Simon [00:00:09] Welcome to this Global Innovation video, in which Storm and Clare are going to talk about the latest developments in AI infrastructure. Storm, let's start with you. You've been talking about AI infrastructure over the past couple of years. Can you just remind us why it's so important and what's been happening recently in this field?
Storm [00:00:27] AI infrastructure has seen a tremendous buildout over the last two years. We fundamentally believe this is the fourth industrial revolution and it's going to require a buildout which will take at least 10 years before we move into a cyclical type market. The reason why there's so much structural demand for AI infrastructure is because of three different reasons. One is because the larger the AI infrastructure, the smarter the models. Then what we can do is we can use a lot of AI infrastructure to then train and focus these models on doing particular tasks. That's called reinforcement learning. The next major vector of demand for AI infrastructure is reasoning. So that's if you give a task to an AI system, it does that task for you. Maybe it takes you two weeks to normally do that task. Maybe it take the AI system about 30 minutes. But that requires a significant amount of compute. So these three major vectors of demand mean that over the next five years, it requires a very large infrastructure buildout. We've had a significant breakthrough with AI systems, with Grok 4 coming out. We also know that OpenAI has also made another major breakthrough as well. What that means is there's going to be broader adoption across the entire economy. So that's a fourth vector of demand for AI infrastructure compute buildout. This is the reason why we're five minutes into the football match for this particular infrastructure buildout. [26.0s] This is a reason why we believe it's going to be the biggest infrastructure buildout in history. So what is AI infrastructure? Well it's data centres, it's GPUs, it's CPUs and it's networking. This is what we don't have enough of and this is what needs to be built out over the next five, ten years.
Simon [00:02:05] Clare, Storm talked about the pace of development in this area. How does this affect who's going to be the winners? Are you seeing a change in the winners?
Clare [00:02:13] We're certainly seeing a change, Simon. And we're of the firm belief that the winners of this new technology cycle are going to be very different to the winners of the last technology cycle. We don't hold the likes of Alphabet, Microsoft, Amazon, Apple, across any of our Funds. When we're confronted with platform technology transitions, what we see time and time again is that infrastructure demands change. This happened in the era of the internet, it happened with the cloud, and we're now seeing this happen with the AI. So there's a new cohort of companies that are taking market share from incumbents when we're talking about AI-specific workloads. If we think about the hyperscalers who actually provide the compute to end customers that Storm was just talking about. They provide the computer networking for customers, enterprises, to rent. Oracle has got about 5% market share of cloud revenues. So it's pretty small. To put this into context, AWS, Amazon has got 44%, Microsoft 30%, Google Cloud 20%. Yet when we look at AI workloads and GPU allocations from NVIDIA, Oracle has about a 20% share. So they're taking a disproportionate share of AI workloads. And the reason for this is that they've purpose-built their infrastructure for accelerated compute. They've mastered something called rack technology, something the other hyperscalers have failed to master to date, and this delivers performance for AI workloads that is 50% to 60% cheaper, 30% to 40% faster. CoreWeave similarly is taking a disproportionate share of AI workloads. They're a neo-cloud, which again delivers about a 40% improvement versus Azure or AWS. Why this matter is going forwards is that we're seeing very, very large projects being announced by sovereign AI customers. These are brand new customers. This is an entirely fresh market, and they're not hindered by legacy data centre footprints. They are allocating their projects to the infrastructure providers, suppliers, that purely deliver the best technology on a price performance basis. So we think that this market share shift is going to continue and actually gather pace over the next five years.
Simon [00:04:46] Storm, you've both talked about the pace of development in this area and the pace of demand. Is this underappreciated? Is any aspect of AI infrastructure underappreciated, do you think?
Storm [00:04:56] Yeah, well that's the best thing about an emerging new industry that's very different to the past in that there's new winners going forward. So the initial winner was obviously NVIDIA, providing the GPUs for accelerated computing that enables AI systems to work on top of. As the AI infrastructure builds-out and becomes more important, there's new winners that emerge within this buildout. So originally, about 50% of the value accrued to the GPU, as these GPU clusters get larger, more of that value goes to networking. In fact, we heard from Alphabet that a third of the cost for CapEx going forward is actually going to networking. The reason why is because as these clusters get larger, and then they get packed closer together, there's more networking required. So this is companies like Broadcom, other companies like Credo, and companies like Amphenol. We're building an entirely new type of compute infrastructure, and the way I like to sort of draw a contrast is really traditional compute is like a bicycle, accelerated computing is like a Ferrari. Completely different types of infrastructure and that's the reason why there's new winners going forward compared to the past.
