SpaceX Has Spent $329 Million on Tesla Megapacks to Power xAI Compute
SpaceX has bought $329 million worth of Tesla Megapack battery systems in the first half of this year, regulatory filings show, as it powers up data center infrastructure for its absorbed xAI operations.
SpaceX has sharply increased its purchases of Tesla Megapack industrial battery systems, spending $295 million in the second quarter alone. Regulatory earnings filings submitted to the United States Securities and Exchange Commission show the company spent $329 million on Tesla battery hardware across the first half of this year.
The disclosures underscore the expanding internal trade between companies controlled by Elon Musk. SpaceX absorbed his artificial intelligence venture, xAI, earlier this year. Before that consolidation, xAI had separately purchased $430 million worth of Megapacks to support its own compute facilities, on top of $34 million in battery purchases it made during the first quarter. The filings also show that SpaceX acquired $131 million worth of Tesla Cybertrucks as of December 2025.
The aggressive battery deployment reflects the power demands of modern AI infrastructure. High-performance graphics processing units used to train and run large language models create volatile electricity loads that spike sharply during intensive compute tasks. Large-scale battery arrays let data center operators smooth peak demand, prevent grid overloads, and avoid punitive surge charges from local utilities.
xAI has faced scrutiny for running dozens of natural gas turbines at its primary Colossus data center in Mississippi. Even so, battery storage remains operationally critical. Industrial battery systems provide sub-second backup power when local grids fail or when generation output drops during heavy model training runs, keeping workloads online without interruption.
The Cognarah Angle
The scale of battery infrastructure required to keep xAI compute running confirms something African AI founders and data center operators already live with daily: compute is, at its core, an energy problem. Across West and East Africa, national grids suffer chronic outages and voltage instability. Running serious AI workloads on raw grid power is not a cost optimization question; it is a reliability question. Where Silicon Valley operators use massive battery installations to shave utility bills and smooth demand spikes, African data center operators deploy off-grid batteries and solar arrays simply to guarantee baseline uptime for local inference nodes.
The broader concern here is structural. The consolidation of satellite connectivity, AI models, data center compute, and battery storage within a single corporate entity creates steep barriers for any independent infrastructure provider trying to compete. SpaceX handles the last-mile internet link through Starlink. xAI supplies the models. Megapacks keep the power on. When one firm controls that entire stack, smaller operators are not partners in an ecosystem; they are tenants with no lease protections.
For African governments and tech investors serious about sovereign AI infrastructure, this should sharpen the urgency of domestic energy storage investment. The gap between African AI ambition and African AI capability is, in large part, a battery gap. Fixing it requires treating energy storage as core tech infrastructure, not a utilities footnote.
If African tech hubs cannot close the energy storage gap, every local AI compute project risks depending on the very foreign monopolies it should be trying to route around.
Reporting sourced from TechCrunch. Analysis and Cognarah Angle are Cognarah's own.
Written by
StacyAI-assisted news curation. Every story is reviewed by our editors before publication.
