Yves here. Readers who have taken note of the performance of DeepSeek AI or other Chinese technical feats, like their super cheap EVs, dark factories, number of patents issued would not be surprised to learn that the US is lagging China. This Tom Neuburger post provides some details.
By Thomas Neuburger. Originally published at God’s Spies
It’s not difficult today to find articles on China’s great leaps forward in high tech development, and indeed, in development of all kinds. Ian Welsh, for example, has been ringing this bell for a while. A good example is here:
Forget GDP, it’s completely misleading. China is ahead in everything that matters: 80%+ of tech fields, has more population and the largest industrial base in the world and it’s the main trade partner of more nations than anyone else, including America.
This graphic is illustrative, but it applies to everything except planes and launch capacity, and soon it will apply to them too[.]
The graphic referred to is sobering:
For more of Welsh’s China reflections, go here.
China and Tech Innovation
But one story in particular struck me as beyond remarkable:
World’s fastest memory writes 25 billion bits per sec, 10,000× faster than current tech
PoX could be the key to unlocking performance bottlenecks caused by storage limitations in AI hardware
A research team at Fudan University has built the fastest semiconductor storage device ever reported, a non‑volatile flash memory dubbed “PoX” that programs a single bit in 400 picoseconds (0.0000000004 s) — roughly 25 billion operations per second. The result, published today in Nature, pushes non‑volatile memory to a speed domain previously reserved for the quickest volatile memories and sets a benchmark for data‑hungry AI hardware.
I’ll spare you the jargon. Non-volatile flash memory is memory that persists when the power is turned off. It’s the BIOS at the core of your laptop, the flash drive the size of your thumb. As you know, writing to flash drives is slow. What the Chinese have done, is, with new technology, increased the write speed by a factor of 10,000.
Notice the use of “graphene,” by the way, in the design of this product. Graphene is a miraculous and under-used substance — under-used because its use involves great technical difficulties. But the promise is great, as the above report shows. China is not alone in graphene development, but they’re at the head of the pack.
This is an artist’s representation, but accurate. It’s truly two-dimensional:
Yet it’s considered the strongest substance in nature at a molecular scale, with incredible electrical properties. There’s more on its promise here.
China, Pride and the West
My point is not about tech, not primarily. It’s about China. While the U.S. was busy making its rich people richer (when you own the government, you can make it dance to your tune), China was making China great again.
I mean that literally. The history of China is a story of national greatness: the earliest dynasty, the Xia, dates to 2000 BC; the first Empire, the Qin (“chin”), dates to the Roman conquest of Greece. By the end of the 18th century, the Qing (“ching”) dynasty ruled a third of the world’s population; and its country contained earth’s largest economy.
That greatness was followed by a fall: subjugation to the will of the West, primarily Great Britain at first, as discussed in works like Alfred McCoy’s To Govern the Globe, the story of the West’s brutal treatment of, well, everyone else in the world. (Our ongoing discussion of McCoy can be found here.) From the mid-1800s on, China became a client state of the West.
The Chinese are keenly aware of this great embarrassment, this blot on their historical record, and many have argued that this, more than wealth, drives Chinese policy today. You can see Chinese pride, for example, in the current government’s approach to corporate corruption, a striking contrast to our own, where corruption has almost become a fourth branch of the state. (Welsh comments on that briefly here.)
All this is to say that China is technically leaps ahead of the rest, as you may have read in stories about DeepSeek AI or the stunning BYD all-electric cars, whose selling price in China starts around $8,000. In contrast, the cheapest vehicles in the U.S., all gasoline powered, start around $20,000.
China’s Post-Silicon World
A final story, this from a former Intel executive, now CEO of a Chinese ATE company. (ATE means “automatic test equipment,” the chip-test industry.) Here’s Part 3 of his discussion of memory chips (reformatted slightly for clarity):
PART 3: Moore’s Law is dead. China’s not mourning—it’s moving on.
While the West clings to silicon and EUV [extreme ultraviolet lithography, a chip-making technique], Beijing is building a post-silicon future—with light, atoms, and brains.
If you thought atomic memory was scary… wait till you see what’s next.
Thread:
Most people think chip wars are about EUV, 3nm [3 nanometers, the size of a single transistor on a chip], and NVIDIA. That’s yesterday’s game. China’s playing on a different board now: Post-Silicon Computing.Photonics
Neuromorphic chips
Spintronics
2D [two-dimensional] materials
[2/15]Why ditch silicon? Because physics won’t play ball anymore.
We hit atomic limits
Heat throttles performance
Quantum effects break transistors
Silicon is choking. Moore’s Law is over.
[3/15]China saw this early. Instead of racing TSMC [Taiwan Semiconductor Manufacturing Company Limited] to 1nm, it’s building alt-tech ecosystems:
Memory from 2D crystals [graphene]
Processors that use light not electrons
AI chips that mimic the human brain
All skirting U.S. sanctions.
