Wood Mackenzie: The Challenges Facing Data Centres in Space

It's an idea that sounds more like science fiction than site selection: putting data centres into orbit.

And according to a new report from Wood Mackenzie, the idea is gaining traction because terrestrial data centres are running into mounting obstacles around power and cooling.

As AI workloads grow more demanding and power constraints tighten across major markets, some of the world's largest firms are exploring this as a potential solution.

Yet despite growing interest, the economics of orbital infrastructure remain a major hurdle.

An overview of energy consumption

Global data centre power demand stands at 460TWh in 2026, which is the equivalent to around half of Japan’s total power generation.

Wood Mackenzie projects that demand will rise to 1,280TWh by 2030 and 3,700TWh by 2040.

That represents a 703% increase from current levels, with an annual growth of 16%.

The scale of expansion is concentrated in a handful of markets.

The US and China account for 78% of the global planned data centre pipeline, placing even greater pressure on power grids and infrastructure supply chains.

What are the barriers to building orbital data centres?

The report highlights a growing list of constraints that terrestrial developments are currently up against.

In the US, securing a grid connection can take as long as seven years. The lead times for gas turbine equipment, meanwhile, often stretch through to 2030.

Those challenges are encouraging technology companies and investors to examine orbital data centres as a potential long-term alternative.

However, the financial gap between the two approaches is enormous.

Wood Mackenzie estimates that a hypothetical 1GW orbital data centre would cost somewhere in the region of US$170bn to build, more than three times the cost of an equivalent terrestrial facility.

Launch and satellite expenses account for roughly 60% of the total investment.

To achieve cost parity with facilities on Earth, orbital deployments would require launch-related costs to fall by around 70%.

According to the report, that outcome depends heavily on whether historical trends in launch economics continue.

There are signs of momentum in the space sector. Global orbital launch attempts reached 324 in 2025, up 25% year-on-year.

Commercial operators account for 70% of those launches.

Satellite deployment is also accelerating, with a record 4,517 satellites entering orbit in 2025, an increase of 58% from the previous year, with private organisations owning 87% of them.

Space ambitions grow despite economic realities

Some companies are already outlining ambitious plans for orbital computing capacity.

In November 2025, SpaceX and xAI's CEO Elon Musk set out an ambitious vision for orbital data centres.

On X, he wrote: "Starship could deliver 100GW/year to high Earth orbit within four to five years if we can solve the other parts of the equation."

That figure is ten times larger than the combined announced pipeline of every other orbital data centre developer globally.

Even besides SpaceX, the orbital market is heavily concentrated in the US. Non-US companies account for less than 0.5 GW of planned orbital capacity, highlighting how early-stage and geographically concentrated the sector remains.

Despite the higher costs, Wood Mackenzie expects launch activity among the five leading companies in the sector to begin accelerating between 2027 and 2028.

At the same time, investment in terrestrial infrastructure shows no sign of slowing.

According to Bloomberg, Anthropic committed US$45bn over three years to SpaceX for access to the 300MW Colossus 1 terrestrial data centre.

Wood Mackenzie forecasts cumulative capital expenditure of US$9tn between 2026 and 2040 to deliver around 395GW of new terrestrial data centre capacity under its base-case scenario.

Why terrestrial infrastructure is the priority

Despite growing discussion around orbital facilities, the report makes clear where most investment is expected to flow.

Robert Liew, Research Director at Wood Mackenzie, says: “The constraints on terrestrial data centres are genuine, and they are not going away quickly.

“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 US$9tn 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.”

Wood Mackenzie’s base-case energy transition outlook does not include large-scale orbital data centres. The report notes that no gigawatt-scale orbital or terrestrial facility currently exists.

For now, the numbers point firmly towards Earth.

While orbital computing attracts attention as power demand climbs and infrastructure bottlenecks persist, the industry’s expansion continues to depend on a vast build-out of terrestrial capacity.

With trillions of dollars earmarked for ground-based facilities over the next decade and beyond, conventional data centres are still the primary destination for investment even as the prospect of computing in space comes to fruition.