THURSDAY, MAY 21, 2026 · BRISBANESUBSCRIBE →

THE AI POST

INTELLIGENCE. CURATED.

SpaceX rocket launch at night representing the merger of xAI into SpaceX
BreakingMay 8, 2026

Musk Killed xAI. Every Co-Founder Left. Now Anthropic Runs His Supercomputer.

xAI is dead. All 11 co-founders are gone. The supercomputer built to beat Anthropic now powers it.

Elon Musk just did something nobody expected: he killed his own AI company.

xAI, the company Musk founded in 2023 to challenge OpenAI and Anthropic, will be dissolved as a standalone entity and absorbed into SpaceX under a new division called SpaceXAI. Musk confirmed the move on X on May 6: "xAI will be dissolved as a separate company, so it will just be SpaceXAI, the AI products from SpaceX."

The timing is brutal. This comes after a merger announced in February 2026 that valued xAI at $250 billion and SpaceX at $1 trillion, creating a combined $1.25 trillion entity. But a sky-high valuation means nothing when nobody wants to stay.

The Exodus That Ended xAI

Every single one of xAI's 11 original co-founders had left by March 28, 2026, according to The Next Web. All of them. Musk is the sole founder remaining. He acknowledged the scale of the failure publicly, telling the tech community that "xAI was not built right first time around, so is being rebuilt from the foundations up."

Fast Company reported in April that more than 80 xAI staffers left in recent months beyond the co-founders. The Financial Times traced some departures to Musk making "unreasonable demands" to improve Grok's technology as it fell further behind Claude and ChatGPT. The result: a company that was valued at a quarter-trillion dollars but couldn't keep anyone around to build the product.

The Anthropic Irony Is Almost Too Perfect

Here is the part that makes this story extraordinary. Colossus 1, the Memphis-based supercomputer that xAI built for over $1 billion with more than 220,000 Nvidia GPUs (including H100s, H200s, and GB200s), was supposed to be Musk's competitive weapon against Anthropic. Instead, SpaceX is now leasing that entire facility to Anthropic to power Claude.

Musk spent years publicly attacking Anthropic and its backers. He called Dario Amodei's safety rhetoric performative. He said Anthropic "hates Western civilization." Three months later, his rocket company is directly powering one of his main competitors. Both companies framed the deal around shared ambitions in "orbital infrastructure," but the practical reality is simpler: Colossus 1 generates revenue for SpaceX by serving the rival Musk spent two years trashing.

Grok Is Bleeding Users While Claude Surges

The user numbers tell the story the valuation won't. Forbes reported that Grok's daily active users on mobile dropped from 13.9 million in March to 12.2 million in April, a 12.5% decline. In the US alone, daily app users fell 15.6%. Grok dropped from second place among chatbots in January to fifth by April, losing ground to Claude, Gemini, and DeepSeek.

Meanwhile, Claude's average daily active users worldwide jumped from 16 million to 23 million between March and April, a 44% spike, according to Similarweb data. Claude's worldwide web visits surged 39% in the same period. Year over year, Claude's mobile app users are up 1,205%. Grok? Up 83.5%. That is not a competition. That is a rout.

What SpaceXAI Actually Means

Grok is not going away. Development continues under SpaceXAI, and xAI's products remain live. But the company that built them no longer exists, its founding team has scattered, and its flagship compute asset is partly rented out to a competitor.

CNBC reported the restructuring was framed internally around "speed of execution." That reads less like strategy and more like damage control after a mass exodus. Whether a rocket company can stabilize an AI product roadmap is the question heading into the second half of 2026. The answer, based on the trajectory so far, is not encouraging.

Musk built Colossus to beat Claude. Now Claude runs on Colossus. Sometimes the story writes itself.