Elon Musk has spent the past two weeks laying out, in increasingly specific terms, a vision in which robots and artificial intelligence don’t just reshape the global economy — they double it. The pitch, delivered to a room full of technology ministers and then compressed into a single social media post, is forcing governments to reckon with numbers that sound more like science fiction than policy planning.
The G20 Pitch
The clearest version of Musk’s argument came on September 1, during a virtual appearance at the G20 Innovation Ministerial in Chapel Hill, North Carolina. Speaking to assembled technology ministers, Musk laid out a two-part thesis: artificial intelligence software alone could expand global economic output by 20% to 30% — somewhere between $20 trillion and $30 trillion a year. Robotics, he argued, is where the multiplier really kicks in.
He told ministers he was confident enough in the robotics prediction to bet serious money on it: within a decade, there will be at least one billion humanoid robots in operation, and each will produce roughly five times what a single human worker does. Combined, he said, that fleet would out-produce the entire human race.
Musk described the dynamic as recursive — once robots become capable of manufacturing other robots, growth that starts slowly begins to compound explosively. He broke down a robot’s usefulness into three multiplying factors: the AI software running its “brain,” the onboard chip powering that software, and mechanical dexterity — particularly hand movement — needed to perform physical work.
The Power Problem
Musk didn’t stop at economic upside. He also delivered what amounted to a warning: the industry’s ambitions are on a collision course with the electrical grid. He cited an industry consensus pointing to a shortfall of roughly 15 gigawatts in AI-related power capacity as soon as 2027, driven by chip production growing 40% to 50% a year while electricity generation outside China expands at only 10% to 20% annually.
He framed it starkly, telling ministers there is “quite a crisis of power” building — one that could arrive far sooner than most grid forecasts anticipate. Musk pointed to his own company, SpaceX, as an example of how serious the constraint has become: it built its own power plants to keep pace with AI infrastructure needs, and other companies, including Google and Anthropic, are now leasing capacity from those plants rather than waiting on traditional utilities.
His advice to the ministers in the room was direct: build out power generation, host AI data centers domestically, and tax or charge fees for the privilege — rather than cede that capacity, and the economic activity that comes with it, to countries able to move faster. He also argued for lighter-touch regulation generally, telling the ministers that new technologies should be “default legal” rather than default illegal.
David Sacks, a venture investor who has advised the Trump administration on science and technology policy, echoed Musk’s framing at the same event, arguing that AI-driven demand could ultimately push electricity costs down rather than up if new generation capacity comes online alongside it.
The Follow-Up Post
Eight days later, on September 9, Musk distilled the pitch into a single sentence on X, the platform he owns: AI and robots, he wrote, will more than double the global economy in under a decade.
That claim is bolder than it first appears. The International Monetary Fund puts current global GDP at roughly $126 trillion for 2026. Doubling that figure means the world would need to add a second $126 trillion in annual output by the mid-2030s — a pace of compounding growth well beyond anything in modern economic history, and far above the IMF’s own baseline projections for the period.
Not a New Argument
Musk’s robot-economy framing builds on positions he has staked out for months. In April, he argued on X that a federal “universal high income” — not merely a universal basic income — would be the right response to AI-driven job losses, on the theory that a surge in robot-produced goods and services would create disinflationary, not inflationary, pressure. Printing money to hand out as income, he reasoned, wouldn’t chase scarce goods the way it normally would, because AI and robots would be expanding supply faster than the money supply grows.
He has also previously suggested that Tesla’s Optimus humanoid robot alone could help “eliminate poverty,” with output gains he’s pegged at 10 to 100 times current productivity levels once the robot reaches mass deployment — though Optimus remains in early stages of commercial rollout.
The Reality Check
For now, the robotics industry is scaling nowhere near the trajectory Musk describes. China is expected to ship around 50,000 humanoid robots in 2026, according to Morgan Stanley figures cited by CNBC — a sharp increase from a January forecast of just 14,000, but a figure still many orders of magnitude short of the one-billion-unit fleet Musk envisions within ten years.
That gap is the crux of the skepticism surrounding Musk’s numbers: the compounding growth rates required to hit either the robot production targets or the GDP-doubling timeline are steep even by the standards of past technological booms. Musk has called his own billion-robot forecast conservative, betting that the compounding effects of AI, chips, and manufacturing dexterity will accelerate faster than critics expect.
Whether governments treat the message as an actionable warning or an overstated sales pitch, the core ask Musk left with the G20’s technology ministers was concrete: invest in power generation now, or risk watching the robot economy — and the tax revenue that comes with it — build out somewhere else.