SpaceX's $2.8B Gas Turbine Bet: Fossil Fuel Stopgap, Space Solar End-State, and the Tension with Tesla's Clean Energy Mission
Summary
- • SpaceX disclosed $2.8B in gas turbine commitments to power xAI's Colossus data centers, contradicting Musk's own Tesla clean energy vision
- • xAI spent $697M on Tesla Megapacks but has bought no materially significant solar panels from Tesla, relying instead on unregulated natural gas
- • SpaceX's IPO filing frames gas turbines as a conscious stopgap, with space-based solar — generating 5x terrestrial output — as the intended long-term power source
- • SpaceX projects "terawatt-scale annual AI compute growth," dwarfing the ~40 GW total consumed by all data centers globally today
Details
SpaceX committed $2.8B+ to gas turbines across two deals in March and April 2025 — $805M in March through 2029, $2B in late April still pending
Both deals were struck with unnamed vendors. The scale of the commitment signals that fossil-fuel-powered temporary infrastructure is a core, not incidental, part of xAI's near-term data center strategy.
xAI's Colossus data centers collectively draw approximately 1 gigawatt of power — Colossus 1 in Memphis, TN; Colossus 2 in Southaven, MS. SpaceX has $14B+ in construction in progress.
The scale of construction in progress indicates the buildout is far from complete. Combined, the two facilities represent a major share of dedicated AI compute infrastructure in the US.
SpaceX is leasing Colossus server capacity to Anthropic for $15 billion annually
This positions SpaceX as a major commercial AI infrastructure provider — not merely an internal compute operator for xAI. Musk has signaled more such third-party leasing deals are planned.
xAI spent $697M on Tesla Megapacks over two years but has not purchased materially significant solar panels from Tesla
SpaceX also spent $131M on 1,279 Cybertrucks. The intra-Musk-empire purchasing pattern conspicuously excludes Tesla's solar products, despite Musk's public commitment to solar as the foundation of a clean economy.
The NAACP and other groups sued xAI over 27 unpermitted gas turbines in Memphis. SpaceX added 19 more turbines after the suit, bringing Colossus 2 to 46 units.
Portable turbines can legally operate for one year without a clean air permit. The decision to expand turbine count after the lawsuit suggests SpaceX is prioritizing speed-to-capacity over regulatory compliance, with potential public health implications in a majority-Black city.
SpaceX is using gas turbines as a deliberate grid-bypass strategy amid US electricity shortages, explicitly framing them as a stopgap pending space-based solar
The IPO filing characterizes terrestrial power constraints as a fundamental bottleneck, arguing third-party demand estimates are artificially low because they are bounded by what the terrestrial grid can supply.
SpaceX's IPO filing identifies space-based solar as its intended long-term power source for AI data centers, citing 5x the energy output of terrestrial arrays
The 5x advantage comes from 24/7 illumination in orbit, versus the day/night cycle and atmospheric losses affecting ground-based panels. SpaceX is positioning orbital infrastructure as the eventual replacement for its current gas turbine buildout.
Economic and technical challenges make space-based data centers difficult to realize near-term: Starlink power costs run multiples higher than terrestrial rates, chip hardening is expensive, and AI training distribution across satellites is unsolved
These are not minor obstacles. The cost gap between orbital and terrestrial power delivery remains large, and the question of whether large-scale AI training — which benefits from tight interconnects — can be distributed across geographically separated satellites is an open research problem.
SpaceX's IPO filing projects 'terawatt-scale annual AI compute growth,' versus roughly 40 gigawatts consumed by all global data centers today
If this projection is directionally correct, today's entire global data center fleet would represent a rounding error in future AI compute demand. This framing justifies both the urgency of the gas turbine buildout and the long-term case for space-based solar.
Tesla's founding 'Master Plan' explicitly calls for transitioning from a 'mine-and-burn hydrocarbon economy' to a 'solar electric economy' — a vision xAI's gas turbine reliance directly contradicts
Musk authored and published multiple Tesla Master Plans over the years with electrification and solar as the consistent throughline. The xAI buildout represents a notable departure from that stated philosophy, even if Musk views it as a temporary phase.
SpaceX is preparing an IPO on the Nasdaq; turbine commitments, Anthropic revenue, and energy strategy disclosures all came via the prospectus
The IPO filing is the primary source for most of the financial and strategic details now public about xAI's infrastructure. Investors will need to weigh a fast-growing AI infrastructure business against environmental liability and regulatory exposure.
Financials = spending/revenue figures, Infrastructure = physical buildout, Legal = litigation/regulatory, Strategy = business positioning, New Tech = emerging capability, Insight = analytical observation, Stat = key figure, Context = background, Industry Update = corporate events
What This Means
Musk is making a calculated bet that gas turbines are an acceptable bridge fuel for the next few years while SpaceX develops the capacity to put data centers in orbit, powered by space-based solar that the company says will generate five times the energy of terrestrial panels. The problem is that this bridge involves unpermitted fossil fuel infrastructure in communities that are already suing, and the space-based endgame faces real economic and technical hurdles that SpaceX has not resolved. Meanwhile, the absence of solar panels in xAI's supply chain — despite hundreds of millions flowing to Tesla for batteries and to SpaceX for trucks — exposes a widening gap between Musk's public clean-energy identity and the actual operating choices his companies are making. For investors evaluating the Nasdaq IPO, the story is a fast-growing AI infrastructure business with a compelling long-term vision, meaningful near-term regulatory liability, and an ideological contradiction that the market will eventually have to price.
Sentiment
Mostly skeptical of environmental and regulatory risks, with focus on aggressive cloud/AI expansion
“SpaceX just pledged over $2.8 billion for gas turbines to juice its AI data centers, right as critics hammer the carbon emissions. The bigger twist is what this says about its cloud ambitions.”
“🟠 SpaceX drops $2.8B on gas turbines to power AI data centers because apparently training Grok is more important than SpaceX's green rocket PR”
“DEAL: xAI is buying $2.8B in natural gas turbines for AI data centers over three years, disclosed in SpaceX's IPO filing. The company faces a lawsuit over pollution from its existing generators.”
“SpaceX S-1: xAI plans to buy another $2.8B worth of turbines for its data centers, including a $2B deal for mobile gas turbines, the type it's being sued over”
Split
~60/40 skeptical vs neutral on risks vs speed-to-capacity (limited deep takes from major voices)