Cameron Gleeson

When we move to a world where some of these agents are working independently or always on, you might be talking about consumption of, say, 100,000 tokens.

You can just see how the growth in the demand for AI compute is leading to this huge potential for monetization.

That's where we sit right now.

Yeah.

Incredible, isn't it?

Yeah.

And it just shows, like, the power of AI because, you know, not only are they producing, you know, this technology, but they're also using AI to build, you know, the engine.

So it's almost like this sort of self-fulfilling prophecy in a way, and that allows them to operate a very light machine.

A very light model.

And clearly one of the areas where AI has proven itself is in terms of a productivity dividend has been in code, has been writing software.

You talk about other use cases, whether it be in terms of maybe customer service.

And, you know, there are, you know, advances in terms of, you know, being able to deliver really good solutions there or potentially, you know, data entry.

But coding, you know, given the nature of coding is very logical, AI has absolutely excelled in that.

And an organization like Anthropic is able to use that to its advantage in terms of building the leading models.

And, you know, obviously the Opus model with Claude there is a product of that.

Yeah.

Yeah.

I guess there's a few things that you're looking out for.

And one thing is clearly, you know, is this massive investment.

And I think the expectations for CapEx this year from the large US hyperscalers sits north of 700 billion US dollars, which is enormous.