Anthropic struck a deal this week to take over all the compute capacity at xAI's Colossus 1 data center in Memphis, Tennessee. On the surface, it's a win-win: Anthropic gets desperately needed compute for its enterprise AI push, and xAI monetizes infrastructure it wasn't fully using for Grok.

But TechCrunch's analysis suggests something more cynical is happening. In a brutal takedown on their latest Equity podcast, the team called this "a major heat check before the IPO" and questioned whether xAI is even a real AI company anymore.

"Why be positive when you can be cynical?" asked TechCrunch's Sean O'Kane. "This seems like a more believable business in the near term, but it's also not the kind of business that's going to draw the same outside investment that we're seeing go into all the frontier labs."

The context makes the deal look worse for xAI. Grok — xAI's chatbot — is known more for generating illegal content than cutting-edge capabilities. It's not being used for enterprise work. Reports emerged that xAI employees weren't even using Grok internally, leading to a company-wide shakeup that saw all co-founders except Elon Musk leave the company.

Now Musk is dissolving xAI as a separate entity entirely, folding it into SpaceX ahead of the planned IPO. He's rebranding the whole thing "SpaceXAI" — another example of Musk "ruining a brand that has some value to it," as O'Kane put it.

The neocloud business model — buying GPUs from Nvidia and renting them out rather than using them for your own AI training — suggests xAI isn't prioritizing frontier model development. Most companies would choose to use their compute for internal training if they had genuinely competitive AI models in development.

"When you are positioning your company as a forward-looking, innovative company, that's tougher to sell if you are simply just renting out your GPUs and not using them for that innovation," noted TechCrunch's Kirsten Korosec.

The timing is particularly awkward. SpaceX paid $250 billion for xAI in the run-up to the mega-IPO, positioning it as a key strategic asset. Now that asset is essentially becoming a data center rental business.

There's also the environmental lawsuit hanging over Colossus 1. xAI is facing legal action for operating over 400 MW of gas turbines without proper permits, adding regulatory risk to what was supposed to be a clean tech showcase.

The deal does give Anthropic access to significant compute capacity as it competes with OpenAI and Google. Anthropic has been aggressively expanding its infrastructure partnerships, with $200 billion committed to Google Cloud and a $1.8 billion deal with Akamai for edge inference.

But for xAI and SpaceX, the Anthropic partnership looks more like a face-saving pivot than a strategic masterstroke. The company that once promised to challenge OpenAI and Google is now settling for being their landlord.

The broader question for investors: Is a data center rental business worth the $250 billion SpaceX paid for it? And more importantly, does this kind of infrastructure play generate the excitement needed for a successful IPO in a market obsessed with frontier AI capabilities?

As TechCrunch's analysis suggests, the answer might be a resounding no. The Anthropic deal solves xAI's immediate revenue problem, but it doesn't solve its innovation problem. And in the AI race, infrastructure without cutting-edge models is just expensive real estate.