[4/15]Example: Photonic Chips
Light-based logic, no heat, near-infinite bandwidth. Tsinghua + Huawei built photonic tensor cores for AI—10x the speed at 1% power. No EUV. No Intel. No problem.
[5/15]Example: Neuromorphic Computing
Think chips that think like brains. China’s Tianjic and Darwin chips use spiking neural networks—ideal for edge AI and robotics. Forget GPU farms—this is on-device intelligence.
[6/15]Example: Spintronics
Using electron spin instead of charge. Fudan and CAS are prototyping memory and logic devices with zero leakage, zero volatility, and atomic scale density.
[7/15]And it all connects to the memory bomb we dropped in Part 1 & 2:
2D memory = foundation
Photonics/Neuromorphics = compute layer
All built without TSMC, ASML, or ArmChina’s breaking the U.S. tech siege by jumping over it.
[8/15]Why this matters: The West has bet everything on scaling silicon and squeezing EUV. But China is building a parallel ecosystem, born in national labs, designed for strategic decoupling.
[9/15]Western analysts dismiss it as “not ready.” That’s what they said about:
Chinese EVs
5G
Solar
DronesNow those industries are owned by China.
[10/15]And while the U.S. throws $52B at TSMC Arizona delays, China is:
Publishing thousands of post-silicon papers
Funding 100+ new labs under “863” and “Key R&D” programs
Launching “More than Moore” pilot fabsThis is national strategy, not startup vaporware.
[11/15]Even Huawei—banned, sanctioned, bleeding—just released AI chips rumored to contain photonic interconnects and 2D SRAM. They didn’t die. They evolved.
[12/15]Remember this: The U.S. controls the past of computing. Silicon Valley is named after a dying substrate [silicon]. But China may control what comes after—and it won’t be built with silicon.
[13/15]You don’t need to beat NVIDIA. You just need to make NVIDIA irrelevant. China’s post-silicon roadmap is a bet against the status quo—and it’s accelerating.
[14/15]Next up (Part 4):
If China controls post-silicon computing, who controls the new standards?
Quantum encryption, chip-to-chip photonics, edge AI models—
The next war isn’t on the chip. It’s on the protocol. #MemoryWars continues.
[15/15]
I really don’t want the focus to be on Trump. He’s not the fire; he’s just an accelerant. The fire is the greed of the rich, their capture of government, and their constant shortsightedness when faced with the gleam of more wealth.
How crass, how tragic, how low, how mean their desires. That we went along, that we fed them, how sad for ourselves.
Wow. An amazing amount of info about what is coming shortly.
Yves, you are getting me depressed. I can’t imagine China is the only one with this degree of forward thinking.
Other people are, but other research had to be done first.
https://graphene-flagship.eu/materials/health-and-safety/
I wonder to what degree the West’s reaction has been because the Chinese are not white. Most normal people are happy when their neighbours are successful, bring soup when their neighbours are down.
Especially when they’ve been kind to you, have been willing to share that success, and when it could save the world.
The US business community and press was not happy when Germany outstripped the US in manufacturing in the 1970s, above all in cars. But US has 38,000 active duty soldiers there, so they are also subject to US influence, unlike China.
Thank you, Yves. Plus Japan.
“All this is to say that China is technically leaps ahead of the rest, as you may have read in stories about DeepSeek AI or the stunning BYD all-electric cars, whose selling price in China starts around $8,000. In contrast, the cheapest vehicles in the U.S., all gasoline powered, start around $20,000.”
That’s MSRP, but in practice good luck finding a vehicle for under 25K given how many fees dealerships are now getting away with adding since the quarantine.
*******
Overall this is a great article to write and circulate. We’ve had so many comments in the past two decades about Chinese cheap labour, cheap Chinese crap, Chinese only know how to copy, Asians have a lower creativity IQ, etc. It was all nonsense racist cope.
Pretty sure my kid’s not gonna be learning Mandarin in small town USA, unfortunately
Thank you.
Mauritian government schools abandoned the teaching of German and even Spanish a dozen years ago and offer Mandarin, Hindi and Russian instead. Many, if not most, of my young relatives, say under 25, speak French, Creole, English, Hindi and Mandarin.
Some 25-30 years ago, my Australian colleagues told me that there it was already pretty common to study Asian languages (Mandarin, Indonesian/Malay seemed to be the most popular) instead of, say, German.
I tell my US high school students that they should be learning Mandarin. Opps, my school district, one of TX’s largest, doesn’t offer any Asian languages.
However, when I taught in Thailand, Thai students studied English and Mandarin in addition to their native tongue.
