Alice, I mean, obviously, they still face significant constraints on the hardware side, access to the best chips, etc., right? But it hasn't stopped them from developing some pretty remarkable LLMs. And you know, you and James on your pod, China Geopolitics, were just talking about this recently, and you noted that up in Silicon Valley, you know, some studies have shown that probably half the LLMs that folks are using there are Chinese LLMs. And it's pretty easy to see why they're so attractive. It's because they're either extremely low cost or in many cases free. Does this surprise you that it's happening in Silicon Valley?
No, I mean, in many ways, Silicon Valley is motivated by market forces. And the fact that you have these Chinese models that are generally open weight, meaning that it's pretty easy to access. Yes, we don't see the underlying code and data on basis in terms of some of these LLMs, but they're pretty easy to access given that they tend to be open weight and sometimes open source. This makes it easy to build on for a lot of these developers in Silicon Valley. And again, much cheaper, orders of magnitude cheaper than what you can see coming out of OpenAI, out of the big LLMs in the states. So, it doesn't surprise me that developers and even CEOs in Silicon Valley are using Chinese LLMs. What I think will and expect to happen is that there will be a political movement amongst the more hawkish elements in DC to try to block this. Because they do believe that there are security, data security, and privacy concerns associated with US companies, US officials using Chinese LLMs. I think that this gets politicized pretty quickly in 2026. But to your question, Martin, it doesn't surprise me. Silicon Valley has always been motivated by market forces.
And sorry, I didn't mean to jump in there. I thought we'd finished. Alice, may I jump in here as well? You know, pertinent to the chip conversation is of course Taiwan. And you know, given that conversation overnight between Presidents Xi and Trump, President Xi made it certainly very, very clear that his aim is still to reintegrate Taiwan back into mainland China. What do you think, Alice, is Trump's stance towards Taiwan? In other words, if mainland China did make a move on Taiwan, what do you think the United States would do?
Well, this is I think the biggest strategic question and challenge that the Trump administration faces. And frankly, the biggest challenge at the heart of the US-China relationship for the foreseeable future. It is clear that Xi Jinping prioritizes Taiwan as part of his legacy. Whether or not he acts during the Trump administration remains to be seen. But I think on balance, the Trump administration is far more reluctant to sacrifice blood and treasure when it comes to the Taiwan question than his predecessors. I think that the Trump administration and Trump in particular are more interested in rebalancing the trade relationship, trying to find areas in which they could potentially work together, maybe on fentanyl, for instance. I don't sense from this administration, especially from Trump himself, that Taiwan is going to be top of mind for him. In fact, he's trying to extricate himself from these global commitments. So, I think in the next few years, it could well be that the Trump administration may subtly hint that Taiwan is not going to be an issue for them. The Chinese, I think, not quite this was not a coincidence. They were basically signaling over the summer that they would apply pressure on the Americans to say that they oppose Taiwanese independence, not just the Biden line of not supporting. I think this issue could come up again because whether rightly or wrongly, the Chinese believe that the Trump administration doesn't care that much about Taiwan. I think again, on balance, the Trump administration is far more preoccupied with trade, with technology, with energy than the issue of Taiwan itself.
Yeah, and on the issue of trade and also on the issue of the realignment of the global order of trade flows. I mean, I think, you know, the Europeans haven't been particularly impressed by Trump's treatment of Ukrainian President Zelenskyy. You know, the periodic berating of him saying that he's not grateful enough and various other things. Do you think this is going to push the Europeans closer to trading partners like China and further away from the United States?
Well, I would bet that Europe remains somewhat in the middle, but awkwardly so. Because the Europeans, I think, are very eagle-eyed about the fact that Trump signals a more protectionist America, an America that is more in favor of retrenchment from its commitments overseas. That is pretty obvious when it comes to the Russia-Ukraine conflict. But on the flip side, the Europeans, and I speak to them quite regularly being living in the UK, are very worried about the strategic and especially economic challenge that China poses. Europe generally has run trade surpluses with the rest of the world. Germany has been chief among them in this enterprise. And they're having their lunch eaten by the Chinese, where China is increasingly going up the value chain, competing with them and substituting for a lot of the goods that Europe has traditionally been quite good at. And at the same time, China has shown its willingness to deploy export controls for rare earths and critical minerals, where the Europeans again are in a position of vulnerability. I think they had more of a wake-up call over the last month or so when China did declare that this was going to be an extraterritorial set of export controls. They've since walked that back. But certainly, that creates a huge issue for the Europeans. Because on the one hand, they have an America that is isolationist and doesn't seem to care. And on the other hand, they have China that is competing with them on so many different levels.
You know, Alice, talk about eating other people's lunch. I mean, it all started several months ago with DeepSeek, right? And then more recently, there was another LLM which just dropped, which is catching on like wildfire, Moonshot's Kimi 2, which apparently costs what, just a little over 4 and a half million dollars to train, which is just a sliver, just a fraction of what names like OpenAI happen to spend, right? So, I want to get to the two competing models, the US AI model and also the Chinese model. And talking about lunch, the way I like to think about it, tell me whether it resonates with you, is what the US is trying to do is moving more closely towards say AGI. This is full fat. This is wagyu. It's expensive, it tastes good, but you have to wait, it involves a lot of work. What China is trying to do is more like, you know, lean turkey, a little bit cheaper, faster, and easier to raise. Doesn't really taste good, but it gets the job done, it feeds you. What do you think?
