It really couldn't be a bigger story. Already Jensen Huang, the CEO of Nvidia, has said that his company has largely conceded — those are his words — largely conceded China's AI chip market to Huawei. In 2026, so far, Huawei has captured around 50% of the Chinese market for AI chips. What this really means is that chip companies in the future might stop obsessing so much as they do right now about how small the transistors in a semiconductor should be, and they might focus more instead on how fast the data moves through a semiconductor.
Welcome to China Decode. I'm Alice Han, and I'm James Kynge. In today's episode of China Decode, we're discussing the latest sign that China's chip industry might overtake Nvidia, a growing trade war with the European Union, how Hong Kong has become more popular than Switzerland for the mega rich. That's all coming up. But first, let's do a quick check-in with how the Chinese markets are starting the week. On Monday, markets were down today as investors turned cautious over renewed China-EU trade tensions. The Shanghai Composite Index dropped 0.73% while the Shenzhen Composite Index fell 1.81%. Overall in May, the Shanghai Composite lost 1.06% while the Shenzhen component was up 3.1%.
All right, let's get right into it. There are more signs coming to light of the Chinese chip industry's ability to potentially take down US competitors in the not-so-distant future. Huawei, thanks in large part to its head of chip development, He Cangbo, has announced an antidote to Moore's law. That's the long-in-practice mechanism by which semiconductors continue to improve and get smaller over time. But increasingly, it appears that we've hit the Moore's law limit. Instead, Huawei is beginning to operate under a different kind of scaling law, and the innovative approach might just catapult Huawei into serious competition with US chip makers, the leading of which is Nvidia. I love the fact that she's now called the chip queen and she's somebody who's generally shied away from the spotlight, but she's been responsible for running this internal semiconductor unit since 2003. So, nearly two decades with an annual budget of about $400 million, and she's now become a viral sensation in China, and I think rightly so. Can you tell us a little bit about what you think about this new towel scaling law that Huawei's just launched and whether or not this is a serious threat? Everyone's going to ask this multi-trillion dollar question: Is this a serious threat for Jensen Huang?
I think, first of all, as you've described it, this story is partly about He Cangbo, the chip queen as you put it, and it's also partly about the business of the world's biggest company, that's Nvidia, in China, the world's biggest emerging market for semiconductor chips. So, it really couldn't be a bigger story. And I think it's really key because already Jensen Huang, the CEO of Nvidia, has said that his company has largely conceded — those are his words — largely conceded China's AI chip market to Huawei. And he may well say that because in 2026 so far Huawei has captured around 50% of the Chinese market for AI chips. And just to give you a number here, the size of the AI chip market in China is supposed to be about 21 billion US this year, but by 2030, according to Morgan Stanley, it could be worth about 67 billion US. So we're talking really big dimensions as usual. And now what He Cangbo has said is that we're moving to a new type of way of assessing semiconductor performance, and she's called it the towel or the scaling law. What this really means is that chip companies in the future might stop obsessing so much as they do right now about how small the transistors in a semiconductor should be, and they might focus more instead on how fast the data moves through a semiconductor. So that puts the focus away from the little transistors that go into a semiconductor and more into the clever way that you can arrange semiconductors on a chip and the wires that you use and the memory systems that you use so that you get greater efficiency on that chip. At the moment, as I'm sure you're aware, these transistors that go into CPUs or GPUs — those are the chips within AI systems, right? — they're so small that you can literally get hundreds of billions of them onto a chip. Now, that means that these transistors are invisible to the human eye. So we're reaching the physical limits of miniaturization. And that's why, as you said in the opening, Moore's law, which used to say that a transistor that would go into a chip would endlessly shrink, so you could get more and more of these transistors onto a chip and that of course boosts the compute power of that chip. That's why this has always been the aim. But we're now reaching the physical limits of Moore's law. We simply cannot expect physics to allow chips to get smaller and smaller and smaller now that they've already reached the size effectively of a human virus. So that's the key background to what He Cangbo is saying here. She's saying we're moving to a new way of assessing the compute power of a chip. And that way is not to miniaturize the chip because we're already reaching the limits of physics. It is to cluster the chips, to stack the chips in a clever way and focus on the other parts of the chip such as the wires and the arrangement, design, all of those things to increase the compute power of the chip. So, I do think it's an interesting development. I have no way of assessing whether or not this is going to be something that really has traction. I have no way of assessing whether the towel or this towel scaling law is really going to be the way that we assess chips in the future. What do you think, Alice?
