The vulnerabilities holding back Chinese industry
Despite its prowess, China has not been able to overcome dozens of ‘choke points’ that are the essential building blocks of modern manufacturing
Edward White in Shanghai and Harry Dempsey in Tokyo
In the early 1950s as the Communist party set out to rebuild a China ravaged by years of war, Mao Zedong, the revolutionary hero and party chairman, bluntly assessed the bleak state of the nation’s industry: “We can make tables and chairs, teacups and teapots . . . But we cannot make a single motor car, plane, tank or tractor.”
Soon the No 1 Tractor Plant and an adjacent factory producing ball bearings was established in Luoyang, an ancient city in the central Henan province — one of 156 industrial projects conceived under the party’s First Five Year Plan, according to Karina Khasnulina of Leipzig University.
Just a few years later, in 1958, the first Chinese-made tractor, named Dongfang Hong, “the East is Red”, rolled off the production lines.
Seventy years on and China has risen to be a global powerhouse, the world’s second-biggest economy and one of two true military superpowers.
And yet, as the country’s leaders in Beijing are acutely aware, the nation has not been able to overcome dozens of industrial “choke points”.
From a western standpoint, the good news is that these choke points are ultimately holding back China’s quest for independence and leave it vulnerable to US pressure in an era of trade wars and export controls.
As well as cutting-edge computer chips, they also include a series of obscure components and materials that are essential building blocks of modern manufacturing.
The bad news, for the west and its companies at least, is that China is addressing these problems in a methodical and systematic manner and is using tools such as AI to advance more quickly.
Leading European, Japanese and American rivals, which have long been insulated from competition from China because of problems with quality and inconsistent yield in Chinese factories, are now on high alert.
“What should give us pause is that the dynamic of China’s technological dependency on the west is rapidly changing,” says Elisa Hörhager, the Beijing-based chief representative in China of the Federation of German Industries, known as BDI.
“Many foreign companies still hold a clear edge when it comes to high-quality industrial products, thanks to their reputation for precision and engineering excellence.
But Chinese competitors are catching up fast.”
While several choke points, such as advanced semiconductor manufacturing, remain years, if not decades, from being conquered, others appear close to being resolved.
These products include carbon fibre used for aviation and ball bearings.
Last month, Chinese leader Xi Jinping chose the factory floor of the current iteration of the Mao-era Luoyang factory, now operated by the state-backed Luoyang Bearing Group, to issue a rallying cry for China to address any remaining potential choke points.
The visit to Luoyang by Xi, China’s most powerful leader since Mao, was both symbolic and timely.
Successive American presidents have steadily expanded export controls to restrict access to cutting-edge technologies, fearful that China’s technological might benefits the country’s military and threatens US national security.
These controls — or the threat thereof — have driven Beijing to throw more resources into mitigating its vulnerabilities than they would have otherwise, says Kyle Chan, a researcher in Chinese industrial policy at Princeton University.
The motivation for Xi’s quest for industrial self-sufficiency has only intensified since Donald Trump returned to the White House and launched a global trade war that threatens to accelerate the decoupling of the world’s two biggest economies.
“From the past reliance on imported matches, soap and iron, to now becoming the world’s largest manufacturing country with the most complete industrial categories, we have taken the right path,” Xi told ball bearing engineers and technicians in Luoyang.
“China must keep improving our manufacturing sector, insist on self-reliance and self-improvement [and] master key core technologies.”
At a nondescript business park on the outskirts of Hangzhou, eastern China, there are signs that a quiet industrial revolution is taking place.
A team of Chinese engineers at DeepVision Technology is redeploying image processing technology, originally developed for gastrointestinal surgery, to solve a problem that has stubbornly thwarted Chinese engineers for decades: how to make high-quality ball bearings, and at scale.
High-end ball bearings are crucial for reducing friction in everything from high-speed trains and tunnel boring machines to electric vehicles, humanoid robots and drones.
Cosimo Ries, an energy analyst with consultancy Trivium China, points out that for machinery like offshore wind turbines — which are now being built close to 200-metres tall and need to last for around 25 years — bearing manufacturers face “incredible reliability requirements” as their products must withstand “huge” amounts of weight and pressure.
