Elon Musk’s declaration that Tesla will spend more than $25bn on artificial intelligence and robotics this year — three times last year’s capital outlay — has landed with the force of a rocket launch inside the boardrooms of Wall Street and the quiet corridors of Beijing’s EV strategy teams. The announcement, embedded in quarterly results that otherwise beat expectations, signals that Musk is now betting the company’s future not on electric vehicles but on a moonshot in autonomous driving, humanoid robots, and the sprawling neural networks needed to power them. Viewed from Washington, this is a high-stakes pivot that risks alienating the four million Tesla owners still waiting for the “full self-driving” promised on hardware they bought years ago. Analysts in London note that the hardware arms race is intensifying: Tesla is upgrading its AI4 computer to a 64GB, dual-SoC system dubbed AI4 Plus, designed to close the memory bandwidth gap with Nvidia’s Orin and Thor chips. Yet Musk himself admitted that even current hardware cannot deliver Level 4 autonomy, and that regulatory approvals in Texas and California are unlikely before 2027 — a timeline that sits awkwardly with the billions being poured into compute infrastructure now.

Meanwhile, a separate narrative is gathering speed on the other side of the Musk empire. SpaceX, valued at a rumoured $1.75 trillion, is preparing an initial public offering that could be the largest in history, luring investors with a more predictable leadership under Gwynne Shotwell and fewer of the quarterly earnings dramas that plague Tesla. From Zurich to Singapore, fund managers are closely watching the filings, which reveal that SpaceX will retain “controlled company” status, giving Musk and insiders super-voting shares and shielding them from independent board oversight. The contrast with Tesla’s recent financial discipline — the EV maker still generated positive cash flow of $1.4bn in the first quarter on $22.4bn in revenue — only sharpens the question of where the Musk premium should be allocated. Compounding the complexity, SpaceX is in talks to acquire or partner with the AI coding startup Cursor in a deal worth up to $60bn, and Musk’s xAI has held discussions with France’s Mistral to form a three-way alliance against OpenAI and Anthropic. Viewed from Paris, this coordination between Musk’s entities raises legitimate competition concerns; from San Francisco, it looks like the assembling of a private-sector AI Avengers team.

What emerges from the cacophony of quarterly reports and IPO teasers is a clear pattern: Musk is building a closed-loop ecosystem where Tesla’s cash flow funds SpaceX’s AI ambitions, and SpaceX’s eventual public listing may cannibalise Tesla’s investor base. The looming $3 trillion IPO wave — also including OpenAI and Anthropic — promises unprecedented market liquidity but also a brutal reckoning with fundamentals. As one strategist put it, once the excitement fades, profitability will be demanded. For Tesla, that means proving that its $25bn AI splurge can eventually produce a self-driving service that regulators approve and customers trust. For SpaceX, it means showing that a rocket company can deliver software margins. For Musk, it means managing the impossible geometry of three high-stakes ventures at once — a balancing act that will define not just his net worth, but the next chapter of the global technology race.