Yeah, it's so complicated and insofar as I understand it, you know, Moore's law is predicated on a doubling of the number of transistors on a chip and to your point, it's ever getting miniaturized, smaller and smaller in terms of the components on the chip. Now apparently TSMC is understanding that it's hitting the upper limit of Moore's law and is now going into 3D stacking which is not too dissimilar I think from this towel philosophy in terms of increasing efficiency on a chip. But if TSMC is going in that direction because it recognizes it's hit the upper limit, my question is, you know, can Huawei and other players, you know, Hua Hong and SMIC, who are pure-play foundries as well in the space, can they reach the aggregate amount needed to meet Chinese demand? I thought it was interesting and again here I'm citing the Council on Foreign Relations so you've got to take it with a grain of salt, but insofar as they understand it through their own research, you know, the best US AI chips are still five times more powerful in compute than what Huawei can currently do and apparently according to them that gap is widening. And according to their calculations, only around, you know, in the most bullish scenario, Huawei can only produce about 4% of the aggregate AI compute that Nvidia can produce. I think that is a huge takeaway because even if they are getting more efficient in the short term, the big question, trillion, multi-trillion dollar question, is can Huawei produce enough to meet domestic Chinese demand which is only going to increase as we've previously talked about, you know, because of the massive public and private sector investment in data centers and AI across the stack. I'm a little bit worried there, but I don't bet against the Chinese in terms of finding ingenious ways to work around existing technological constraints, but in the short term, I'm a little bit worried about whether or not they can meet the demand currently. One last data point I thought was super interesting. I'm still bullish, maybe this is out of consensus, Jensen Huang in China. Jensen Huang just got announced that he is on the board to Tsinghua University's School of Economics and Management, on which the Apple CEO Tim Cook is also a member. Maybe this is a signal that things are heading in a more positive direction, that some of the floodgates may be opened softly. But my number one question is can China meet the demand that is only rising for AI compute with the existing domestic competition?
Well, that's the $64 million question or probably $64 billion question. I have no way of knowing but my sense of this is that the running is going in China's favor. And I think that this statement by He Cangbo, Huawei's chip queen, gives a sense of confidence. It gives a sense that China's got an alternative method, an alternative route towards creating the types of compute power that it needs. And in this regard, this is this scaling technique that she is describing. So I think that actually although I believe China would probably like to get its hands on the latest Nvidia chips, the Blackwell chips which are considerably more powerful chip for chip than anything that China's got or anything that Huawei is even close to producing, I think that even if China doesn't get its hands on these Blackwell, it can stack these chips and scale these chips, the Huawei indigenous chips, in such a way that it can probably have a useful workaround. And then if you add into the equation things that we've spoken about on China Decode before, the fact that electricity is so much cheaper in certain parts of China and that, you know, you can run data centers close to the Gobi Desert using solar power that provide incredibly cheap electricity and thereby overcome the fact that you're having to use more power to power the chips to run your large language models. I think overall China is in a pretty good spot. But I'm also interested — you mentioned He Cangbo at the beginning. This is a great human story as well, isn't it, Alice? I mean, as I understand it, she was put in charge of Huawei's chip development way back in 2003 when Huawei was really nowhere when it came to making chips. They were really nowhere. But she had a $400 million US budget annually and she's become one of the company's most important executives and Huawei has to rank as China's most important tech company if not China's most important company, full stop. And now He Cangbo is on the Huawei board. She's one of only two women to be on the board. It's really an incredible story, isn't it, Alice?