While China is by far the biggest single ball bearing market in the world, the $53bn global bearing industry is dominated by Sweden’s SKF along with Germany’s Schaeffler, US group Timken and Japanese companies NSK, NTN and JTEKT.
The top six manufacturers account for about 55 per cent of the global rolling bearing market, according to SKF’s annual report. Chinese groups hold about 25 per cent.
A 2020 study by researchers at the Wuhan University of Science and Technology noted a “significant gap” between Chinese companies and their Swedish and Japanese rivals in terms of controlling the degree of microstructural defects and dimensional accuracy.
According to a research paper from Kai Yuan Securities, a Chinese investment group, published four years later, mid to high-end products still accounted for only about 20 per cent of total production among Chinese bearing manufacturers.
Sitting in an office above his factory, Wang Shuailin, the cheerful 38-year-old founder of DeepVision, sketches out a drawing of how his team combines an AI chip with image sensors to identify defects in a bearing’s shape, size and structure as small as 2 micrometres — 0.002 millimetres.
DeepVision’s inspection system, Wang says, is radically improving quality control for his clients, which include Luoyang.
People used to have the impression that Chinese products, not just limited to bearings, were cheap but lower quality . . . I don’t think there’s much difference anymore
One Chinese customer, he says, has seen the “qualification rate”, meaning the percentage of the ball bearings it makes that are good enough to be sold to customers, jump to 97 per cent, from below 90 per cent.
Another would previously receive as many as 400 customer complaints about quality per year, but now receives just two or three, while its staffing requirements for inspection have fallen from 150 to just a few people.
“There is a stereotype that China doesn’t have the ability to produce such high-precision machines,” Wang says.
While foreign rivals have accumulated a technological advantage over a long period of time, he believes that when it comes to introducing AI into manufacturing, which kicked off around 2017-18, everyone has been “starting from the same point”.
“And China is now adopting these technologies faster, so our products will be better,” he adds.
Research by China Policy, a consultancy, commissioned by BDI and published in May, showed that China’s 4mn factories are only in the early stages of seeing the impact from the kind of industrial application of AI showcased by DeepVision.
But that could soon change as Beijing has set a target of having an advanced level of smart manufacturing in most big factories in the next 10 years.
Still, as it stands today the world’s biggest foreign ball bearing suppliers retain a significant technical advantage over China.
This is particularly true when producing components vital for machinery that is required to function for long durations and in extreme environments or temperatures, such as in space or chemical plants.
But even the most dominant players are wary of Chinese progress.
Vivian Wang, head of marketing in China for SKF, points to the Gothenburg-headquartered group’s history of selling bearings in China, which dates back to around 1912, just as the 268-year Qing Dynasty came to an end.
She notes the group’s large local footprint, which includes nine manufacturing units, 6,000 employees and a high localisation rate across its Chinese supply chain.
“We acknowledge the growing capabilities of our competitors, including those in China, but we act from a position of strength,” she says.
Sadatsune Kazama, deputy general manager of product design for bearings at Osaka-based NTN, the world’s fourth largest ball bearing manufacturer, says that Chinese groups are now on par with the industry leaders on manufacturing technology but still lag behind on design capabilities.
“People used to have the impression that Chinese products, not just limited to bearings, were cheap but lower quality . . . I don’t think there’s much difference anymore.”
However, he says Japanese companies were still better at long-term quality, reliability and after-market support.
“We’ve made bearings for 100 years and we do it based on a large amount of experience that we've built up”.
A Japanese government industrial policy official, who asked not to be named, suggested that Chinese quality still needed to stand the test of time.
“It’s amazing how far China has come.
But it’s not even 20 years since they started making precision technology.
The goods need to last for more than 30 years, so we’ve not yet seen how durable their products are,” they say.
Over the course of 2018, Science and Technology Daily, a Chinese state-run newspaper, published a series of 35 articles outlining the country’s most acute industrial vulnerabilities.
Ben Murphey, a former CIA linguistics expert, translated the articles at the Center for Security and Emerging Technology, a research group within Georgetown University’s Walsh School of Foreign Service. CSET analysts narrowed the 35 identified choke point technologies down to the 14 that had been “most vexing”.
