How Elon Musk’s One‑Line Tweet Marked the End of xAI and Shifted 220,000 GPUs to Anthropic

Announcement

On May 6, 2026, after the U.S. stock market closed, Elon Musk posted on X: “xAI will no longer operate as an independent company. It is now SpaceXAI.” No internal email, farewell ceremony, or commemorative video accompanied the statement.

The $250 billion‑valued AI company, founded to “understand the true nature of the universe,” was effectively written into history with a single tweet.

A Revenge’s Starting Point

Three years earlier, in July 2023, ChatGPT sparked a global AI boom. While Microsoft, Google, and Meta went all‑in, Musk was absent from the boardroom. He had been the largest backer of OpenAI but was forced out of its board in 2018, which left him angry.

Musk publicly announced the creation of xAI, promising a “non‑politically‑correct” AI to counter OpenAI’s “woke culture.” The flagship model, Grok, was positioned as a direct competitor to ChatGPT.

Rapid funding rounds lifted xAI’s valuation from $10 billion to $2.5 trillion. Grok‑1’s benchmark results approached GPT‑4, and the model was fully integrated into the X platform, guaranteeing massive traffic.

Media narratives framed the venture as “Musk’s AI ambition: make xAI surpass OpenAI.” The hype, however, masked an underlying vulnerability.

Three Fatal Flaws

1. Grok’s awkward positioning. xAI marketed Grok as “fun” and “anti‑woke,” appealing to a vocal segment of the public but offering little value to enterprises or power users who demand stability, safety, and professional capability. Consequently, Grok failed to secure a solid foothold in either consumer or B‑to‑B markets.

2. GPU stack without talent. xAI amassed a world‑class super‑computer of 220,000 Nvidia GPUs delivering 300 MW of compute power, costing hundreds of billions of dollars. Yet AI research success depends more on people who can harness the hardware than on the hardware itself. Musk’s high‑pressure, flat‑management style—effective for rockets—proved unattractive to top AI researchers, prompting a mass exodus of the founding team.

3. The SpaceX acquisition sealed xAI’s fate. In February 2026, SpaceX announced a full acquisition of xAI at a combined valuation of $12.5 trillion. Once integrated into a larger conglomerate, xAI’s strategic priorities, resource allocation, and culture were reshaped, effectively ending its independence.

220,000 GPUs Handed to the Enemy

Following the acquisition, SpaceX struck a deal with Anthropic: the entire Colossus 1 compute cluster—220,000 GPUs and 300 MW—was leased to run Anthropic’s Claude models. Anthropic immediately doubled Claude Code limits and increased rate caps for Pro and Max plans.

Ironically, three months earlier Musk had criticized Anthropic as “anti‑humanist.” In a later post he noted that his discussions with Anthropic’s leadership were “competent, focused on doing the right thing, and did not trigger my ‘evil detector.’”

The move also aligned Musk’s interests with OpenAI’s chief rival, positioning SpaceX and Anthropic as allies against a common adversary.

Epitaph or Prospectus

From one perspective, Musk appears to have lost: $42 billion spent over three years, Grok failing to win any decisive battle, the entire founding team departing, and xAI being dissolved.

From another angle, SpaceX stands to gain. The company is eyeing an IPO with a target valuation of $17.5 trillion, leveraging a closed‑loop ecosystem of AI compute, rockets, satellite communications, and global data. The GPU assets now generate revenue from Anthropic, bolstering SpaceX’s financial narrative, while plans for “gigawatt‑scale orbital AI compute” promise a new frontier beyond terrestrial power and cooling limits.

Thus, Musk’s three‑year saga turned xAI into a strategic asset rather than a standalone AI champion—its GPUs fuel Anthropic’s Claude, and its integration fuels SpaceX’s broader AI‑space ambitions.