It really is. And I've been reflecting a lot on Huawei's evolution, that Huawei was late to the game to what Samsung had long been doing because Huawei obviously started off doing telecommunications and then has expanded now across, you know, phone sets to now chip production and Samsung did this fairly early on but I think increasingly, you know, my take on this and maybe you think this is right or wrong, is that Huawei is increasingly becoming the Samsung of China which is this global technological leader in the broader ecosystem, but that has a complete supply chain or hopefully resilience in that it's touching many, many different sectors in the technology stack. What's your take on that?
Yes, I would definitely agree with that and I would say the other thing about Huawei is that it's the symbol of China's resistance to American pressure under the export control regime. I mean Huawei has been the main target of American sanctions. It has been unable to get hold of a lot of crucial American technology and yet it seems to have prevailed and so the development of various different semiconductor chips, He Cangbo's confidence in coming up with this new law, the scaling law, I think is just another indication that Huawei is prevailing against, you know, American sanctions and American pressure. So, it's a geopolitical story as well.
And I have to say that I'm happy about my prediction that semiconductor stocks could rally. I was just looking at year-to-date performance on one of the ETFs and it's up at least 40% year to date. It's outpaced the AI stocks and the Shanghai and Hong Kong composite indices. Clearly a sign both of confidence, growing confidence in the domestic chip industry, but also, you know, foreign investors I think are looking into it as a play, as a more pick-and-shovel play for AI as Nvidia has been in the US context. It's interesting that we'll have CXMT which is a big memory DRAM supplier coming online probably this year and then I just heard the announcement that MiniMax which is a big AI model company is looking to list additionally in addition to Hong Kong also in mainland China on the Shanghai STAR board rather. So we're seeing a lot of momentum, I think, in this AI hardware story and the chips definitely are a beneficiary of that. Okay, we'll be back with more after a quick break. Stay with us.
Support for the show comes from LinkedIn Ads. There's no worse feeling than making a major investment in something only to realize it didn't exactly live up to the hype, such as buying a nice piece of tech that ends up in storage collecting dust, or taking a business workshop where your main takeaway was little more than a few motivational words. If you work in marketing, this can happen with ads. You optimize for the numbers that look great — impressions, reach, and reactions. But when they don't show revenue, well, that can turn into an unfun conversation with the CFO. LinkedIn has a word for that: bull spend. Reach the right buyers with LinkedIn Ads and invest in what looks good to your CFO. According to the 2026 Dream Data Benchmark report, LinkedIn Ads generated the highest ROAS of all major ad networks. It's 121%. You can target by company, industry, job title, and more. It's time to cut the bull spend. Advertise on LinkedIn, the network that works for you. Spend $250 in your first campaign on LinkedIn Ads and get a $250 credit for the next one. Just go to linkedin.com/scott. That's linkedin.com/scott. Terms and conditions apply.
Welcome back. In another evolution, following the increased sanctions between the US and China, there are now apparently growing worries in the European Union as to their economic dependence on China, which means a new trade war is just about to brew. James, as the resident European, I have to get your hot take on this. This has been a long time coming because China for years has been running increasing trade surpluses with Europe. It just apparently hit over $400 billion last year, the highest level in history and probably will increase again in 2026. But I mean the Europeans should have foreseen this, that China would increasingly become Germany in the global supply chain and that it's replacing it in the areas in which German and European manufacturing have traditionally been strong, that is cars and chemicals in particular. I found it interesting that the majority of the trade surplus is in cars and machinery right now but China's making a lot of inroads in the chemicals industry as well. So it wouldn't be surprising to me that if we have this conversation in 5, 10 years that it is the global dominating supplier in the chemicals industry. But as a European I have to ask you, you know, where did we go wrong? Where did the Europeans miss the train?
