Elon Musk used a Davos conversation with BlackRock’s Larry Fink to lay out a tight set of commercial milestones and a striking strategic claim: the immediate limit on large-scale AI deployment is not chips or algorithms but electricity, and China — through massive solar deployment — is already best placed to supply it.
Musk distilled the economic case for robots into a single equation — economic output equals per-robot productivity multiplied by robot numbers — and predicted a surge in humanoid robot production that could, in his ideal scenario, exceed the global population. He painted an optimistic picture of robots caring for children and the elderly, easing labour shortages and delivering “sustainable abundance,” while acknowledging the philosophical question of human purpose if machines can do virtually everything.
On autonomous vehicles he was equally forward-leaning. Musk said full self-driving (FSD) is effectively solved at a product level, pointing to weekly software iterations, insurer-backed safety claims and a plan to roll out RoboTaxi services in the United States “very, very widely” by year-end. He also forecast regulatory approvals in Europe and China in the near term, even suggesting a Chinese approval could arrive “next month,” signalling a countdown for FSD’s official entry into those markets.
The speech pivoted to energy, where Musk issued a counterintuitive warning: semiconductor capacity is scaling exponentially but grid power is not. He argued that electricity generation is likely to grow only a few percentage points annually while compute demand for AI will accelerate far faster, creating a real risk of “chips but no power” unless generation and storage are dramatically expanded.
Musk singled out China’s solar deployment as the practical solution. He credited China with vast annual additions of photovoltaic capacity and estimated that, coupled with storage, Chinese solar could generate a steady-state output comparable to roughly half of U.S. average demand. He also criticised tariffs that keep solar equipment costs artificially high in the United States, noting that Chinese factories currently supply most solar panels worldwide.
To quantify the scale, Musk argued that a roughly 160-kilometre square of solar arrays could cover U.S. electricity demand, and revealed that Tesla and SpaceX are aiming to build large domestic manufacturing lines — on the order of 100 GW a year — to close the gap. He framed the barrier less as technology than as economics and policy: manufacturing scale and tariff regimes will determine how quickly countries electrify for an AI future.
Pushing the horizon further, Musk proposed space as the next platform for cheap, high-intensity compute. Space-based solar panels avoid weather and benefit from better cooling in shadowed environments, he said, making orbital AI data centres economically attractive once launch costs fall. That fall, he argued, depends on demonstrating full reusability for Starship, a move he predicts would cut access-to-space costs by two orders of magnitude and make spaceborne AI infrastructure commercially viable within two to three years.
He offered tight timelines: AI that outperforms any single human could appear by the end of this year or next, aggregate AI surpassing collective human intelligence by around 2030, Optimus humanoid robots entering public sale by late 2027 and industrial deployments of robots beginning earlier. While Musk wrapped his forecasts in an optimistic “curiosity philosophy,” his calendar remains ambitious compared with many independent assessments.
The implications are immediate and geopolitical. If Musk is right, countries that can scale power generation and storage cheaply — and build the supply chains for solar and batteries — will enjoy a competitive advantage in the coming AI-industrial wave. U.S. tariff policy, Chinese manufacturing capacity and regulatory timelines for autonomous systems are not just economic issues; they are strategic levers that will shape which firms and nations capture the gains of robot-driven productivity.
Whether one embraces his timetables or treats them with scepticism, Musk’s Davos remarks focus attention on three urgent policy choices for governments and corporate boards: accelerate grid and storage investment, reconcile industrial policy with trade realities, and prepare regulatory frameworks for autonomous systems that balance innovation with public safety. Those decisions will largely determine whether the age of abundant machine labour arrives as promised, and on whose terms.