MobbleOpen in Mobble ⇢
Business · Corporate earnings · published 2026-10-10 · via Crypto Briefing

Goldman flags SpaceX GPU rental prices as sign of AI profitability strain

Image via Crypto Briefing
Image via Crypto Briefing

SpaceX’s AI unit earned $2.56 billion in Q2 2026, up 247% year over year, by renting surplus GPU capacity from its Colossus data centers. Goldman Sachs argues that current compute rental rates are far above the revenue AI companies need to become profitable, suggesting a scarcity premium rather than sustainable pricing. The bank warns that high rental costs could make it difficult for AI labs to reach profitability.

Expanded Detail

SpaceX's AI division generated $2.56 billion in Q2 2026, a 247% increase from a year earlier, by leasing excess GPU capacity from Colossus data centers. Its adjusted EBITDA turned positive at $1.1 billion.

Customers include Anthropic at about $1.25 billion monthly through May 2029, Google at roughly $920 million monthly, and an unnamed buyer committing $1.11 billion monthly from December 2026. Goldman estimates hyperscalers face a $230 billion yearly revenue gap versus investment requirements. SpaceX listed on Nasdaq in June 2026 as SPCX, with Goldman as the offering's main underwriter.

Context

If compute rental costs remain elevated, AI developers may pass expenses to users or delay projects, affecting consumers, researchers, and smaller startups. SpaceX investors could gain near term, yet reliance on a few large customers and future capacity additions may weaken pricing. Wider society might see AI services become costlier or more concentrated among well-funded firms, shaping access and innovation.

Expanded detail and Context are AI-generated analysis; the linked article remains the authoritative source.
Read the full article at Crypto Briefing →
This summary is Al-enhanced to contain extended analysis and broader social context. The original is {NAME); the linked article is the authoritative source. Original headline: “Goldman says SpaceX’s GPU rental boom exposes AI’s revenue gap.” Browse more stories.