Simon [00:06:08] So many investors talk about companies having moats to protect themselves. Are there any companies developing that in this field or is it completely open?
Clare [00:06:18] It's pretty early days for the majority of companies in the space. NVIDIA certainly has built out a very formidable moat and that is because of their full stack offering. So a lot of those companies that Storm just mentioned are the networking companies, the likes of a Broadcom, Credo, and Amphenol. These mostly support an open architecture when it comes to networking, which has really been the status quo over the past decade, think about traditional ethernet. Now, NVIDIA has got its own networking business. We actually think this is an underappreciated growth lever for the company in our view. But NVIDIA has traditionally been a closed ecosystem. You buy the GPUs, the AI chips, you also have their CUDA offering, which is their software, and networking is really the final component of this full stack offering. So you can view this opportunity as a competition between the NVIDIA ecosystem and the open architecture players. Ultimately, this market is so vast, we think that both are going to win, that is NVIDIA's advantage of being able to fuse together all of those components and deliver it to customers with a price performance advantage.
Simon [00:07:44] Thank you Clare. Thank you Storm. And thank you for watching, we'll see you for the next Global Innovation video.
KEY RISKS
Past performance does not predict future returns. You may get back less than you originally invested.
We recommend this fund is held long term (minimum period of 5 years). We recommend that you hold this fund as part of a diversified portfolio of investments.
- All investments will be expected to conform to our social and environmental criteria.
- Overseas investments may carry a higher currency risk. They are valued by reference to their local currency which may move up or down when compared to the currency of the Fund.
- This Fund may have a concentrated portfolio, i.e. hold a limited number of investments. If one of these investments falls in value this can have a greater impact on the Fund's value than if it held a larger number of investments.
- The Fund may encounter liquidity constraints from time to time. The spread between the price you buy and sell shares will reflect the less liquid nature of the underlying holdings.
- Outside of normal conditions, the Fund may hold higher levels of cash which may be deposited with several credit counterparties (e.g. international banks). A credit risk arises should one or more of these counterparties be unable to return the deposited cash.
- Counterparty Risk: any derivative contract, including FX hedging, may be at risk if the counterparty fails.
The issue of units/shares in Liontrust Funds may be subject to an initial charge, which will have an impact on the realisable value of the investment, particularly in the short term. Investments should always be considered as long term.
DISCLAIMER
This material is issued by Liontrust Investment Partners LLP (2 Savoy Court, London WC2R 0EZ), authorised and regulated in the UK by the Financial Conduct Authority (FRN 518552) to undertake regulated investment business.
It should not be construed as advice for investment in any product or security mentioned, an offer to buy or sell units/shares of Funds mentioned, or a solicitation to purchase securities in any company or investment product. Examples of stocks are provided for general information only to demonstrate our investment philosophy. The investment being promoted is for units in a fund, not directly in the underlying assets.
This information and analysis is believed to be accurate at the time of publication, but is subject to change without notice. Whilst care has been taken in compiling the content, no representation or warranty is given, whether express or implied, by Liontrust as to its accuracy or completeness, including for external sources (which may have been used) which have not been verified.
This is a marketing communication. Before making an investment, you should read the relevant Prospectus and the Key Investor Information Document (KIID) and/or PRIIP/KID, which provide full product details including investment charges and risks. These documents can be obtained, free of charge, from www.liontrust.com or direct from Liontrust. If you are not a professional investor please consult a regulated financial adviser regarding the suitability of such an investment for you and your personal circumstances.
A collective redress mechanism by consumers in respect of infringements of applicable Irish or EU laws is available under the Representative Actions for the Protection of the Collective Interests of Consumers Act 2023 which transposes Directive (EU) 2020/1828 into Irish law. Further information on this collective redress mechanism is available from Representative Actions Act - DETE (enterprise.gov.ie).