On May 13, 2026, The Wall Street Journal reported that Google is in advanced talks with SpaceX to launch data centers into orbit — not as a moonshot, not as a research project, but as a forward operational plan. The framing inside the discussions, according to people familiar, is striking: SpaceX is pitching orbital compute as the lowest-cost path forward, positioning the offering as an anchor narrative ahead of its planned $1.75 trillion IPO later this year.
Read that again. The lowest-cost path. Not the most ambitious. Not the most futuristic. The cheapest.
For two decades, the conventional wisdom held that data centers got cheaper as you got closer to power, fibre, and water. Hyperscalers chased Iceland's geothermal grids, Virginia's tax abatements, and Phoenix's industrial parks. Now, the same companies that built the playbook are pricing in the cost of gravity — and finding the math no longer works on Earth.
This is the moment terrestrial AI growth hits its ceiling in public view. When Sundar Pichai and Elon Musk's lieutenants sit across a table and conclude that a 60-tonne payload to low Earth orbit is the rational response to power constraints, the conversation about "scale" stops being a marketing exercise. It becomes a capital-allocation problem. And every growth-stage business is now downstream of it.
