Google & The Great AI Reversal

Nov 25, 2025

∙ Paid

I’m writing this as Google is preparing, for the first time in its history, to turn upside down an assumption that has made it what it is: keep the strategic hardware in-house.

Indeed, Google is preparing to sell its own AI Chip, the TPU, which I’ve been covering for years on my blog and newsletter as the critical enabler of Google’s vertical integration, to the once-enemy, competitor: Meta.

The thing is, we live now in a different world, where the new platform for domination is AI compute, and as such, the TPU is the best lever Google (Alphabet) has to lock in the AI market, in a chokepoint, which can make it the most crucial player for the next decade.

I’ve been arguing in this newsletter for the last three years that if Google started rolling out the TPU to the market, it would make Google Cloud, by far, the dominant player in the AI space.

Well, here we are!

And yet, to understand how we got here, you have to see how this transition played out, what, why, and how Google transformed, and it all started with the ChatGPT moment.

In November 2022, OpenAI’s launch of ChatGPT created what many called a “code red” moment at Google. The company that had pioneered transformer architecture, the very foundation of modern AI, suddenly found itself playing catch-up in the consumer AI race. Wall Street openly debated whether Google would become “AI roadkill.”

Three years later, the narrative has completely reversed. Google’s Gemini 3 model now tops 20 industry benchmark tests. The company’s market cap has surged past $3.6 trillion, surpassing Microsoft for the first time in seven years. Stock prices have risen over 60% since the summer of 2025.

This analysis examines the strategic playbook Google executed to achieve this turnaround, a playbook that offers lessons for any incumbent facing a paradigm shift.