00:00 Speaker A
I was watching this clip of Michael Dell play where he was talking about on-premises AI. That’s exactly what you’re talking about.
00:06 Speaker A
Well, why is that important? For example, why can’t companies run their operations through data centers run by hyperscalers? What’s in it for them? And when you say on-premises, do you mean physically located in the same building or adjacent to the company that has your servers?
00:27 Speaker B
Oh, those are all great questions. Now let’s tackle the second thing.
00:31 Speaker A
Sorry, there were a lot.
00:34 Speaker B
No, it’s okay, it’s okay. Well, on-premises is traditional, it’s a company’s data center. Although they may be physically located in the same location, we are now seeing more and more companies using what are called colocation centers. Companies like Equinix and others essentially host and physically manage the servers, which are typically owned or leased according to arrangements by organizations like Coca-Cola.
01:04 Speaker B
That’s the on-premises concept. It could be physically there or elsewhere, but kind of under their control. Um, the reason they started doing this is, again, there’s a capacity issue, um, and an issue of how much of this is accessible. But the most important thing is what people call tokenomics: how many tokens will you generate and how much will it cost to generate those tokens?
01:33 Speaker B
And we’ve also seen Dell announce, and Google talked about this at IOL last week (which I attended), the ability for some of their largest Gemini models to run on-premises. Therefore, certain types of workloads can run entirely within their own on-premises environment.
1:53 Speaker B
Most of the biggest frontier models and features will still be accessed on the web, but the idea is what I call a hybrid architecture. I think you and I talked about this the last time we joined. The idea is to start running some of your AI workloads in the cloud, some on-premises, and some on your devices.
2:09 Speaker B
By the way, the other thing about Dell’s earnings is that everyone was shocked by that because PCs, not just servers, smashed their earnings. Of course, HP Inc just announced a very strong quarter in the PC space, and Lenovo also advertised, but they followed up with HP. What’s happening in the PC market is that ASPs are going up due to memory pressure and other reasons. In other words, the number of units is decreasing, but overall revenue is increasing. And in fact, both of these companies were predicting an even bigger year than they originally thought. This definitely surprised a lot of people. Because everyone is out of memory and all those sales are going to get crushed. And the fact is, they’re going to, yeah, they might have fewer units, but you know, they’re going to make up for it in revenue.
