Latest / China Economy with Fexingo: Beijing's Policies, Trade, and Asia-Pacific Markets / Why China's Memory Chip Push Is Running Into Headwinds
Transcript
- Lucas: Luna, there's a headline from yesterday that caught my eye: the memory shortage shaking Apple and Microsoft is being called an 'existential crisis' for smaller players. And I kept thinking — where does this leave China's multi billion dollar push to build its own domestic memory chip industry? Luna: Right, because on paper this shortage should be a tailwind for Chinese chipmakers. If demand is outstripping supply, local producers should be able to step in and capture market share. Lucas: That's the theory. But the reality is more complicated. China has two main memory chip champions — Yangtze Memory Technologies Corp, or YMTC, which makes NAND flash, and CXMT, which focuses on DRAM. Both have been under intense pressure from US export controls that restrict the advanced equipment they need to produce cutting-edge chips. Luna: And it's not just equipment. The talent gap, the software ecosystem, the yield rates — those are all bottlenecks that don't go away just because there's a shortage in the global market. Lucas: Exactly. I was looking at YMTC's trajectory. They made impressive progress a couple of years ago, getting to 232-layer NAND — that's competitive with what Samsung and Micron were doing. But then the US tightened export controls in 2022 and again in 2023, and they lost access to key tools from companies like Applied Materials and Lam Research. Luna: So they had to basically freeze their next-generation roadmap. Meanwhile, Samsung and Micron kept moving — Samsung's now shipping 300-layer-plus NAND, and YMTC is stuck. Lucas: Stuck is the word. And this memory shortage we're seeing now — it's partly driven by demand from AI data centers, which need high-bandwidth memory and high-capacity NAND. Those are exactly the premium products where YMTC and CXMT can't yet compete. They're more competitive in older generations, where margins are thinner. Luna: So the shortage ends up benefiting the leaders — Samsung, SK Hynix, Micron — more than the Chinese challengers. That feels like a pattern we see across the semiconductor space. Lucas: It is. And the Chinese government knows it. They've been pouring money into domestic alternatives — subsidies, state-backed funds, preferential loans. But building a memory chip ecosystem takes years, and the technology gap is real. I saw one estimate that YMTC's current generation is about two to three generations behind the leaders. Luna: Which in memory chips is a big gap. Because the market cycles — when demand is strong, you want to be on the latest node to capture the highest prices. By the time you catch up, the cycle may have turned. Lucas: Exactly. And speaking of cycles, let's look at the macro picture. The yuan has weakened to 6.77 per dollar as of last week. That's near multi-year lows. In theory, a weaker yuan helps Chinese exporters, including chipmakers, by making their products cheaper globally. Luna: But memory chips are priced in dollars globally, so if your costs are in yuan and your revenue is in dollars, depreciation actually widens your margins — if you can sell. The issue is if you can't produce competitively, the currency tailwind doesn't matter. Lucas: Right. And the trade-weighted US dollar index is at 120.4 — that's very strong. So the dollar is strong broadly, not just against the yuan. That makes dollar-priced imports more expensive for everyone, including Chinese companies that need to buy foreign equipment. And since export controls already limit access, the strong dollar adds another layer of cost pressure. Luna: So the Chinese memory chip strategy is basically caught between geopolitics and market realities. Washington isn't going to ease export controls anytime soon, and the technology gap isn't closing fast enough. Lucas: And there's another angle: the global memory market is notoriously cyclical. When demand softens — and it will at some point — the leaders cut prices to maintain market share, which crushes margins for latecomers. YMTC and CXMT have to survive those down cycles without the scale or technology edge to compete. Luna: You know, it reminds me of what happened with China's display panel industry a decade ago. BOE and others eventually caught up in LCDs after years of investment, but memory chips are much harder because the technology changes faster and the capital intensity is higher. Lucas: That's a good parallel. In displays, the technology curve was more mature and the supply chain less restricted. For memory, you're dealing with sub-10-nanometer lithography, extreme ultraviolet tools, and a tightly controlled equipment ecosystem. It's a different league. Luna: And you have the US, Japan, and Netherlands all coordinating export controls. It's not just the US — ASML from the Netherlands and Tokyo Electron from Japan are key. So Beijing can't easily source alternatives. Lucas: Which is why we're seeing China pour money into developing domestic equipment. But that's a multi-year, maybe multi-decade effort. In the meantime, the memory shortage benefits the incumbents. And it's worth noting that Chinese tech companies like Huawei and Xiaomi are among the biggest buyers of memory chips — so they're feeling the pinch too. Luna: Right, Huawei's server business and smartphone division need a lot of high-end memory. If they can't get enough from Samsung or Micron, and Chinese suppliers can't deliver, that's a bottleneck for their own products. Lucas: Exactly. So the shortage that's helping Apple and Microsoft — they have the pricing power to secure supply — is a real problem for Chinese OEMs. And that's where the 'existential crisis' headline really lands. Luna: It's a good reminder that trade wars and tech decoupling have real consequences. Sometimes the policy goal and the market outcome don't align. Lucas: Speaking of which — and I don't say this often — but if you've been following this show and finding these conversations useful, there's a small way to keep it going. If today's episode was worth a coffee to you, you can find us at buy me a coffee dot com slash fexingo. It's a simple way to support ad-free, independent economics coverage. Luna: Yeah, it genuinely makes a difference. And we don't run ads, so listener support keeps this whole thing running. Lucas: Right. So back to memory chips. Another data point: Chinese chip stocks have been underperforming lately. The KWEB — that's the China internet ETF — is down about 4.4 percent over the past five days. And Alibaba, JD, Baidu all lower. That broad weakness reflects a lot of concerns, but the semiconductor sub-sector is particularly sensitive to these export control headlines. Luna: And the broader economic backdrop isn't helping. China's trade surplus has been shrinking — the latest goods and services trade balance was a deficit of about 56 billion dollars in April, actually slightly better than the prior month but still a deficit. And the current account deficit widened to nearly 227 billion in Q1. Lucas: Those numbers show the economy is still importing a lot — including a lot of high-tech components. So the push for self-sufficiency in memory chips isn't just about prestige; it's about reducing a structural import dependence. Luna: But here's the question: can they get there before the window closes? The memory market is consolidating — Samsung, SK Hynix, Micron are investing huge sums. They're not standing still. Lucas: They're not. And I think the realistic answer is that China will make progress in legacy nodes — DDR4-style DRAM, older NAND — but the frontier will remain out of reach for a while. That's not a failure, but it means the strategy has to be more about resilience than dominance. Luna: And resilience is what matters in the long run. If they can at least supply the domestic market for older chips, that reduces vulnerability to supply disruptions. Lucas: Exactly. So the memory shortage is a stress test for China's chip ambitions. It's revealing the gaps, but also the determination. The next few years will tell us whether that determination can overcome the technology barriers. Luna: And whether the global trade environment allows it. I think that's the big unknown. Lucas: It is. And we'll keep watching it episode by episode. For now, that's the picture on China's memory chip push — ambitious, under pressure, and far from settled.