Yeah, those are great questions, Alice. I would say some of the biggest things to change recently was a statement by French President Emmanuel Macron. He called on the EU to take inspiration from the United States on trade measures aimed at protecting strategic industries in Europe. And then he had this quote and this really sums up the realization that is dawning not only on Emmanuel Macron of course but on many policymakers in Europe. He said, quote, "China is literally killing a large part of European industry." And I must say he couldn't be more correct. We've all heard about floods of Chinese electric vehicles coming into Europe and really dealing body blows to European car makers, but that is by no means the only sector in Europe that is getting hit. And you mentioned the issue of whether or not Europe should have woken up to this a long time ago. And you know, it's so clear that they should. I have been so exasperated looking at Brussels, you know, the center of EU power and looking at the national governments that are members of the European Union and looking at the government here in the UK and wondering why is it that they are failing to heed so many warnings and so many research papers written by think tanks all over Europe talking about the rise of China's tech challenge and the way that this is going to impact one industry after another across Europe. And you know, until people like Macron a couple of months ago started to talk in the way that I've just mentioned, we really didn't get hardly anyone sounding the alarm. But now we have, and I guess that's why this is topical at the moment. We've got a French-led coalition of five EU countries, that is France, Italy, Spain, the Netherlands, and Lithuania. And they are urging Brussels to rapidly deploy stronger cross-sector tariffs and defensive trade tools to protect European industries from this surge of Chinese state-subsidized imports. And the trade deficit that Europe has with China is enormous. It looks like there are forces in Europe, let's put it that way, that are limbering up for a trade war. They want to have a trade war with China. They want to restrict Chinese imports with tariffs and other defensive trade tools. There is going to be a G7 summit. That's the summit of leaders of the G7 countries. And that will be held in France on June the 15th. And I think that's partly why we're seeing Emmanuel Macron, the French president, start talking about this because I believe that this G7 summit may talk a lot about reducing economic reliance on China and talking about the hit that many industries in countries across the G7 are getting from cheap Chinese imports. And we have officials from the European Commission's Directorate General for Trade saying that they are drawing up, quote, "more assertive and effective trade defense policy tools" with Beijing. So, we've got quite a lot of things going on. Personally I think that what we may see is some new trade tools being used by Europe to protect their domestic industries against Chinese competition. But I very much doubt whether these will work. And the reason I say that is primarily because many of Germany's biggest companies have China as their biggest market in the world. And those big German companies are lobbying the German government to ensure that a trade war or trade friction between the EU and China doesn't get too nasty because they fear that the Chinese government could turn against their interests in China. So this is a convoluted web. It's hard to know how it's going to evolve, but what are you seeing on this topic, Alice?
Okay, I better be a bit of a Debbie Downer, James. I go to so many conferences, I was just in Poland where the standard fare is Europeans complaining about the G2, US-China dominating the world and they're stuck in the middle. And meanwhile, you know, very little of the Draghi report has been adhered to. There's a lot of talk about export restrictions and tariffs on China. And yes, you know, over a year ago, they did put restrictions on EVs with tariffs, tariff measures, but we still continue to see record Chinese exports of EVs into Europe. I thought it was interesting, the Rhodium report just had recent data suggesting that while there was a downward cycle in Chinese FDI, both M&A and greenfield since 2018, over the last two years, there's been an uptick. So last year, FDI grew 57% year-on-year. And a lot of that is actually driven by both M&A activity, so buying up European companies, and greenfield investment into places like Hungary traditionally, but also increasingly France and Germany in the automotive and green tech sector. So automotives and batteries were the highest share of that greenfield FDI investment breakdown. So my big cynical take and question of sorts is at the end of the day, do these barriers matter if China is just going to put up shop and factory in Europe? Do the kind of reverse tech transfers that part of Brussels actually wants and bring jobs, manufacturing jobs to Europe in the way that Japan did in the 80s? Remember when Reagan put on voluntary export restrictions, basically tariffs on Japanese goods, in particular automotives, and then Toyota and other conglomerates responded by setting up factories in middle America. Maybe that is the future and all this tariff stuff is just a nothingburger, but maybe I'm overly cynical.
So, you're saying, Alice, that it won't save European industry anyway because Chinese investors will move into production sites within the European Union like Hungary and whatever.
Correct. And those guys are going to go out of business anyway. I mean, the Germans and the French.
