We are very skeptical about xAI and Anthropic

Anthropic and xAI he announced a big deal this week, it’s Anthropic to buy all the power at xAI’s Colossus 1 data center in Tennessee.

On the latest episode of TechCrunch’s Equity podcastKirsten Korosec, Sean O’Kane, and I discussed what this partnership could mean for xAI’s parent company SpaceX, as SpaceX prepares to go public with apparently planning to dissolve xAI as a separate entity.

Kirsten did her best to give a “good idea” to the deal – after all, it’s a new way for xAI to make money. But he said that this also shows that xAI isn’t doing much when it comes to training its AI models, and it’s hard for the company to position itself as a “forward-looking, innovative” business when it does.

Then Sean asked: “Why do you have to be positive when you can be cynical?” In his opinion, this appears to be “examining the heat before the IPO.” Yes, being a neocloud it may be a “credible business in the short term,” but it is unlikely that foreign investors will be happy in the long run. (And there is environmental crime that xAI is targeting Colossus 1.)

Read on to see our discussion, edited for length and clarity.

Sean O’Kane: I always like surprises, especially when everyone’s eyes (are) on the other ball, great trials what is happening. It seems that this week, SpaceX and therefore its AI company xAI – which seems to no longer exist, or will not exist, as far as we can reach – partnered with Anthropic.

Basically, the real nature of the deal is that Anthropic is taking over all the computing at a data center known as Colossus 1 in Memphis, Tennessee, to focus on Anthropic’s business-focused AI business. There have been a lot of reports about how (Anthropic’s) have been wanting more information (…) and it seems like an escape valve to be able to get involved in the project and get a chance to read all of it.

Recently, for xAI and SpaceX, yes, they are neocloud now, in the sense that they had to do something with all these computers that they build, because it seems that they would not need Grok – which, outside of X, does not set the world on fire until it becomes a new bot to buy consumers.

Kirsten Korosec: And we must say that according to what neocloud is, for those who don’t know, it has the idea of ​​buying GPUs from Nvidia and so on, and renting them to outsiders rather than using their AI, training their AI models.

So this is another type of business, and a point our AI editor, Russell Brandom, makes is that many companies are building data centers, but if given the choice, they rent them (or use them to train their models), still prioritizing the use of this computer for their internal AI training. I think this is an important point and one that shows that maybe xAI is not doing much in AI education (side)

Anthony Ha: Right, and as Sean says, most people wouldn’t think of Grok as – not only that he’s known for being unpleasant, if not more. absolutely illegalcontent, but not too important. Especially if we start talking about business AI, which I know we’ll get to later in this episode, you don’t hear much about people using Grok for critical tasks.

So the question becomes: How can xAI make money? And obviously just selling the damage would be one of the main ways to do that.

Kirsten: And you can feel good about it, right? He found a way to make money. But I think when you position your company – in this case, SpaceX-slash-xAI – as a forward-looking, innovative company, that’s a hard sell if you’re just renting out your GPUs and not using them for new things.

Sean: But why should you be positive when you can be cynical? Which means this looks like a big heat check before the IPO we’re about to see hit the markets with SpaceX.

Anthony, you didn’t just say that Grok wasn’t being used for large enterprise applications, there have been reports that xAI employees were using other examplesthey weren’t using (Grok) inside, and that caused a big shakeup inside xAI, post. purchase from SpaceXwhich really affected all the co-founders are leaving except Elon Musk(and) in fact they say that they are starting again on xAI, although SpaceX paid $ 250 billion to get to this mega-IPO.

And now he says that they have eliminated xAI as a special group within SpaceX altogether. He’s started calling the whole thing SpaceXAI, because this guy loves nothing more than destroying a brand that has value – check out Twitter.

This could be a promising business in the near future, and to some extent, I can see this being attractive to investors come IPO time, because it is more reliable and realistic than them being a laboratory manufacturer. But it’s not the type of business that can match – at least, in a normal environment – the foreign investment that we see going into all the laboratories.

This is probably one of the biggest problems we have seen happen during this IPO.