How Can China Win the AI Race Against the US?

Joining us now. You you heard what Shepard was was reporting on there. You know, we often frame, The US China AI race in terms of who has better chips, who is doing the most work on advanced models. But access to capital seems an important swing factor right now. I mean, that's always been the case that capital matters and particularly in a high CapEx or CapEx intensive industry like artificial intelligence. I'd I'd maybe take a step back. I think for long term investors, what really matters is more finding outliers, and those outliers will then naturally have greater access to capital because they are fundamentally better businesses. In some ways, I actually think I might be a bit of a boring guest for you today because I'm gonna be repeating quite a lot of older information rather than giving you new news. But I think increasingly market participants tend to be conflating new and important, and it's particularly easy to do in AI given the pace of change and growth, and it's exacerbated by structural market dynamics, I think. You know, 60% of US investing is done through passive vehicles. 75% of trading volume is from quant funds, and there's increasing retail participation. So markets have never been faster and noisier in being able to take that step back and focus on the fundamental attributes of a company. What are what industry are trying to disrupt? What new industry are they creating? Are they investing in innovation over the long term? And long term here means five or ten years because true change takes decades and not quarters. I think those are the kinds of businesses that are going to grow well in the long term regardless of which industry they're in. Okay. So I'm happy to park the new for a second and stick with the important. I did spend a bit of time going through the holdings in your fund. So two interesting parallel case studies is TSMC is a top holding right as is SK Hynix on the memory side. The case studies in that big take they're outlined are CXMT on the memory side compared with SK and also China's efforts to have domestic manufacturing capacity. How would you sort of interpret that? I think China's been a tremendous engine for growth and disruption and innovation for a very long time, and I certainly wouldn't want to bet against China. Indeed, we have holdings in China such as BYD and Pinduoduo and Tencent, and they're tremendous companies. But I think companies like TSMC and SK Hynix and ASML are almost in a league of their own. TSMC in particular has created effectively monopoly at the leading edge of chips, and that's going to be very difficult to disrupt, not just because of access to capital, which is a a limiting factor still given the amount of money it takes to build a leading node fab, but also because there's so much inherent process knowledge, that TSMC has built up over decades. And that is incredibly hard to replicate even if Chinese companies had access to a lot of the equipment that they need, which they simply do not because of export controls. I think a similar story, is probably going to happen with SK Hynix where a growing proportion of their revenue and earnings are coming from HBM, high bandwidth memory, which is very complicated to make. Again, requires quite complex processes and equipment, and CSMT doesn't actually have access to a lot of that equipment yet. Now will they be able to replicate it over time? Perhaps. But that will take If if I was gonna give it the sort of umbrella label of the AI trade, you know, your non US focus is TSMC, SML, SK Hynix, very, very chip concentrated. Where is the non US AI opportunity outside of semiconductors that you look at right now? I mean, there there's a there's a number of them, but I'd perhaps pull out two really exciting ones that actually reported relatively recently. So I am going to give you some new news. Please. And that's Shopify on the one hand and MercadoLibra on the other hand. I'd contrast them again because I think that there's there's some recognition that what Shopify is doing with AI is bearing fruit. So, fundamentally, the immediate complexity for merchants and consumers, and what AI does is it introduces ever more complexity into the shopping journey. So increasingly autonomous agents are capable of making decisions for consumers. That means that you need more of a source of ground truth for those agents to be able to rely on when offering up potential purchases to their users. So catalog Shopify's record of all the items that merch merchants has and all the metadata associated with that is showing two times higher conversion on orders than simply scraping website data when AI agents are involved. And that's that's tremendous for a company that, you know, monetizes based on GMV. On the other hand, I put MercadoLibre where it's almost a classic example of, again, a lot of investors are not taking the long term view here because what MercadoLibra is doing is something they've done multiple times in the past where they're investing substantially in growing users because they see this as a really critical inflection point in the markets. And so they're seeing a very strong GMV growth, very strong user growth on the back of, lower shipping, requirements, and lower take rates for for merchants. And at the same time, they're investing in AI in order to speed up the cadence of development internally. So they've seen 75% higher deployment of code internally at the same time as the number of rollbacks has gone down. So not only are they moving faster, they're moving at a higher quality. And I think that's that's really exciting, and the market probably isn't recognizing that.