If you can’t beat ’em — or suspect they might pose a challenge at some indeterminate future date — buy ’em.
There are ostensible legal hurdles to the aggressive execution of that strategy because more times than not, it comes at the expense of competition. Less politely, it’s monopolistic.
But who doesn’t love a monopoly, right? Shareholders are certainly ok with them, particularly if they buy back a lot of stock. And when you mint money, you can afford great legal counsel, which helps inoculate you in the (unlikely) event the government gets around to scrutinizing your business.
With that in mind, and without casting aspersions, Nvidia decided to — how should I put this? — preempt prospective competition with a massive investment in AI startup Groq, not to be confused with Grok, or “MechaHitler.”
Jensen Huang’s spending — and this is pretty remarkable — $20 billion to license the company’s technology and poach hire its top executives.
In a short press release, Groq described the tie-up as “a non-exclusive inference technology” deal which “reflects a shared focus on expanding access to high-performance, low cost inference.”
“Shared focus” indeed. As The Information put it, Groq is “one of the best-funded startups trying to challenge Nvidia’s dominance.”
Now, Huang will be able to incorporate — “exclusively” or not — the company’s design innovations into his product offerings, an integration effort which should go swimmingly given that the startup’s CEO, its president and “other members of the team” now work for Nvidia.
Although Groq will carry on as a stripped-down, independent entity, the scope and (sparse) details of this deal are eyebrow-raising. In their coverage, The Information used the word “stunning,” noting that at least for now, no one’s sure what the $20 billion headline “licensing” fee actually includes, nor how it’s structured.
What we do know is that Huang just agreed to pay triple the startup’s most recent valuation (from September’s Series E round anchored by Dallas-based Disruptive, Blackrock and Neuberger Berman) just to license its technology and acquire its leadership team, including founder Jonathan Ross.
As Bloomberg helpfully noted, Ross was instrumental in Google’s Tensor Processing Unit development. That technology (the TPUs) was used to train Alphabet’s Gemini 3, which shook up the AI universe late last month by outperforming OpenAI’s models on most key industry benchmarks.
Disruptive’s chief Alex Davis told CNBC that Nvidia’s “getting all of Groq’s assets” except for its “nascent cloud business.” In an internal memo, Huang said Nvidia’s “adding talented employees and licensing Groq’s IP,” but isn’t “acquiring Groq as a company.” From what I can tell, that’s a distinction without a difference.
In addition, it looks like the succession plan to Jensen Huang, as CEO, just got a little more clarity. Always good to have someone really capable to hand over the reins to. Given that Jensen Huang’s share of Nvidia is worth $165B, paying $20B to secure stewardship for that wealth doesn’t seem unreasonable. Now there will be plenty of time to get Jonathan up to speed on “all things Nvidia”.
In the future, when the machines are doing all the work, and humans have nothing left to do, Groq will give us some Grog.
I have been expecting Apple–with that mountain of money it is sitting on–to make a similar kind of move for some time now. I can now envisage a WWE Battle Royale/”Beyond Thunderdome” type scenario where some twenty AI combatants enter the cage and only a handful manage to survive. It is all very 1990s.
Or wait until this initial stage of leverage collapses under its own weight and buy up some assets at nickels on the dollar. Why nickels? Because inflation…and we don’t make pennies anymore.
Apple’s strategy here has been maddeningly slow to reveal itself. There’s a lot of speculation as to why but my best guess is that — keeping it simple — Tim Cook is a finance guy and he’s playing the long game. Apple’s “privacy first” model makes monetizing early AI difficult. And overpaying for IP would make it even more so.
Siri was below Apple’s standards for most of her product life. I’m hoping they’re taking their time and the replacement will actually be worth the wait.
If one wishes to know what Siri’s replacement is like, you can discuss it with your Model Y on your next trip to the grocery store.
If a company, or any of its assets, is not technically acquired then are there any regulatory requirements?
That was exactly my question when I finished reading this.