e/content/images/2026/11/cover-330.png)

His race map is a sequence of problems, not a ship date. The home, in his telling, is a different and less specified problem than the factory floor.

## 1\. What he will call a home launch, and what he will not

He splits the work into two tracks: commercial robots, and robots in the home. Commercial work, he says, is in some ways easier because the task can be written down — and harder because speed and reliability are unforgiving. The home is the reverse. You do not need factory tempos, but “every home is different,” and the jobs “change quite dynamically.” Kids and dogs are in the way. Figure will ship, he says, when the machine can “walk into the home, talk to you, be able to do real work,” without being teleoperated, and without stopping every time it is unsure — “an Apple-style quality product.” He thinks other firms will ship sooner, and that those earlier devices will fail the way some much-hyped AI gadgets failed. That last comparison is a judgment, not a result.

Pressed on when we will see a robot that can drop into any home “for like 10 hours and it can do work,” he describes “stair steps” over the last three years, including “zero shot” behavior in new rooms. “We’ve arguably seen some of that this year in 2026,” he says. “I’d be very shocked if we exit 2027 \[and\] we haven’t seen this whole picture.” That is not a promise that a product will be in living rooms by the end of 2027\. It is his forecast that the capability picture should, if the trajectory holds, be visible by then. He also talks about a Figure 4 as an “iPhone one moment” after a “flip phone era.” Those are product teases from the person who would benefit if you believe them.

On price, a host asks whether he is “still tracking towards a $30,000 price tag, $300 a month type of lease.” Adcock says the home “will have to make the robot really affordable,” and talks about “some sort of like $300 to $500 a month lease,” “10 bucks a day,” “under a buck an hour.” That is a target, not a current offer. Further out, he “arguably see\[s\] a path” — “a really hard path,” if they solve the chapter-three problems they are working on — to “a million units a year out of a facility fairly quickly.” That sits on the far side of the same conditional: the intelligence and manufacturing loops have to work first.

![A timeline with no hard ship date: a 2026 step labeled ‘some of that already,’ a 2027 band labeled ‘shocked if we have not seen the 10-hour picture,’ and a later, fainter mark for an affordable home lease.](https://storage.ghost.io/c/d7/93/d793216c-cf13-441f-bdae-4f3aef998c0e/content/images/2026/11/fig1-99.png)

The only date he is willing to be shocked about is a capability picture, not a product sitting in the hallway.

## 2\. The factory loop, and the claims to hold lightly

The scaling chapter is where the conversation gets most speculative. A host asks about BotQ and “RSI in the physical world.” Adcock’s useful correction, borrowed from a Gigafactory intuition, is that the near-term loop is not robots literally winding their own actuators. It is “robots controlling machines that build robots.” If that loop tightens, production can move faster than a headcount would suggest. If it does not, the million-unit path stays a sketch. He also mentions large compute plans, including a deal he describes as 100,000 GPUs and “three plus billion dollars” for the second half of 2027\. Figure’s own [announcement of a partnership with Nscale](https://www.figure.ai/news/figure-and-nscale-sign-strategic-partnership?ref=longtalkshort.com) describes up to 100,000 NVIDIA Vera Rubin GPUs, targeted for initial deployment starting in the second half of 2027, and an initial commitment of $3.5 billion of compute. That corroborates the deal’s existence and size as a commitment and a plan, not capacity delivered or robots built with it.

Tesla is in the title, and almost offstage in the conversation. The people with a direct stake in the claims are already in the room: a founder selling a timeline, a venture host, and a show with its own products and sponsors.

![A factory loop: a humanoid at a console controlling a machine that builds another humanoid, with a caption that this is ‘robots controlling machines,’ not robots building themselves by hand.](https://storage.ghost.io/c/d7/93/d793216c-cf13-441f-bdae-4f3aef998c0e/content/images/2026/11/fig2-99.png)

The important distinction in the scaling talk is that one. A tighter control loop is not the same thing as a self-replicating machine.

## Weighing it: what holds up

This is a founder interview on a show that celebrates abundance, recorded at the company’s site, and it should be read as advocacy with numbers attached. Adcock’s characterizations of Chinese demos, Tesla’s visibility, home-readiness, lease prices, and unit paths are his claims, most of them conditional, some of them promotional; the compute deal is corroborated by Figure’s own announcement, but only as a commitment. The 2027 line is about seeing a capability picture, not about taking orders. Blundin’s firm and Diamandis’s ventures are commercial interests, and Link Ventures is also a sponsor; we have not established whether they hold stakes in Figure. The episode already says this is not financial advice. Auto-generated captions are close, not official.

**Related reading:** [Robots by the Hour](https://longtalkshort.com/tutor-intelligence-josh-gruenstein-robots-by-the-hour-a16z/)

## Long Talk Short

- Adcock frames a four-stage race — hardware, an AI recipe, then scaling intelligence and scaling robots — and treats China, more than Tesla, as the competitor he is watching.
- His account of Chinese “dancing” demos is a competitive brief: mocap, open-loop replay, not, in his words, a net that is learning on the fly. It is not a census of the field.
- He will not, he says, ship a home robot until it can walk in, talk, and do real work without teleoperation. Other firms, he thinks, will ship earlier and regret it.
- He would be “very shocked” if, by the end of 2027, we have not seen the picture of a robot doing hours of work in an unseen home. That is a capability forecast, not a launch date, and the affordable lease he describes is a target.

**The one thing to remember:** A robot that can do a factory task you can write down is not the same as one that can walk into any home — and a race is not a delivery date.

*Independent summary; not affiliated with Figure, Moonshots, Link Ventures, or any person named. Not investment advice.*

Source: Moonshots with Peter Diamandis, episode with Brett Adcock and Dave Blundin, on YouTube (linked above); Figure’s own announcement is linked in the article. Independent summary; not affiliated with the show or its guest.