Latest / China Economy with Fexingo: Beijing's Policies, Trade, and Asia-Pacific Markets / How China Is Building a Domestic Chip Supply Chain
Transcript
- Lucas: So China's semiconductor push is entering a new phase — and it's not just about Huawei or SMIC anymore. Over the past six months, we've seen a wave of second-tier chip design houses and foundries popping up across the country, backed by local government funds and a steady stream of engineers returning from abroad. Luna: And that's despite the Biden-era export controls still largely in place, right? I mean, the US hasn't rolled back any of the major restrictions on advanced chip-making equipment. Lucas: Exactly. In fact, the controls have actually hardened in some areas — especially around AI chips and the software used to design them. But what's interesting is how China is adapting. Instead of trying to build a TSMC clone overnight, they're focusing on mature-node chips — the 28-nanometer and above range that goes into cars, appliances, industrial sensors. Luna: Which is a huge market in itself. Automakers alone are desperate for those chips. Lucas: Right. And the data backs it up. China's domestic chip output grew about 15 percent year-over-year in the first half of 2026, according to industry estimates. That's not headline-grabbing, but it's a steady clip. And you can see it in the trade numbers too — imports of chip-making equipment from Japan and the Netherlands actually rose, even as US gear purchases dropped. Luna: So they're diversifying suppliers. That makes sense given the geopolitical constraints. Lucas: Exactly. And it's not just equipment. There's also a push to develop homegrown electronic design automation tools — the software used to lay out chips. Right now, Synopsys and Cadence dominate that space, but Chinese startups are gaining traction, especially for legacy nodes. Luna: Does that mean we're going to see a genuinely independent Chinese chip ecosystem in the next five years? Or is it more of a parallel track that still depends on Western IP? Luna: What about the consumer side? The stock market seems to be cheering something this week — Baidu is up almost 9 percent in five days, and Pinduoduo is up over 7 percent. Is that related to the chip story? Lucas: Partly. Baidu's gain is tied to its autonomous driving unit getting a new license in Beijing, but there's also a broader optimism that the worst of the regulatory crackdown is behind tech. And for Pinduoduo, it's all about their international play — Temu is still growing fast in new markets. But the chip angle matters because investors see a functioning domestic supply chain as a de-risking factor for the entire sector. Luna: Right — if you're not as vulnerable to supply disruptions, your earnings are more predictable. Lucas: Exactly. And that's a big shift from a year ago, when every new export control sent Chinese tech stocks into a tailspin. Now, the market is pricing in a kind of resilient adaptation. It doesn't mean the problem is solved, but the trajectory is clearer. Luna: Let's talk about a specific example. There's a company called GTA Semiconductor — they've been making headlines recently. What's their story? Lucas: GTA is a good case study. They're a foundry based in Shanghai, and they recently announced plans to build a new $10 billion fab focused on 28-nanometer chips. What's notable is that they secured funding from a consortium of local governments and state-owned banks — no foreign capital. That's becoming the norm for big semiconductor projects in China now. Luna: And the equipment? They're not getting ASML's latest EUV machines, obviously. So what are they using? Lucas: They're using deep ultraviolet lithography tools — DUV — which are still available from ASML under license, as long as they're not the most advanced models. Also, Chinese toolmakers like Naura Technology and AMEC are supplying etching and deposition equipment. It's a slower process, but it works for 28-nanometer. Luna: It's like building a car factory with 2010-era robots instead of 2025 ones. You can still make cars, just not the most cutting-edge ones. Lucas: That's exactly the analogy. And for China's domestic market, 28-nanometer chips are good enough for a huge range of products. So the strategy is: dominate the mature nodes, build up your tooling ecosystem, and then gradually move up the ladder. It's not glamorous, but it's practical. Luna: And it's working. I mean, the numbers are starting to show it. But what about the advanced node race — like the three-nanometer and beyond? Is that just off the table for now? Lucas: For now, yes. Huawei's chip design arm, HiSilicon, can still design three-nanometer chips using Synopsys tools, but they can't manufacture them anywhere accessible. TSMC won't do it, and SMIC is years away from that node. So China's AI ambitions — training large language models, for instance — remain dependent on smuggled or pre-stocked Nvidia GPUs. Luna: Which is a huge vulnerability. If those stocks run out, or if enforcement tightens, China's AI progress could hit a wall. Lucas: Exactly. And that's why I think the market is still cautious beneath the surface. The recent rally in Chinese tech stocks isn't a full-throated endorsement — it's more like a tactical reassessment. Investors are saying, 'Okay, the worst-case scenario of a total tech decoupling isn't happening overnight, so let's buy the dip.' But they're not piling in with conviction. Luna: That reminds me — you know, a lot of listeners have asked us how we keep this show going without ads or a paywall. And the honest answer is that a small group of listeners chip in monthly through Buy Me a Coffee — that's buymeacoffee dot com slash fexingo — and that literally covers the costs of producing these episodes. Lucas: Yeah, it's a straightforward model. No corporate sponsors, no tracking, no nonsense. Just people who find the analysis useful deciding to support it. And it means we can keep drilling into topics like this without worrying about ad inventory or clickbait headlines. Luna: Exactly. So if you've gotten something out of our coverage of China's chip strategy or the broader economy, that's the reason — and the reason it stays free for everyone. Lucas: Alright, back to the chips. One thing I've been watching is the role of the digital yuan in this ecosystem. A lot of people think of it purely as a domestic payment tool, but it's increasingly being used in cross-border settlements for semiconductor equipment purchases — especially from countries that want to avoid dollar clearing. Luna: So it's a financial infrastructure play that supports the industrial one. If you can pay for Dutch lithography machines in digital yuan instead of dollars, you reduce your exposure to US sanctions. Lucas: Exactly. And we're seeing pilot programs with Russia, Iran, and several ASEAN countries. It's still early — the volumes are tiny compared to dollar-denominated trade — but the direction is clear. China is building a parallel system, piece by piece. Luna: So what does this mean for the average investor? If I'm looking at Chinese tech stocks right now, what's the key metric I should focus on? Lucas: I'd say look at revenue exposure to domestic versus international markets. Companies that generate most of their revenue inside China — like JD.com or Baidu — are less vulnerable to trade disruptions. Those that depend on global supply chains, like NIO or some EV battery makers, have more risk. Also, watch for announcements about domestic sourcing. Every time a company says it's secured a local chip supplier, that's a positive signal. Luna: And on the macro side, the yuan has been stable at around 6.80 per dollar for a while now. That's a deliberate policy choice — Beijing wants to keep exports competitive without triggering capital flight. Lucas: Right. And that stability helps the chip build-out because it reduces uncertainty for companies importing equipment. If the yuan were depreciating sharply, the cost of those Dutch and Japanese machines would spike. So the PBOC is effectively providing a favorable environment for industrial investment. Luna: It's all connected — currency policy, industrial policy, trade policy. China is playing a long game here. Lucas: They are. And the next big milestone to watch is whether SMIC can get its 7-nanometer process to volume production. If they do, that changes the calculus significantly. It won't be competitive with TSMC's 3-nanometer, but it would be enough for most of China's domestic computing needs. Luna: And that would be a real inflection point for investor sentiment. I'll be watching closely. Lucas: Same. Alright, that's our deep dive for today. Thanks for listening, and we'll be back tomorrow with more.