
Nov 21, 2026 · 54 min
China builds an AI stack beyond NVIDIA
How AI Is Forcing China to Reshape Industrial Policy
China’s push for technological self-reliance could reshape semiconductor competition, but industrial strength has not reliably produced profits or household prosperity.
- 1China’s interconnected tech sectors let firms share supply chains, talent, and capabilities across AI, EVs, robotics, and semiconductors.
- 2Export controls accelerated domestic chip development while leaving Huawei constrained by software gaps, production capacity, and lithography bottlenecks.
- 3Beijing prioritizes strategic self-reliance and economic diffusion even when overcapacity, weak moats, and thin margins hurt investor returns.
Don't miss
Kyle Chan frames China’s effort to catch TSMC and NVIDIA as trying to catch a moving bullet train, with lithography as the central bottleneck.
The brief
China’s AI strategy is part of a broader industrial system linking semiconductors, EVs, robotics, smartphones, biotech, and consumer electronics through shared capabilities.
Kyle Chan of Brookings explains why Chinese firms build across the stack: supply-side resources, adjacent-industry spillovers, market incentives, and government policy signals.
Huawei sits at the center of the effort, but matching NVIDIA requires more than capable chips; software, production volume, networking, HBM, and lithography remain major constraints.
The core tension is speed versus sovereignty: Chinese AI companies want NVIDIA hardware for frontier performance, while Beijing wants an independent supply chain.
The episode’s sharpest economic question is whether technological progress counts as success when overcapacity and intense competition produce globally competitive products without durable profits.
What was said on this episode
34 statements · 22 positive · 8 negative · 3 mixed · 1 neutral
China’s technological development is trending toward vertical integration and self-reliance.
“Vertical integration and self-reliance.”
Listen at 2:36
Most physical products are likely to be built in China in the future.
“if there is something that gets built, if it is a built product, there's a good chance that in the future it will be built in China.”
Listen at 3:48
Chinese technology sectors may deliberately generate spillovers across adjacent industries.
“how do they actually have all these spillovers? Are those spillovers deliberate?”
Listen at 7:34
Xiaomi’s EV entry drew on China’s existing supply chains, expertise, and resources.
“He was drawing on a kind of supply chain foundation and a sort of pool of expertise and resources that already existed in China”
Listen at 9:46
Some Chinese technology companies control factories and manufacture components themselves.
“some of these other companies in China, they're controlling the factories, they're building even further down to the components.”
Listen at 12:23
Chinese firms may move from low-margin EVs into faster-growing markets such as robotics.
“there are reasons to want to move away from a low margin industry like EVs and try to tap into maybe the next big thing, maybe robotics or some other faster growing market.”
Listen at 13:55
China’s five-year plans identify chips, AI, and robotics as priority areas.
“There's a five-year plan that comes out and it literally lists out. what some of these priority areas are, whether they're chips, AI, robotics”
Listen at 14:42
Chinese firms entering priority sectors can unlock government resources and investment funds.
“if you go in this direction, that will unlock a whole bunch of resources for you”
Listen at 15:03
Huawei Ascend chips previously lacked sufficient reliability and functionality for Chinese AI companies.
“the main problem was getting Huawei's AI chips, the Ascend series, to be reliable enough and functional enough to be used by Chinese AI companies”
Listen at 19:57
DeepSeek and Huawei have a close partnership to optimize Huawei chips for DeepSeek models.
“you have a really tight partnership between DeepSeek and Huawei”
Listen at 21:02
Huawei Ascend chips support inference but remain inadequate for training massive models.
“the Huawei Ascend chips were good enough for inference, but were not quite good enough for training”
Listen at 21:42
Huawei can combine less powerful chips into larger supercomputing pods.
“you could bundle them together and you could start to build up these sort of super pods.”
Listen at 22:20
The next DeepSeek model may be fully pretrained on Huawei chips.
“there's rumors that the next DeepSeq model, for example, will be fully pre-trained on Huawei chips.”
Listen at 22:40
Huawei’s AI-chip production is currently constrained by limited capacity.
“at this point, Huawei is constrained.”
Listen at 23:06
Huawei’s production scale depends on equipment and supply-chain constraints.
“the big question is... how quickly can you get that production scale up over time?”
Listen at 23:45
U.S. export controls strongly accelerated China’s semiconductor self-reliance drive.
“US expert trolls... had a massively galvanizing effect on China's push for semiconductor self-reliance.”
Listen at 25:32
China is attempting to reproduce the entire global semiconductor supply chain domestically.
“China is almost trying to replicate the entire global semiconductor supply chain in China”
Listen at 26:06
Domestic semiconductor capability is China’s top five-year-plan priority.
“it's priority number one on the five-year plans.”
Listen at 27:49
A self-reliant semiconductor supply chain would let China advance AI despite U.S. policy.
“will basically allow China to continue to make AI progress regardless of what Washington decides.”
Listen at 29:49
Beijing’s long-term goal is supply-chain independence and sovereignty.
“The goal is to build up that kind of independence and sovereignty, essentially, in supply chains in the long run.”
Listen at 30:34
Beijing prioritizes high-tech self-reliance and leadership in future industries over GDP growth alone.
“the goal is high tech self-reliance of sort of like winning the industries of the future.”
Listen at 33:01
Winning high-tech industries and building complete supply chains will eventually produce growth and jobs.
“then everything else will kind of follow. You will end up having the kind of robust economic growth. You will have jobs.”
Listen at 33:50
China is making an all-in effort to become a dominant global player in key industries.
“there is this sort of all-in push to try to become, if not the dominant, a dominant global player in a number of these key industries.”
Listen at 34:18
Beijing intends AI to accelerate drug discovery, healthcare, and manufacturing.
“I think they want to integrate this into everything else. They want it to turbocharge drug discovery and healthcare innovation and manufacturing.”
Listen at 38:36
China continues to designate multiple technology sectors as critical future pillar industries.
“these different technology sectors are still laid out as like critical, you know, so-called future pillar industries”
Listen at 39:14
China’s chip industry faces a moving frontier that is difficult to catch.
“I kind of compare the whole chip industry to like trying to catch a bullet train.”
Listen at 41:15
China may produce sufficient chips for Chinese AI to maintain its current progress rate.
“will they be able to have sort of like good enough chips? that have decent enough supply where Chinese AI can sort of continue roughly at the pace that it's going at now”
Listen at 42:04
Lithography machines are the key bottleneck for China’s chip advancement.
“Ultimately, I think the key bottleneck is the lithography machines and whether China can get there.”
Listen at 43:20
Chinese companies generally struggle to maintain durable competitive advantages.
“it's hard to have a sustained competitive moat.”
Listen at 44:37
Low entry barriers create Chinese industrial overcapacity alongside globally competitive industries.
“they end up getting kind of flooded with, you know, overcapacity, redundancy, but then also building these overall very, very globally competitive industries”
Listen at 45:39
Unsustainable long-term profits make high valuations difficult to justify.
“If you don't have long-term profits or don't expect that to really be sustainable, it's hard to justify rising or high valuations.”
Listen at 45:56
Huawei may eventually match NVIDIA’s capability and achieve large-scale chip production.
“it seems that Huawei would get as good as NVIDIA and eventually you have the scale to produce as much as TMC on a long enough timeline.”
Listen at 47:34
Most economically productive activities can absorb only a limited amount of additional AI intelligence.
“most economically productive things in the economy can only absorb so much. intelligence”
Listen at 49:49
Frontier AI may have a small market while broader AI becomes commoditized and low-margin.
“The market is smaller than the rest of it gets really commodified and there's no money.”
Listen at 50:10
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.