Why Musk Is Merging xAI into SpaceXAI to Weaponize Compute Against OpenAI

Elon Musk announced that xAI will be dissolved as an independent company and merged into SpaceXAI, coinciding with a new compute partnership with Anthropic. The announcement states that SpaceXAI will provide Anthropic with access to Colossus 1, one of the world’s largest AI supercomputers.

Colossus 1 houses over 220,000 NVIDIA GPUs, including densely deployed H100, H200 and next‑generation GB200 accelerators. The cluster delivers extreme parallel performance for large language models, multimodal systems, scientific simulations and generative AI. To power the 300 MW facility, a gas‑turbine generator is already in use, highlighting the massive energy demands of such a system.

As part of the agreement, Anthropic announced that its Claude Code model’s rate limits have been doubled for Pro and Max users, peak‑time throttling removed, and the Opus API quota significantly increased, directly attributing the performance boost to the additional compute from SpaceX’s infrastructure.

The article notes that xAI is projected to lose $13 billion in 2025, burning roughly $1 billion per month, and that GPU utilization has been low despite the massive hardware count. These financial pressures are presented as a key driver for the “golden handcuff” strategy of renting idle compute to Anthropic.

Strategically, Musk is portrayed as using the compute lease to both monetize excess capacity and to create a competitive lever against OpenAI. By binding Anthropic to SpaceX’s hardware, Musk aims to weaken OpenAI’s market dominance and pressure Microsoft’s Azure cloud business.

The piece frames the development as a three‑pole competition: (1) OpenAI + Microsoft, (2) Google with its TPU and DeepMind assets, and (3) the newly formed SpaceXAI‑Anthropic alliance, which plans to extend AI compute into orbit using Starship‑launched servers that benefit from vacuum cooling and uninterrupted solar power.

Looking ahead, the article speculates that if SpaceXAI can deliver orbital AI training and inference services, traditional cloud providers such as AWS, Azure and Google Cloud could face severe challenges, especially as SpaceX’s valuation approaches $1.25 trillion and an IPO is rumored for 2026.