These technologies, the analysts said, reflected three characteristics: the technology is the exclusive preserve of a small number of US, European or Japanese companies; few Chinese providers are making progress developing high-end versions of the technology; other unique factors making import substitution unusually difficult.
They also noted that rather than acting patriotically, Chinese companies, including state-owned groups, “harbour doubts” about the quality of products from their compatriots.
While there is evidence that China is edging closer to eliminating many of the long-held industrial choke points, the thorniest import dependencies, requiring the most investment to overcome, relate to chipmaking.
According to official Chinese trade data, the value of imported photolithography machines, which are crucial for making the most advanced semiconductors, almost quadrupled to $47bn, from $12.5bn, in the past 10 years.
Chan, the industrial policy expert at Princeton University, has analysed the S&T series and notes that while the choke points are still seen by Beijing “as a major problem”, there are “clear signs” of progress.
Among those choke points identified in 2018 that have since been resolved are high-end radio frequency components and operating systems — technologies where Huawei has become self-sufficient — as well as lithium battery separators, where Chinese suppliers like CATL and BYD are now world leading, and lidars, the laser sensors used in self-driving cars, which are also dominated by Chinese suppliers.
Others are in the process of being mitigated.
This includes vacuum evaporators, which are used to fabricate layers of organic film and metal electrodes in OLED panels, as well as photoresist material used to transfer circuit patterns for chips.
Chan says that beyond being motivated by the threat of western export controls, the lack of access to foreign suppliers and their personnel during the pandemic was another driver that sharpened Beijing’s focus.
But he also notes that a key dynamic holding back progress has long been from Chinese companies buying these components and materials.
They have often preferred to stick with foreign suppliers, which are considered more reliable, rather than “take the leap with less tested Chinese suppliers”.
One such area is carbon fibre composite cascades, which are used in an engine’s casing to help aircraft land safely.
Japanese group Nikkiso has a market share of 90 per cent.
Takeshi Iwaoka, head of the aerospace division at Nikkiso, dismisses the threat from China, saying that rival companies have been trying for 40 years to catch up without success.
“The process involves laminated moulding into a complex shape, which is quite a sophisticated manufacturing method,” he says.
“We have a lot of proprietary knowhow when it comes to shaping and forming this kind of product, and other companies just can’t easily replicate it.”
He notes that the Japanese government is stepping up efforts to protect its crown jewel technologies related to aerospace and other strategic sectors from industrial espionage, citing hearings on economic security conducted by the National Public Safety Commission, related to the Cabinet Office.
“In the past, we weren’t receiving questions and inquiries from NPSC,” he said.
“The aerospace industry is one area we must protect and keep proprietary technologies.”
Following the breakout success of DeepSeek — the AI developer that rocked the tech world with advances achieved with far less computing power than US rivals — there is, domestically at least, a renewed sense of confidence in China’s homegrown technological capabilities, even in the context of being cut off from US technology.
In February, Ren Zhengfei, founder of tech group Huawei, was among a select group of business leaders to meet Xi in Beijing.
Seated opposite the Chinese leader, Ren told Xi that Huawei is spearheading a coalition of more than 2,000 companies aiming to boost Chinese semiconductor supply chain self-sufficiency to 70 per cent by 2028.
The 80-year-old former People’s Liberation Army engineer carefully drew on Chinese homonyms as he said hitherto concerns of the country’s “lack of heart and soul” — meaning weakness in hardware and software — had eased.
Hörhager, China’s BDI head, says that whereas Made in China 2025 — Beijing’s 10-year manufacturing advancement plan launched in 2015 — targeted key sectors with funding and policy support, industrial AI could “go far beyond that”.
This means not just eliminating the remaining choke points but boosting productivity “across China’s entire manufacturing base”.
Angela Huyue Zhang, a professor of law at the University of Southern California and a leading expert on China’s technology regulation and policy, cautions against viewing China’s choke point progress as a quest to be “number one” in a race against the US.
“Washington has misunderstood China’s goals; the real goal for China is to increase self-sufficiency, improve productivity, and drive economic growth — outcomes that will ultimately enhance the legitimacy of the Chinese Communist party,” she says.
“Tech development is only a means to achieve these goals; it is not an end in itself.”
Additional reporting by Ding Wenjie and Wang Xueqiao
0 comments:
Publicar un comentario