Company Interview / M7's race for super-intelligence

Tony Wang, from T. Rowe Price, offers his analysis on the current state of the AI (Artificial Intelligence) market. Tony suggests that market struggles are connected to the immense spending by hyperscale players such as Microsoft (NASDAQ: MSFT), Google (NASDAQ: GOOG), and Meta (NASDAQ: META). To Tony, digital advertising and Tesla's (NASDAQ: TSLA) self-driving project, powered by GPUs, are the immediate promising uses of AI despite significant capital expense, with a potential $200 billion spent on AI infrastructure so far. However, for the larger tech firms, he sees it as a crucial race for superior AI capabilities, too significant to slow their investment pace.

As for individual tech giants, in Tony's view, Google remains committed to AI as it could transform search functionality, while Microsoft can feasibly utilise AI across their office suite for significant improvements. Regarding Meta, Tony notes they have seen improvements from digital advertising, and hence, their large spending is likely to continue.

On the topic of Apple (NASDAQ: AAPL), Tony explains the company's unique position as it integrates AI into its products. Apple chose to be the ecosystem holder, enabling a lower capital expenditure approach compared to other companies. However, Tony questions whether the market's expectations align with the potential gains AI can offer. Touching on Google, he discusses the regulatory issues they face in the wake of a potential breakup. In conclusion, Tony shares that at T. Rowe Price, they are focusing on large cap companies expected to benefit from AI advancements, including Meta, Nvidia and Microsoft.

Full unedited transcript:

0:00

I think that's a totally fair question that the market is currently struggling with. Um, and just given the amount of spending and the capital spending that the hyperscale players are doing, I think that's like a natural question. I think that there are like really interesting use cases. Um, but it takes time. Um, I think that one of the big use cases that we're seeing right now that is working, though, is digital advertising. Uh, as well as right now, Tesla is training full self-driving with, um, you know, GPUs. So I think those are like kind of the lead cases here. I think there is a fair amount of experimentation that is, uh, you know, fueling kind of this, this growth in compute spend. But I think another aspect on top of that is just like, is this kind of race for super intelligence among the big tech firms, um, just too important to, to, you know, take the, the foot off the pedal. So I think that's actually kind of what is more important for the go forward. I think that just given the amount of CapEx spend, like I could tell you, a thousand

0:59

examples probably of enterprise use cases, but I don't think that they would stack up to the $200 billion that's being spent on AI infrastructure. So I think in the meantime, like, you know, the market's probably just struggling with like, you know, how willing are the hyperscale players like Microsoft, Google, um, you know, Oracle, uh, meta willing to spend to, to, uh, you know, kind of have that, um, first place seat in the AI race. Well, given it is a race as such, then Tony juicy that there are winners and losers, um, certainly longer term.

1:37

Yeah. Well, I think that, uh, they're all building large language models that kind of do the same thing. And, you know, in terms of, you know, requiring a lot of compute being a broadly applicable, um, I think that we are kind of in this AI space race and, um, I think it's too early to call winners and losers. Um, but, you know, we are in a position where the gains are could be like astronomical. And there's two important to lose. I think Google on one hand, uh, is is really focused on AI because search could be transformed here. And so I think it's imperative that they continue to invest. I think Microsoft has, um, you know, ability to roll out AI across their Microsoft, uh, office suite, you know, via copilot. And if that works is there's huge rpu, um, improvements there. And so in terms of meta, I mean, they are seeing the improvements there, um, with digital advertising. And so, you know, I think that it it's likely that they'll continue to spend

2:36

in, you know, very large kind of amounts. But I think the question is like, are the expectations, um, you know, low enough so that the stocks can keep working relative to what's going on? How are you viewing Apple at the moment? Obviously, is it seeks to integrate AI, particularly into its devices.

2:55

Yeah, I think Apple is in a very unique position where they've chosen to kind of be the ecosystem holder. Um, and as a result, they can kind of go at it with a relatively low CapEx uh, way like they did, probably in search. So, you know, they didn't have to invent the best search, but they can integrate it with their phones, and then they're able to take kind of their, um, kind of traffic acquisition costs and attack payment from Google. And so I think with, um, AI and generative AI, they've chosen to integrate the best ChatGPT right now for their users. And so I think there's a they're in an interesting place where they can have a low CapEx way to integrate generative AI into their phones. And, you know, as a result, like, they can kind of cause a large upgrade cycle. Their phones are already really, um, kind of their installed base is really aging right now. And so there's an opportunity to upgrade since you need to have the iPhone 14 Pro, um, or above to access the features.

3:55

And so I think it gives a good opportunity that for consumers to see something new, for developers to have a new opportunity to unlock new features in the iPhone. Um, and Apple is kind of in a good position where they don't have to spend, uh, you know, 50, $60 billion of CapEx every year in order to, to have, um, that technology.

4:15

Tony, you mentioned, Uh, Alphabet's Google. There a couple of issues at play, I guess most immediately is the, um, the regulatory threat, given the Department of Justice is looking about a potential breakup there. What should we make of that?

4:30

Yeah. Well, you know, it's interesting. Um, I think there's a lot of different permutations that can result in that. Uh, I think in the near term, if Google doesn't have to pay Apple for, uh, you know, the traffic acquisition cost as much for it, uh, there could be a little bit of savings. And so there could be like, opportunities for earnings, um, improvement in the near term. Unfortunately, I think that comes at the detriment of, uh, the long term terminal value at a time when, you know, searches are already kind of under pressure, perhaps from these generative models like ChatGPT, uh, or perplexity, for example. And so I think that they're happy with the current agreement, otherwise they wouldn't be, uh, paying Apple this much money. And so I think any type of uncertainty around the terminal value for Google on search, if they lose a few points to being, uh, because of this case, I think that could be viewed negatively by the market more than perhaps the savings on paying. Um, you know, Apple or that traffic.

5:30

You know, the practicality of it is that this kind of happened in Europe, and most users did choose, uh, to to stick with Google because it is a superior search product.

5:41

Tony, overall, can I ask what your preferred play there is at zero price in terms of AI, which companies you're focused on?

5:52

Yeah. So I mean, I think that there's um, you know, several of the large cap companies look to benefit. You know, I like Medicare. Um, they've been able to kind of

6:02

kind of grow off of tough comps. Um, and largely because I think they've shifted from CPU to GPU, and they've got kind of really nice advertising platforms and ecosystems to provide a lot of value. And I think we're seeing kind of tick tock share gain, um, kind of kind of, uh, even out here and so I think they're a good beneficiary of digital advertising, and it's kind of like they're they can be front footed. Um, and gain share here. Uh, obviously we own a lot of Nvidia as well. Um, it had been a, you know, we really liked that company for, um, the last seven years. And so they continue to be the tip of the spear on, on AI celebration and GPUs and the leaders of the kind of AI platform ecosystem. Um, and then, you know, we also own a good amount of Microsoft. Um, you know, they kind of have the, the, the leading position within Azure. And being the thought leader with their connection to ChatGPT open AI integration. And so, um, you know, I think those are

7:02

there's something those that would, I would highlight, I mean, there's other, um, kind of smaller cap, uh, companies like, you know, I, I do like apple oven, for example, and Zeta Global where, you know, I think you're seeing that generative AI, you know, meets like proprietary data sets really, um, produce good returns in improving returns for advertisers. So I think there's, uh, you know, we we own several AI beneficiaries, but I, I'd highlight a few of those, uh, for as examples.