SpaceX deployed $15.8 billion toward AI compute infrastructure over three months, exceeding its entire 2025 Space division capex by fourfold, while Musk warns global chip production cannot scale quickly enough.

SpaceX spent $15.828 billion on capital projects in its AI segment during the three months ended June 30, 2026. That quarterly figure is 4.13 times larger than the $3.832 billion its entire Space segment spent across all of 2025.

The comparison derives from SpaceX’s second-quarter Form 10-Q and the segment note in its 2025 audited accounts. While this unaudited quarterly report is not a general measure of AI infrastructure costs, it demonstrates how sharply the company’s centre of capital investment has shifted.

An accounting distinction is necessary here. Capital expenditure funds assets expected to remain useful beyond the current quarter and differs from operating expenses; consequently, the full $15.828 billion did not pass through SpaceX’s quarterly profit-and-loss statement.

For the second quarter, SpaceX reported $18.369 billion in total capital expenditure. AI accounted for $15.828 billion, or 86.2 per cent. Connectivity consumed $1.367 billion, while Space took $1.174 billion.

The pace had already accelerated. AI capital expenditure reached $7.723 billion in the first quarter, meaning it more than doubled in Q2. Across the first six months of 2026, SpaceX invested $23.551 billion into AI capital projects, representing 82.7 per cent of the company’s $28.476 billion total.

When we previously examined the AI infrastructure pivot buried in SpaceX’s IPO prospectus, it remained largely a statement of direction. These first public quarterly accounts now attach a price tag to the opening phase.

SpaceX described roughly $15.8 billion as supporting the build-out of AI compute infrastructure. That wording matters. The figure should not be interpreted as a single purchase of Nvidia chips.

Compute infrastructure encompasses accelerators, servers, high-speed networking, storage, buildings, electrical substations, backup generation, cooling equipment, and construction. During the quarter, SpaceX also disclosed the purchase of $295 million in Tesla Megapacks. Together, these systems transform a collection of chips into a data centre capable of reliably selling computing capacity.

The physical result was 1.4 gigawatts of nameplate compute capacity at the end of June, up from 1 gigawatt in March and 400 megawatts a year earlier. SpaceX stated it expects to exceed 2 gigawatts by the end of 2026.

The AI segment generated $2.561 billion in Q2 revenue, including $1.6 billion from new AI infrastructure, while recording a $1.257 billion operating loss. Depreciation and amortisation within the segment reached $1.885 billion.

Those figures cannot be compared as though they belong in the same column. Servers and data-centre buildings are recorded as assets and generally expensed through depreciation over their useful lives. Capital expenditure indicates how much capacity SpaceX is building; operating income reflects business performance under the accounting costs recognised in that period.

According to a published transcript of the earnings call, SpaceX said its new compute deployments are achieving payback in less than a year. This is a management claim based on early contracts rather than a result independently proven by the 10-Q. The filing also warns that major cloud customers can generally terminate agreements after an initial ramp period by providing 90 days’ notice.

Dividing $15.828 billion by $3.832 billion yields the headline’s “more than four times” comparison. Both figures represent segment capital expenditure, making the arithmetic like-for-like.

This comparison still does not capture every dollar devoted to rockets. Space research and development alone cost $1.076 billion in Q2, much of it linked to Starship, and R&D is classified as an operating expense rather than capital expenditure. Nor does the ratio imply that AI has become four times more important than Space. It simply indicates that, during this specific build-out, significantly more capital was being allocated to AI assets.

That distinction echoes our earlier reporting on SpaceX’s IPO proceeds allocation: rockets, constellations, and compute share one balance sheet, but they are at very different stages of construction.

SpaceX’s prospectus states that orbital AI at scale requires “significantly more” AI chips than are currently available to the company, noting that GPUs and specialised components come from a small group of qualified suppliers.

On the Q2 earnings call, Elon Musk framed the constraint more starkly. He estimated that memory production was growing at approximately 20 per cent annually while demand was rising around 200 per cent. These are Musk’s projections, not audited industry statistics, and forecasts from an interested buyer warrant appropriate scrutiny. His broader point remains less controversial: semiconductor factories take years to plan, equip, and qualify, whereas data-centre orders can expand far more rapidly.

That reality underpins Terafab. Announced by Musk in March, the project addresses the concern that existing manufacturers are not scaling output quickly enough to meet the combined demands of SpaceX and Tesla. Terafab aims to consolidate logic, memory, and advanced packaging under a single programme. It also represents a substantial manufacturing risk rather than an immediate solution to today’s supply shortage.

Most of this quarter’s spending supported terrestrial compute. SpaceX’s longer-term strategy involves deploying computing satellites—an ambition contextualised in our previous analysis of its proposal for up to one million AI spacecraft. That orbital system does not yet exist at commercial scale.

SpaceX closed June with $93.522 billion in cash and $6.487 billion in marketable securities following its IPO and a $25 billion notes offering. This capital position enables an unusually rapid build-out. The unresolved question remains whether chips, memory, power equipment, and paying customers can all scale simultaneously. Q2 spending patterns indicate that SpaceX believes they will.