We’re #1! (in falling behind)
Untold billions have been deployed into making silicon the substrate for western frontier compute stacks, and trillions of dollars of market cap on the Nasdaq rest on the thesis that accelerating towards 1nm is the holy grail. Even an acknowledgement that a post silicon future is possible (and the protagonist at the vanguard is China) would collapse forward multiples so the warnings in this article will be dismissed with derisive laughter by the business press and analysts long Nvidia. They’ll continue breathlessly asserting that “the US is leading the AI race” while deepseek quietly makes algorithmic breakthroughs and disintegrates the economic value of the AI market through open source.
And while nuclear in the US continues to run into nimbyism and regulatory hurdles, China is taking the wraps off a Thorium fuelled small modular reactor (still in the testing phase) and recent satellite photographs have confirmed that the Chinese have the largest nuclear fusion test facility in the world (while billions in western venture capital is pouring into fission startups). Expect western hubris to be wholly dismissive of all this.
Thank you, Thuto.
You forgot to add Uncle Sam’s European vassals, even Blighty, whose comprador jefe says it’s “the AI superpower”. The comprador finance minister says “defence industry superpower”.
As per my comments yesterday and recently, some French and others* no longer believe, are jumping ship and heading our way. Even their sex workers, the canaries in the coal mine, are coming, too. The fairest Cape of all, Cairo, Dakar and Mauritius appear to be the top destinations.
*Engels and Volkers has opened a few miles down the road / west coast. This is also the area where the sex workers base themselves on tour.
Thank you, Colonel.
Yes “the fairest Cape of all” is awash with decamping European glitterati who are snapping up prime real estate on the Atlantic Seaboard. As regards Blighty comprador’s delusions of grandeur about it being an AI and defense tech superpower, it’s comical to watch an empire turned into a third rate power thinking it’s still in its heyday.
CEO’s in China have zero political influence…none…in fact, if they do need something, anything, they have to get on their knees….of course in the US, they and their acolytes run the system…
They’re CEOs.
They’re not suffering.
Although competing would feel like that to all of ours.
Ever since George Bush was President, the Chinese could see the writing on the wall. That America was going to be gunning for them as they would not comply with their demands so they had to be prepared. Things really ramped up with Obama and his ‘Pivot to Asia’ and his ‘Everybody but China’ trade deal aka the TPP. When Trump came in he targeted China with tariffs but in a klutzy way. Biden did not have time for China as he was wrapped around the axle of the Ukraine but with Trump back in with his neocon backers, they are now going after China with a trade war. Point is, China saw all this coming and took steps to secure their economy like Russia did a decade ago. And this included scientific research as China was vulnerable to being cut off at the knees here once. No longer. And if it is going to be China writing the standards of modern computing, well, that is one of the rewards of heavily investing in research and not gutting their scientific establishments. Maybe the US could do a Manhattan Project for computing so that they do not get left behind but I do not think that it will ever happen. Besides, most efforts these days seem to be devoted to AI and cryptocoin because they figured that was where the fast bucks were. And it now appear that we are going to be left in the dust.
“Maybe the US could do a Manhattan Project for computing so that they do not get left behind but I do not think that it will ever happen.”
Response by US Congressional and Presidential policy experts: “Manhattan Project? What’s that? Latest AI or a stock buyback?….I’ll call my broker to see if I should get some of that before news gets out we’re funding it.”
Thank you, Yves.
In the late 1990s, I worked for NatWest Bank and noted how engineers jumped ship from clients Northern Telecom, Sony and Ericsson for the Chinese firms that we later became familiar. Some of the engineers were also clients of the bank and freely, even cheerfully, admitted that their hiring managers told them to build for the future, build the best and not to worry about financial control, a most welcome message after years of having to put profit and shareholder returns ahead of investment and innovation.
A dozen years ago, I recalled these conversations when told by a British automotive engineer based in Germany that in a decade or so China would eat Germany’s lunch as Chinese car manufacturers were investing and innovating, but German firms preferred to bribe politicians for protection and focus on financial engineering. We talked mainly about cars, but domestic appliances were also highlighted. This discussion took place in Brussels when the City, French agriculture and German engineering, Europe’s big three, met to take care of business, i.e. we would ask the European politicians and officials on our payroll to ease up on them if theirs were asked to ease up on us.
A sure way of killing creativity, innovation and surprising new discoveries is to do what is being done now:
-First dumb it down sufficiently so that even the PMC can pretend to understand and measure it.
-Then control all through Key Performance Indicators (KPI) that are so simple to understand that even PMC can understand them, sadly this often means that the KPI are also easily gamed
-And finally act all surprised when, although KPI showed outstanding progress, the actual results are bad.
