On 5 November, markets gave a harsh first verdict on the elections: the Philadelphia Semiconductor Index fell 12.3%, as investors weighed what Democratic gains in state governments could mean for the cost and pace of the data centre build-out, at a time when local communities already question its impact on electricity bills.

AI assistants are challenging Google's dominance in Search. That story is now well known. Far less attention has been paid to what Alphabet has been building on the other side of the AI race: a full stack stretching from its own chips to products used by billions. At its heart is the Tensor Processing Unit (TPU), a chip developed for internal use a decade ago that has since become a product rival that AI labs are willing to pay for. This combination is difficult to replicate and is already driving growth. The catch? The economics are far less attractive than those of Search, with significantly lower profitability.

Alphabet, OpenAI and the battle for the world’s most valuable ad auction