What have you learned about the agent workload in particular that is leading to different design decisions?

Yeah.

So let me start at, like, 10,000 feet, and then I'll give you a couple of specific examples.

The simplest answer is that agents don't sleep, right? We're now living in a world where those agents are constantly feeding the accelerators, constantly reacting.

They are spawning additional agents.

So just because you spawn one agent doesn't mean one agent exists.

Every agent could spawn ten, a hundred, a thousand agents, and each of those could spawn a bunch of agents to kind of go and fulfill your request.

Um, the, I think everybody can agree on is the Gemini sucks.

And I'm wondering in a firm which obviously has those kind of very high, uh, intelligence hurdle problems, do you actually see this uplift from model generation to model generation?

Yeah, I mean, I think you hit the nail on the head.

I think it really comes down to the c- the complexity and the, you know, the sophistication of the engagement with the model.

Um, you know, I-- Listen, if y- if, you know, if I'm at home and I'm, you know, uh, sitting in front of my, my, uh, my, my laptop and I'm just kinda interacting with a chatbot, the model difference oftentimes isn't gonna make a huge, uh, isn't gonna make a huge difference.

Um, for those more sophisticated users, as we start to think about the complexity betwa- uh, around, you know, optimizing software or going off and running complicated a- agents in a, in an engineering environment, um, yeah, you absolutely do notice the difference.

You absolutely do start to notice it.

And it's not linear.

What have you learned about the agent workload in particular that is leading to different design decisions?

Mm

... which is relying on Arm's technology in some way, what they're clearly saying to you, and which is why they all kind of stood up and, uh, supported us, what they're clearly saying to you is that the strength of the Arm ecosystem in this market ultimately expands the pie for everybody.

And I think that's, that's the net effect.

So whether you're a hyperscaler and you're looking to ensure that your customers on-prem, you know, can now have access to the technology so they can adopt more of your technology in the cloud, whether you're an ASIC provider who's off building custom silicon and looking to, uh, to create software synergy that you can kinda, kinda leverage across a broader software ecosystem.

I mean, regardless of where you sit in that value chain, if you're leveraging, if you're building your own silicon and you wanna go leverage some of the tools and software that we're putting forward into the OCP ecosystem, what you're seeing in every one of those cases is that when people objectively look at it, they think, "Hey, there is nobody competing in this space.

They're filling a niche in the market which hadn't been filled yet.

And oh, by the way, this is gonna strengthen the overall ecosystem."

Yeah.