Whittling Down Apple's Big AI M&A Options

In his newsletter this week, Mark Gurman throws out the following eight names for would-be M&A for Apple with regard to AI:

• OpenAI
• Anthropic
• Perplexity
• Thinking Machine Labs
• Cohere
• Sierra
• Databricks
• Mistral

I'll just point out that last March, I made the case for four of five such companies:

• Anthropic
• Perplexity
• Mistral
• Safe Superintelligence
• Thinking Machine Labs

I'm biased, but my list is still better all these months later (and that's even though I should be biased in the other way, since my wife and I are investors in Sierra and she advises the company!). I just think my list is more drilled in to what Apple actually needs and how they might be thinking about this, as I discussed with John Gruber for his Talk Show podcast at the time. And now the actual reporting – much of it by Gurman himself – has proven that out thus far.

To that end, the only real bit of new actual news in this report is the following:

Separately, Apple met earlier this year with Mira Murati — the former chief technology officer of OpenAI — to discuss a potential deal for her new AI startup, Thinking Machines Lab. The talks never progressed to an advanced stage.

This is not surprising given that we also now know that Meta spoke with Murati as well as a part of Mark Zuckerberg trying to build his new 'Superintelligence' Dream Team. But back in March, I made the following case:

Thinking Machines – The newest company of the bunch was started after Mira Murati left OpenAI and we've subsequently been seeing a lot of talent quickly join her. It was reported that her exit was perhaps a reason why Apple ended up not making an investment in OpenAI since she was a key point person for the company, so is it possible that translates directly into a new deal here? Either an investment or an acquisition? The company clearly doesn't have a product yet, but Apple needs talent more than it needs any product. And there's presumably a price that could work here for both sides. The company is said to be raising their "seed" round at a $9B valuation. And given that they're not even six months old, it might take a $20B - $30B+ offer to get them to build within Apple...

That seed round is now complete. $2B raised – perhaps the largest "seed" ever – at a $10B valuation. And it comes with some truly incredible levels of control for Murati, who undoubtedly learned some vital lessons from her time at OpenAI... This price, while obviously outlandish for a company still in the planning stages for their actual products, still wouldn't put it out of the realm of possibility for Apple. But just as with Perplexity, it would undoubtedly take an offer that's an order of magnitude more than any deal they've ever done...

Speaking of Perplexity, Gurman reported at the end of last week on those talks and I extended my own thoughts on the matter here. I've long thought Perplexity would be the most obvious and perhaps most seamless target for Apple – and in part because it's not just about AI, but also the overall Search story here as well. Something that will be vital if/when the Google Search deal is altered. Which led to my closing remarks last January:

Anyway, Apple should probably just buy Perplexity?

Fast-forward back to my case last March

Perplexity – This would also be insanely expensive for Apple, as this company is now said to be raising at an $18B valuation – so we're perhaps talking about a deal in the $40B - $50B range to entice everyone to do it. But beyond their good product chops with AI, this company obviously has focused on search from the get-go and that could be extra useful for Apple if and when the Google Search default deal goes away. My hold up, beyond price, is that the founder has been a bit outspoken and controversial in ways that I think range from stepping-in-it to somewhat understandable for a company in this space trying to make waves. Apple, of course, wouldn't necessarily appreciate that gumption, internally.

The valuation ended up coming in at $14B. And that perhaps makes a "mere" $30B deal – again, an order of magnitude larger than the Beats deal a decade ago – still within the realm of possibility. We also now know Eddy Cue has taken a liking to the company – as he said as much in court under oath! Still, as I laid out last week, I think a partnership here is more likely than an acquisition – and Alex Kantrowitz and I discussed this topic on his Big Technology Podcast a couple weeks ago. Beyond the potential personality clashes, the Samsung deal seems like a hurdle...

Still, Perplexity was the name I threw out there in my predictions for 2025 – alongside Mistral. Here's what Gurman had to say about that one:

Unlike the others, Mistral would give Apple a serious leg up in building cutting-edge foundation models. Its high-performance models are widely respected for their efficiency and speed — ideal for running on devices or in the cloud.

I don't know about a "leg up", but it would get them closer to OpenAI, Google, etc. And it would be "open source" work. And a lot of their talent comes from the team that started the Llama work for Meta, but clearly got fed up and left, leading to Meta's current AI situation... But again, I think my case here makes more sense:

Mistral – The great European hope also lacks some of the massive corporate backing of the others (though they have Microsoft and a few others large players with smaller stakes). More importantly, they seemingly lack some of the traction of the other major players right now. Obviously, they've been going about building a bit differently, being more in line with open source than the others. And this could be an interesting use of any cash Apple has overseas. The last round valued Mistral north of $6B, so a deal would probably need to be in the $15B+ range. Decidedly doable. Then again, Apple has basically been at war with the EU – it may or may not be thawing with a regime change – and it's not clear if such a deal would help or hurt that relationship. Or if the EU would ever dare bless it, of course...

They're clearly also fundraising again right now, on the heels of some big EU deals. And they're seemingly in a good spot from that perspective with the rest of the world – and Europe in particular – worried about American dominance in AI. Again, that probably speaks to why Apple wouldn't be allowed to buy them...

It's interesting that Gurman doesn't mention Safe Superintelligence here – perhaps because it's already valued at $32B. But he himself has noted that Apple has tried to convince co-founder and CEO Daniel Gross to come back to Apple rather than go to Meta in another wild "hackquistion". While Meta is "hackquiring" Gross' VC fund, what if Apple "hackquired" his AI startup and got a guy named Ilya Sutskever in the process?

Would Sutskever ever agree to that? No idea. But certainly Apple has to be more appealing to him than Meta would be. And if he needs to hitch his company to a cash wagon... There are worse ideas is all I'm saying.

Speaking of cash wagons, while Gurman dismissed Anthropic (and OpenAI) as being too expensive for Apple – undoubtedly true, certainly with OpenAI, which is why I never brought them up in my piece, despite their partnership with Apple – the bigger barrier might be the fact that Amazon and Google own something just shy of 40% of the company. Said another way, Anthropic already hitched their wagon to a cash partner. As did OpenAI with Microsoft, until that fell apart in (ongoing) spectacular fashion, so now it's largely in the hands of SoftBank.

But yes, I continue to think that one of the following would be the most interesting deal for Apple to do:

• Perplexity
• Mistral
• Safe Superintelligence
• Thinking Machine Labs

Remember, it's not just about product – as the last two, as far as I know, don't yet have any – it's about talent and showcasing to the market that Apple is serious about AI. I don't think they need to "win it" – certainly there's a case to be made for partnering right now, just as they're doing with OpenAI, and soon perhaps Google and Anthropic as well. I just think they need to be some level of player here. And I don't think they want to leave themselves open to be fully at the whims of others – especially if those others are largely aligned with rivals...

That's the real M&A case.