Today, Google is the digital “front door” for 5 billion consumers who use it for search monthly. But it’s pretty clear the search landscape is about to be upended by AI chatbots. You might think that’s already happened - ChatGPT, Claude or Gemini might be your daily driver, having already replaced much of your use of old-school search. But the reality is you’re an early adopter, and this shift is still nascent. An AI chatbot hasn’t become a habit for most internet users, if they’ve even tried one.

Google built one of the worlds most valuable businesses by giving advertisors the opportunity to get in front of consumers at just the right moment. They can do this because, in a fairly course-grained way, they can understand “intent” from the search query a user has entered. In a world where AI assistants overtake traditional search, that understanding of intent is only going to become vastly better, and vastly more valuable.

I think the future of consumer-facing AI assistants looks somewhat like search - a free product supported by ad and partnership revenues. If that’s right then Google looks better placed than anyone else.

Here’s why - if we walk through their “AI assistant stack” it’s unrivalled at all layers except the application layer (admittedly crucial - but also easiest to fix).

From the bottom up, they have the only custom AI silicon that can rival Nvidia. Their TPU v5p chips now rival or exceed NVIDIA’s H100 in raw throughput, memory, and interconnect bandwidth.

At the data center layer, Google are a “hyperscaler”, being the world’s third‑largest cloud provider. They’re well placed for a massive data center build out that can keep up with demand for AI training and inference

The Gemini foundation model is close to or on par with anything OpenAI or Anthropic has to offer. And in DeepMind Google have one of the world’s leading AI research labs.

It’s at the application layer where Google are lagging today. Gemini usage lags ChatGPT. But these are still early days, habits are not formed, and many users haven’t even tried out this tech yet. Google can borrow or acquire missing UX pieces as they emerge, but will need to overcome the inertia of a corporate giant.

At the top of the stack, to monetise a free consumer product you need massive distribution and advertisor partnerships. Google are way ahead. They have thousands of account executives who sell into almost every significant enterprise globally. An AE might be selling AdWords right now, but in the near future may be selling some sort of ad placement within a Gemini interface. Meanwhile, 13.7 billion daily searches speaks for itself in terms of consumer reach.

Google are so well positioned that even if they just fast-follow ChatGPT’s product innovations, they should win this space. If OpenAI or another competitor do end up being the dominant consumer assistant a few years from now, it’ll be interesting to look back and see just how Google managed to slip up from such a strong base.