SpaceX plans to significantly expand its compute capacity, targeting more than fivefold growth by the end of 2027. Elon Musk outlined the goal during the company's Q2 earnings call, stating the capacity should surpass two gigawatts. The company currently operates at 1.4 gigawatts, according to its quarterly report. The expansion will rely entirely on Nvidia chips, with Musk emphasizing the importance of the upcoming Vera Rubin platform. He described it as the best architecture available, indicating its critical role in meeting the ambitious targets.
The current and planned Colossus buildout includes roughly 100,000 H100, 110,000 GB200, and 110,000 GB300 processors, with an additional 220,000 GB300 units planned. This totals about 540,000 GPUs. If the expansion fully utilizes the Vera Rubin platform, the number of GPUs could exceed two million. SpaceX has not disclosed how much of the buildout will use Rubin, but Musk's AI company xAI merged into SpaceX in February 2026 through a stock-funded deal. The transaction valued SpaceX at $1 trillion and xAI at $250 billion, totaling $1.25 trillion.
SpaceX's AI segment, which includes xAI and X, reported $2.56 billion in revenue during Q2 2026 alongside an operating loss of $1.26 billion. The segment's losses were significant, with $3.73 billion in operating losses for the first half of 2026. Most of the revenue increase came from new cloud contracts leasing Colossus compute capacity, with a single customer accounting for about $1.52 billion of that revenue. That customer is likely Anthropic, which uses a mix of AWS Trainium, Nvidia GPUs, and Google's TPUs.
Source: thedecoder
