We’re feeling cynical about xAI’s big deal with Anthropic

Anthropic and xAI announced a major collaboration this week, with Anthropic purchasing all of the computing capacity at xAI’s Colossus 1 data center in Tennessee.

In his last episode TechCrunch’s Equity podcastKirsten Korosec, Sean O’Kane and I discussed what the deal could mean for xAI’s parent company SpaceX as SpaceX prepares to go public and apparently planning to disband xAI as a separate organization.

Kirsten did her best to offer “a positive view” of the partnership — after all, it’s a new way for xAI to make money. However, he also noted that this also suggests that xAI isn’t doing much in terms of training its own frontier AI models, and it’s harder for the company to position itself as a “progressive, innovative” business when that’s the case.

Sean then asked, “Why be positive when you can be cynical?” In his view, this looks like “an important pre-IPO heat check.” Yes, becoming a neocloud may be “a more believable business in the short term,” but it’s less likely to excite outside investors in the long term. (And then there’s the environmental lawsuit facing xAI over Colossus 1.)

Read on for a preview of our conversation, edited for length and clarity.

Sean O’Kane: I always like surprises, especially when everyone’s eyes [are] in another ball, a great test that happens. Seemingly out of nowhere this week, SpaceX and thus its artificial intelligence subsidiary xAI — which apparently no longer exists, or will soon exist, that we can reach — struck a deal with Anthropic.

Basically, the real version of the deal is that Anthropic takes over essentially all of the computing in the data center known as Colossus 1 in Memphis, Tennessee, to focus on Anthropic’s more business-focused AI products. There have been many reports of how [Anthropic’s] I was looking for more calculations […] and it seems like an escape valve for them to be able to get that deal and get access to all that computation.

In the short term, for xAI and for SpaceX, yes, it’s now a neocloud, in the sense that they had to do something with all that computing they were doing, because it sure looks like they wouldn’t need it for Grok — which, aside from X, isn’t setting the world on fire as far as the new hot consumer chat bot goes.

Kirsten Korosec: And we have to say that in terms of what neocloud is, for those who don’t know, that’s the idea of ​​buying GPUs from Nvidia and the like, and renting those as opposed to using those for their own AI, training their own AI models.

So it’s a different kind of business, and the point our AI editor Russell Brandom makes is that a lot of companies are building data centers, but given the choice, do they rent them? [or using them to train their own models]they still prioritize using this computation for their own in-house AI model training. I think this is an important point and one that suggests that maybe xAI is not doing that much for training the AI ​​model [side]

Anthony Ha: Right, and as Sean alluded to, most people wouldn’t necessarily consider Grok – not only is it known for some pretty nasty, if not outright illegal content, but it’s also not necessarily super cutting edge. Especially if we start talking about business AI, which I know we’ll get into later in this episode, you don’t hear a lot about people using Grok for mission-critical tasks.

And so the question arises: How can xAI actually make money? And obviously just selling the infrastructure could be one of the main ways to do that.

Kirsten: And you could be positive about that, right? They found a way to make money. But I think when you’re positioning your company—in this case, SpaceX-slash-xAI—as a future, innovative company, that’s a harder sell if you’re just renting out your GPUs and not using them for that innovation.

Sean: But why be positive when you can be cynical? That said, this looks like a major pre-IPO heat check that we’re going to see hit the markets with SpaceX.

Antony, you mentioned not only that Grok is not used for large enterprise work, but it has been reported that xAI employees were using other modelsthey didn’t even use [Grok] internally, and that caused this big upheaval inside the xAI, post acquisition by SpaceXwhich involved essentially all of the departing co-founders except Elon Musk, [and] He’s basically saying he’s starting from scratch on xAI, despite the fact that SpaceX paid $250 billion for it on the way up to the big IPO.

And now he says this they are going to disband xAI as a separate entity within SpaceX as a whole. He’s starting to call the whole thing SpaceXAI, because this man loves nothing more than destroying a brand that has any value – check out Twitter.

That might be a more believable business in the short term, and so on some level, I could see it being perhaps more attractive to investors when it comes time to IPO because it’s a little more credible and certainly more real than those who are borderline lab developers. But it’s also not the kind of business that’s going to attract the same—at least, in a normal environment—outside investment that we see going into all the frontier labs.

This is probably one of the biggest points of tension we’ve seen develop during this IPO process.