Is AI hype or a promising opportunity?

The short answer is I think it’s both. I think it’s really important to distinguish between short-term hype and long-term investment reality and opportunity. I’ll briefly just talk about both.

On the hype side, clearly lots of media attention. I think you’ve seen a lot of technology or AI related technology companies. You’ve seen their share prices increase extraordinarily over the last 12 months or so. What I think you’ve seen there is a lot of hype, a lot of potential future value creation from artificial intelligence has been pulled forward and is being reflected in today’s share prices and today’s market values of these companies. And if that potential future value creation from AI underwhelms, then you could argue that there is actually some, some hype around AI at the moment. I think it’s also important to remember that as of today, there is currently a low understanding and low deployment of artificial intelligence outside of several high profile cases, like embedding these large language models into internet search. So for AI models to become more advanced and more widespread in their use, these models will need access to data sets over and above publicly available internet data. And Joss, that is what chatGPT was trained on, was publicly available internet data. To take narrow artificial intelligence and turn it into general artificial intelligence, these models will need access to financial data, healthcare data, personal, proprietary, fragmented, hard to access, hard to replicate data.

Just in terms of the long-term investment opportunities, 2 areas. We think the most investible areas today in AI come first of all in the companies providing the computational processing power. So semiconductors: semiconductors are the brains, they’re the building blocks of ai. There is an insatiable demand for AI compute right now, both at the training and at the inference or decision making stages of building these AI models.

The second area where we believe there is real long-term investment opportunity as of today are in the AI infrastructure providers. So if semiconductors are the brains behind AI, companies providing the infrastructure are the plumbers, provide the plumbing. So infrastructure providers to AI would include the public cloud hyperscale data centers on which AI workloads run. These are the companies providing access or AI models’ access to the most advanced compute in the world as of today. Finally, we think there are very few opportunities or limited opportunities in the developers or the makers of these AI models.