Predicting Google's AI Earthquake

There have been a few times now where Alex Kantrowitz and I have discussed something on his Big Technology Podcast and it has happened; not just later, but sometimes days, sometimes mere hours later. As was the case with Google last week, as we devoted our last segment to talking through why the company was falling behind (again) in AI and what that might mean...

Sure enough, before Alex could even hit publish on the episode, Demis Hassabis was stepping down ("up") as the head of DeepMind and Jeff Dean, Sanjay Ghemawat, Oriol Vinyals, and Quoc Le – 3 of the 4 leads of Gemini, mind you – were also leaving to found a new startup focusing (at first) on what else? RSI.

So yeah, if Google needed a shakeup (again) in AI, they got it!

Anyway, while Alex kicks off the episode with an emergency interlude from his vacation in Indonesia, we then get the actual episode we recorded, starting with Apple's precarious component situation. Sure, they may have looked great to Wall Street in recent weeks for not spending hundreds of billions on AI build-out CapEx, but as their earnings report made clear, there was still a downside to this moment in time. That is: they can't get enough of the components they need for their other devices because of this AI build-out, and those they can get are now insanely expensive (which they're now forced to pass off to consumers).

It's interesting to think that if Apple had spent more time and resources on the AI build out, if they wouldn't be in a better position now simply with regard to the pecking order for parts. And it's sort of wild that Tim Cook, the logistics expert, in his final days as Apple CEO, didn't see this coming...

As for Alex's conspiracy theory on the price increases, watch the margins. But also, watch the messaging around 'Apple Upgrade'.

Speaking of CapEx, how did Microsoft manage to come out of this most recent earnings report looking so rosy? Well, certainly there's some level of AI potentially fueling cloud growth (as was the case with Amazon as well), but there's also some clever accounting at play here. And I don't mean that deceptively, I just mean it's honestly clever the way Microsoft played a hand they could play at a time when such spend matters more than anything else to Wall Street.

How much does it matter? Well, Microsoft had the biggest single day market cap gain in history following the report...

Anyway, how much are these numbers and growth within the Big Cloud players are simply built around Anthropic and OpenAI? I mean, it's certainly worth asking! And is that part of the reason why we're starting to see Microsoft hard pivot their narrative and model (yet again)? Meanwhile, Amazon is closing out their $50B investment in OpenAI ahead of when they apparently needed to...

Finally, Google. What on Earth is going on? Where is Gemini 3.5 Pro? Will we ever see it? It increasingly feels like "no" – it feels a little Llama-y... And it certainly feels like a lot of bureaucratic in-fighting for resourcing and mindshare internally.

But it also could be as simple as Demis Hassabis not fully being bought into LLMs being the be-all/end-all of AI and wanting to continue to push frontiers while Google overall finds themselves in this current AI dogfight around coding and whatnot which are obviously all tied to the current LLM state of the art. And what if LLMs get us to the aforementioned RSI?

Anyway, the situation corrected itself! Or was corrected. Either way, look for Google to come zooming back in a few months – again.