Big claims about machines and work always arrive with a number attached, because the number is what does the persuading.
Nobody remembers the argument. Everybody remembers the figure.
That is why automation forecasts land differently than interest rate forecasts. A rate call you can check in six weeks. A claim about the shape of the world economy in 2036 sits out there for years, doing quiet work on how you invest, what you tell your kids to study in college, and whether you assume your job survives the decade.
The baseline right now is deliberately boring. Global growth is running near 3% a year, and outside of shocks it has for most of the past decade. Your 401(k) projections, your employer’s hiring budget and your local housing market all rest on some version of that assumption.
Boring baselines are useful. They are also the first thing to get blown up when someone with capital and an audience publishes a different number.
That happened again on Sept. 9. “AI+robots will more than double the global economy in less than 10 years,” Elon Musk posted on X, the platform he owns.
I have read enough forecasts to know the claim itself is never the interesting part. The growth rate the claim quietly requires is. So I ran the timeline against the International Monetary Fund’s own projections, and the gap is wider than one sentence on social media lets on.
Why the robot economy math matters to your money
Global gross domestic product is projected at roughly $126 trillion in 2026, according to the International Monetary Fund. Doubling that means adding a second $126 trillion of annual output by the mid-2030s.
Run the compounding and the requirement gets specific. A clean double in ten years takes 7.2% growth every year. Doing it in less than ten pushes you to about 8% a year over nine years, or 9% over eight.
More Robots:
- BMW doubles down on humanoid robots after a U.S. test run
- Tesla’s Optimus robot plan hits major snag
- Elon Musk has a shocking message on AI and robots
Now the baseline. Global growth is projected at 3.0% in 2026 and 3.4% in 2027, according to the IMF. At 3%, the world economy doubles in roughly 23 years.
My analysis puts Musk’s requirement at about 2.5 times the planet’s current speed, sustained annually, with no recession in the window.
There is one reading that makes the claim far less wild, and almost nobody states it. Musk did not specify real or nominal.
In dollar terms, global output is projected to climb from about $118 trillion in 2025 to about $126 trillion in 2026, according to IMF projections. That is nearly 7% nominal growth, which is within striking distance of the 8% the timeline needs.
Nominal growth counts inflation and a weaker dollar as progress, so it is not the same as the world actually producing twice as much. Read the claim as real output and it is extraordinary. Read it as headline dollars and it is close to the current trend.
That gap is not academic. Wage growth, corporate earnings and the returns sitting inside your index funds all key off aggregate output.
Related: AI Could Blow a Hole in the Federal Budget
Put it in dollars. At 3%, global output reaches about $170 trillion by 2036, adding roughly $43 trillion in annual production. Musk’s version adds $126 trillion over the same stretch, close to three times as much new output for wages, profits and tax receipts to draw from.
That is the real stake in the argument. Musk has made versions of it before, including his case that mass automation would force governments to start handing out cash.
What Musk actually told the G20 about robots
The post was a compressed version of a much longer pitch. “AI will probably increase the global economy by 20% to 30%,” Musk said during a virtual appearance at the G20 Innovation Ministerial in Chapel Hill, N.C., on Sept. 1, according to CNBC.
He put that figure at $20 trillion to $30 trillion in added annual output.
The robot half is where he got specific. There will be at least one billion robots within a decade, and those machines will be “at least five times the output of a human,” he told technology ministers, according to the Daily Tar Heel. He described it as a call he would put serious money behind.
He also handed ministers the constraint. Industry consensus points to a power shortfall of roughly 15 gigawatts in 2027 because AI chip production is climbing far faster than electricity supply can follow, he said, per the Daily Tar Heel.
Here is what the shipment data actually shows:
- China is expected to ship 50,000 humanoid robots in 2026, up from a January forecast of 14,000, according to CNBC’s report on Morgan Stanley (MS) research.
- Chinese annual humanoid shipments are forecast to reach 446,000 units by 2030, according to CNBC.
- China’s humanoid market is valued near $2 billion this year and is projected at $15 billion by 2030, according to CNBC.
- Global output is projected at roughly $126 trillion for 2026, according to the IMF.
What struck me running those figures is how violently the curve has to bend. Reaching one billion units within a decade from a 2026 base near 50,000 requires annual shipment growth of roughly 185%, with something like 650 million robots coming off assembly lines in the final year alone.
Morgan Stanley’s most aggressive published number for 2030 is 446,000 units. That is roughly 1,500 times smaller than the final-year pace the timeline demands, and it comes from the bank that has already revised its own forecast upward twice this year.
What the robot economy means for your paycheck
The forecast is doing real work on the labor market long before any of it arrives. Artificial intelligence was cited in 116,175 announced U.S. job cuts through August, about 22% of all cuts this year, according to Challenger, Gray and Christmas.
That is the leading stated reason for layoffs in 2026, and it is happening while global humanoid shipments are still counted in the tens of thousands. The robots are not taking those jobs yet but the expectation of the robots is.
For Tesla (TSLA) shareholders, the stakes run more directly. Optimus production sits among the milestones attached to Musk’s compensation package, which ties an enormous payout to targets the company has not yet hit.
So watch a different number than the one Musk handed you. Annual humanoid shipments and the electricity available to run them are the two figures that decide whether 8% growth is a forecast or a slogan.
If shipments triple again next year and utilities keep signing data center contracts, the fast scenario stops being rhetorical and current AI valuations start looking defensible. If they do not, the boring 3% baseline holds, and a lot of portfolios are priced for an arrival date that keeps sliding.
Musk’s forecasts are free to publish. The positioning they encourage is not.
Related: AI agents are quietly rewriting how the internet works