One example of unintended effect of poorly defined KPI:
https://pmc.ncbi.nlm.nih.gov/articles/PMC5178044/
I’d say part of the reason why innovation fails is that many managers have difficulty managing people who know more than them. Courses, books and seminars etc etc are needed to teach managers how to manage experts. One example:
https://www.mindtools.com/aebk530/managing-a-team-of-experts
The comic Dilbert made fun of it, the outcome of having managers without knowledge combined with a deep fear of admitting they don’t know everything can lead to bad outcomes
Giving freedom to be creative and innovative does not guarantee success and might lead to costly and shameful failures and therefore it takes courage to give employees freedom to be creative and innovative. Few organisations have the courage to allow even a few employees the freedom to be creative and innovative. Universities used to be more courageous, the trend, as far as I can tell, is to control more and more.
The first Cold War had a Sputnik Crisis that supercharged American tech but that–even the Moon landing–was about fear of Soviet atomic missiles raining down from the sky.
Our new re-enactor version seeks to revive the former FUD but lacks the context of a world that had recently survived WW2. Perhaps all the incoherent noise coming out of the White House these days is merely a Boomer thing. We vividly remember the urgency of the Atomic Cafe. Life via smartphone is so much smaller scale.
Our elites need therapy or maybe should simply pass the baton to a younger generation.
Check out the latest in sodium batteries from CATL. Just amazing
Its a 7 minute video, yes from CATL but its amazing what they are producing, today now.
https://www.youtube.com/watch?v=Na1AwsUcyQY
Everybody says ‘leapfrogging’, but I think this is more accurately described as investment. China invests in technology, industry, efficiency. The BYD case in particular is a clear example of up front investment in productivity yielding returns on efficiency.
In most western economies, R&D investment has to fight with the likes of stock buybacks and austeriy and most of the time loses. It’s no surprise to me to see a country which didn’t go all in on austerity over the last 15 years slowly break ahead while those who did fall further behind. China is just better at productive investment. Western economies are centered around unproductive stock market video games(I not in passing China is also starting to make better video games nowadays too).
What are dark factories? I don’t think I’ve heard the term before.
Max,
It is a “lights out” factory. There are no
regular production workers on the
floor, so there’s ostensibly no reason
to have the lights turned on.
J
Doesn’t sound like that much of a feat, tbh. Sure, it means there are a lot of specialized robots inside and everything is automated but still.
The factory itself is less of a feat, the real heavy lifting is in the product design, engineering and supply chain to enable the entire product(s) to be built with no human intervention. There is no room for any variances, every last component, down to sheets of paper and rolls of tape, has to be precisely spec’d out, documented, and subject to exhaustive QC before it can be allowed in the factory. There’s no ability to implement rolling engineering change orders, there’s no flexibility in accommodating supply chain disruptions (ie new vendors), no rework process to fix minor errors caught by QC.
If you can’t seem to imagine what those are like, this video was posted on this site a few days ago and seems relevant to this topic as well.
It is a promotional video, and I wonder about who might buy the resulting output. There are deep implications for the purpose of an economy if we continue down this road.
https://www.youtube.com/watch?v=yezR-mH12xs
These technologies will diffuse to the U.S. and to the rest of the world, and will require less energy, so it could be good news.
Not if we succeed in containing ourselves!
With respect to “The Chinese are keenly aware of this great embarrassment, this blot on their historical record, and many have argued that this, more than wealth, drives Chinese policy today,” I think it is worth reading Orville Schell and John Delury’s Wealth and Power: China’s Long March to the Twenty-First Century.
Though 11 years old now, the book will remain relevant. One essential aspect of the story they tell is of China’s “Century of Humiliation” or “Century of Shame.” They make a very strong case that the leaders of China, now, are determined to learn from their past and restore their country to Wealth and Power.
In my opinion, it will be hard if not impossible for American foreign policy makers to understand the motivations of the leadership of China. We have a lot of evidence that American foreign policy elites are bad at understanding the world, witness McNamara saying he did not understand Vietnamese motivation in “Fog of War,” the Errol Morris documentary. Even without such evidence, the history of China being reduced to poverty and shame by the Western powers is probably something that is practically speaking unfathomable by the kind of person who gets into power in Washington. Respect for sovereignty of other countries, for example, is probably a deeply held ideal for Chinese foreign policy planners. This concept that other countries should be fundamentally sovereign is one area that Washington may never understand, and will always put it at odds with China.
Chinese have starvation in living memory. The world’s greatest famine. You bet they’re going to do anything to prosper.
First they have us with numbers. Then their government wants people educated and makes it happen.
The US doesn’t want it’s people educated or prosperous. It wants “happy” slaves for Musk.
You can’t deny that China has been uplifted while the US is being downgraded. Rather rapidly.
Exciting times in China. FWIW the quality and originality of work in China in the biomedical sciences has been increasing dramatically as well.
Thanks! Examples?
Thomas