As artificial Intelligence drives an unprecedented surge in computing demand, SpaceX and xAI have announced plans to deploy 100 GW of orbital computing capacity annually — ten times the combined announced pipeline of all other orbital data centre developers globally — highlighting growing industry interest in moving data infrastructure beyond Earth.
But despite the momentum, the economics remain daunting, with a hypothetical 1 GW orbital data centre estimated to cost $170 billion, more than three times the cost of an equivalent terrestrial facility, according to a report by Wood Mackenzie.
The push comes as next-generation AI agents are expected to consume between 10,000 and 40,000 times more computing power per task than today’s chatbots, placing mounting pressure on power systems, grid infrastructure and data centre capacity worldwide.
Global data centre electricity demand currently stands at 460 terawatt-hours (TWh) in 2026, equivalent to nearly half of Japan’s total power generation. According to Wood Mackenzie, demand is projected to reach 1,280 TWh by 2030 and 3,700 TWh by 2040, representing a 703% increase from current levels and growing at an annual rate of 16%.
The United States and China alone account for 78% of the world’s planned data centre pipeline, but expanding terrestrial infrastructure is becoming increasingly difficult.
Grid connections in the United States can take up to seven years, gas turbine equipment faces supply constraints through 2030, cooling systems are competing for scarce water resources in drought-prone regions, and construction costs continue to rise because of higher labour and material expenses, the report said.
These bottlenecks are forcing technology companies to seriously evaluate orbital data centres as a long-term solution to support AI growth.
However, the financial challenge remains substantial.
Wood Mackenzie estimates that launch and satellite infrastructure account for nearly 60% of the cost of a $170 billion orbital facility. To achieve parity with ground-based alternatives, orbital data centre costs would need to decline by approximately 70%.
There are signs that launch economics are moving in that direction. Global orbital launch attempts reached a record 324 in 2025, up 25% from the previous year, while launch costs have already fallen by more than 90% compared with earlier expendable rocket systems.
The pace of commercial activity in orbit is also accelerating rapidly. A record 4,517 satellites were deployed into space in 2025, a 58% increase over the previous year, while 87% of those satellites were owned by private entities.
Despite growing interest in orbital computing, investment remains firmly concentrated on Earth.
Wood Mackenzie forecasts $9 trillion in cumulative capital expenditure between 2026 and 2040 to build approximately 395 GW of new terrestrial data centre capacity under its base-case scenario.
The scale of ongoing terrestrial investment was highlighted recently when Anthropic committed $45 billion over three years to SpaceX for access to its 300 MW Colossus 1 terrestrial data centre, which deploys 220,000 Nvidia GPUs.
“The constraints on terrestrial data centres are genuine, and they are not going away quickly,” said Robert Liew, Research Director at Wood Mackenzie.
“But putting a data centre in orbit still costs at least three times as much as building one on the ground. That gap does not close without sustained and dramatic progress on launch costs. We forecast $9 trillion of terrestrial data centre investment between now and 2040. That is where capital goes first. Orbital data centres are a serious long-term proposition, but right now they remain a bet on the cost curve,” he said.
The report noted that no gigawatt-scale orbital or terrestrial data centre currently exists. While terrestrial expansion is expected to continue out of necessity, Wood Mackenzie concluded that large-scale orbital data centres remain, for now, a long-term technology proposition rather than a commercially viable alternative.