Transcript: Growing bottlenecks pose challenge to AI frontrunners
Kathrin Forrest of Capital Group says constraints include supply of human capital and computer chips, and the build-out of data centres
- Featuring: Kathrin Forrest
- August 31, 2025 August 31, 2025
- 10:01
- From: Capital Group
Welcome to Soundbites, weekly insights on market trends and investment strategies, brought to you by Investment Executive and powered by Canada Life. For today’s Soundbites, we’re talking about artificial intelligence with Kathrin Forrest, equity investment director with Capital Group. We talked about challenges facing the sector, and industry collaborations. And we started by asking about why tech has proven so resilient in the turbulence of 2025.
Kathrin Forrest (KF): You know, it really has been remarkable. During the first quarter, the IT sector within the MSCI All Country World Index was down 11.6% in Canadian dollars. But then it recovered meaningfully in the second quarter, where that same sector subindex was up 16.7%. And looking a little bit closer, market confidence essentially reasserted itself, supported by a couple of points, including the ongoing momentum of the AI infrastructure build-out. Capex — or capital expenditure — plans were reaffirmed, including by some of the largest U.S. tech companies. Four of those have a joint capital expenditure budget of over $300 billion just this year. And with that, just the strong earnings momentum that we’ve seen. The S&P500 IT sector is reporting the second-highest year-over-year earnings growth rate of all 11 sectors, at 21.1%. Semiconductors, within that sector, reporting earnings growth of 33%.
Challenges
KF: A framework that we use to think about that is “the AI stack.” The bottom layer: compute. On top of that: infrastructure. Models as the third layer. And then applications at the top. The task at hand, right now, or one of the key tasks, is scaling up infrastructure and compute. Some of the key challenges are managing bottlenecks around this build-out. Certainly leading-edge chips have been a bottleneck. Land, data centre architecture, all of those have turned into bottlenecks. Around data centre, electricity as a bottleneck. And just the expected power demand growth coming from data centres. And then, around that, we need different types of cooling mechanisms. Liquid cooling is one of the ways to go about that. But that then comes with other bottlenecks, including water, for example. Maybe the last bottleneck to mention: specialized human capital. And that’s anything from welders, electricians, plumbers, all the way to AI researchers. The second challenge: the ongoing demonstration of the economics around this. Clearly, there’s a lot of money being spent. Can we validate the value of that investment? And then, the third challenge, the risk of being left behind, because innovation is blurring competitive lines, and it’s about weighing the risk of spending too much versus the risk of spending too little.
Collaborations
KF: There are a number of examples of collaborations that point to finding new market opportunities. And then also collaborations to address some of the bottlenecks that we discussed earlier. So, in terms of new market opportunities, in IT one of the leading large language model developers is partnering with a large e-commerce platform, and together they’re testing checkout flows within their AI chat bot, allowing users to complete purchases directly in the chat interface. This streamlines the shopping experience and introduces new revenue opportunities. Another example is in the wearable space to combine healthcare, consumer and technology features. That includes the launch of smart glasses that can make calls, that can send texts and do live streaming. And then, looking at some of the bottlenecks, power and electricity appear at first glance to be local issues. But there are a number of examples of collaborations across geographic boundaries, where companies outside of North America are forming partnerships with North American peers.
Tangential benefits
KF: I think there are opportunities for all sectors and industries to see benefits. And over time we might see those benefits move from tangential to more core. Within pharmaceuticals, there are opportunities to accelerate drug discovery through predictive modeling. There are also opportunities for clinical trial optimization. Within energy, you can find examples for predictive maintenance, which creates opportunities to improve operational efficiency and decision making. Within transportation, you can see examples around flight operations and delay management. And then looking beyond that in consumer sectors, you can find examples around inventory management, through tracking product expiration and suggesting promotions to reduce food waste. You can see examples for personalized recommendations and targeted marketing. And then within financials, examples where companies are applying AI tools to facilitate fraud detection and risk management, as well as process automation.
And finally, what’s the biggest takeaway for investors?
KF: There clearly is risk of short-term volatility and overinvestment, as we saw earlier this year. It’s important to pick your spots. The bigger picture really is around broad economic transformation. It’s not necessarily an IT-specific theme. It’s a foundational technology, a platform technology. And if you take a long-term view and analyze companies and their opportunities over the long term, you can find some really interesting investments that are not constrained by sector or geography.
Well, those are today’s Soundbites, brought to you by Investment Executive and powered by Canada Life. Our thanks again to Kathrin Forrest of Capital Group. Visit us at investmentexecutive.com, where you can sign up for our a.m. newsletter and never miss another Soundbite. Thanks for listening.
**
Go back to the article.