Elon Musk told G20 innovation ministers this week that ten years from now there will be well over a billion humanoid robots. He called that a conservative estimate. He said he would put serious money on it.
We already put a number on this. In our on-the-record forecast we said 332,500,000 robots within ten years. That is less than half of his figure, and we still think we are the ones being honest about the calendar.
Key takeaways
- Musk's G20 call: well over one billion humanoid robots in ten years.
- Ours, already published: 332.5 million — less than half of his number.
- He is usually early on every product he launches. The dates slip. The products still tend to win.
- The number we will defend is the one in the middle: hundreds of millions of robots in ten years, and we think that part is certain.
Put the two forecasts on the same table
The remarks were made by video to the G20 Innovation Ministerial in Chapel Hill, North Carolina, at the start of September 2026. The quote, as transcribed, is simple: “Ten years from now, I would say there are well over a billion humanoid robots.” He did not give a Tesla-only shipment number. He was talking about the world's fleet.
| Elon Musk (G20) | RobotTesters | |
|---|---|---|
| Ten-year count | Well over 1,000,000,000 | 332,500,000 |
| How he/we framed it | "Conservative" | On the record, and we are sticking to it |
| Versus the other number | About 3× ours | Less than half of his |
The gap is not a rounding error. One billion is three times 332.5 million. If you take his figure and cut it in half, you still land above us. That is how optimistic the G20 number is. We are not moving our call because he said a bigger one this week.
He is early on every product. The products still win.
This is the Musk pattern, and anyone who has watched Tesla for a decade already knows it. Full Self-Driving. The Cybertruck. The robotaxi. Optimus itself. The number he names on stage is the number the product eventually wants to be, not the number the factory hits in the year he said. The calendar stretches. Then the product arrives anyway, and once it arrives it tends to win the category he pointed at.
That is why we do not treat the billion as a joke, and we also do not treat it as a shipment plan. It is the slide. The factory is slower than the slide. We have been writing that about Optimus for a while, and nothing in a G20 video changes the manufacturing math. A billion humanoids in ten years would be one of the fastest build-outs of a new machine in history. We do not think the world gets there on that clock.
The dates he publishes are almost always too early. The products he bets on almost always get there. Those two facts can sit in the same sentence.
China already treats this as real. Most places still don't.
So here is our call, again, in one sentence. The real number in ten years will be less than half of Musk's billion — around the 332.5 million we already wrote down — and that is still hundreds of millions of robots.
A lot of people, in a lot of countries, still treat that sentence as something that cannot happen. China does not. China is already building the factories, the school programmes, the supply chain. The disbelief is mostly elsewhere. We do not share it.
We think hundreds of millions of humanoid robots in ten years is not a maybe. It is the thing that happens even if Elon is three times too high on the clock. We said so in the 2035 forecast, we said so when we compared our production numbers with Morgan Stanley, and we are saying it again now that he has put a billion on a G20 stage. Check back in 2036. The count will settle the argument better than any keynote.