China has spent 3.6 times more than the US on chipmaking subsidies over the past decade — $142 billion and counting, easily outweighs CHIPS Act
A new study describes the outcome of Beijing's $142 billion drive as a "disruptive failure" at the leading edge.
Semiconductor industrial policy spending in China totaled around $142 billion between 2014 and 2023, roughly 3.6 times higher than the $39 billion committed in the United States over the same decade, according to a recent report published by the Center for Strategic & International Studies in early March. The report argues that China's massive technology spending has given it a better position on the international stage in both influence and power, and it suggests "other governments need to respond pragmatically to reduce the downside costs."
The spending figures are drawn from a chart compiled by Boston Consulting Group and the Semiconductor Industry Association, and report author Scott Kennedy, senior adviser and Trustee Chair in Chinese Business and Economics at CSIS, argues Beijing's chip drive has still amounted to what he calls a "disruptive failure" at the leading edge.
The comparison covers direct industrial policy support across the full semiconductor value chain. South Korea placed second at $55 billion, followed by the EU at $47 billion, Japan at $17.5 billion, and Taiwan at $16 billion. However, the report’s 2014 to 2023 window predates most disbursements from the U.S. CHIPS and Science Act, which was signed into law in August 2022, and it doesn’t capture the third phase of China's Integrated Circuit Industry Investment Fund (‘Big Fund III’), launched in May 2024 at roughly $47.5 billion.
The report draws on 2025 Semiconductor Industry Association data to argue the spending has not yielded a leading-edge breakthrough. U.S.-headquartered chip firms still account for more than 50% of global semiconductor shipments by company headquarters, against 4.5% for Chinese firms.
SMIC's share of global fab production is also placed at roughly 6% as of mid-2025 by the report, ranking it third behind TSMC and Samsung. SMIC sits at least two to three generations behind TSMC on analyst assessments, with reported wafer yields as low as 20% for its 5nm process and 25% to 46% for 7nm, while Intel, Samsung, and TSMC press on with 2nm processes and achieve yields as high as 90%.
SMIC remains realistically locked out of the sub-7nm node without access to ASML's EUV (and, soon, potentially DUV) lithography scanners, the report states, and with Kennedy arguing the gap to the cutting edge may widen rather than close. Chinese labs have tried Frankensteining together reverse-engineered EUV tooling, but these haven’t yet produced a single chip. On the design side, Nvidia's share of the global GPU market sits above 90%, with Chinese entrants including Huawei's Ascend series, Alibaba's T-Head, Cambricon, and Moore Threads all trailing on compute performance. Kennedy cites R&D intensity as a further constraint, with American chip firms reinvesting an average of 17.7% of sales into research and development, against 9.2% at Chinese counterparts.
Beijing's Big Fund III, launched in 2024, was aimed at closing gaps in fab tools, EDA software, and AI accelerators. The report cites analyst Jimmy Goodrich characterizing China as likely to remain a "fast-follower" perennially challenged to keep pace with global leaders for the foreseeable future, a conclusion Kennedy endorses in the face of U.S. export controls on advanced lithography and the widening R&D spend gap.
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Luke James is a freelance writer and journalist. Although his background is in legal, he has a personal interest in all things tech, especially hardware and microelectronics, and anything regulatory.
-
Notton "Center for Strategic & International Studies" is a think tank, and it should be pointed out in the article.Reply
Their funding sources can influence their objectivity.
As in, they might say what their patrons want to hear, so take what they say with a grain of salt. -
pug_s I am sure if Chinese companies are able to buy the latest EUV lithography machines, things would be much different. Instead they are spending billions of R&D to developing their own.Reply -
blitzkrieg316 They are about 3.6 times bigger than the US. The US GDP is also 50% higher. This means China is subsidizing multiple times more per person but from a much smaller tax pool thus plunging their artificially supported economy further into oblivion.Reply
This site is very pro communist China... disturbing. Follow the money I guess... -
usertests Reply
Chyna spending $142 billion on semiconductor infrastructure/subsidies over 10 years isn't going to bankrupt them. Their annual government spending is $4.3 trillion, with $3.5 trillion in revenue (US is $7.1 trillion spending, $5.3 trillion revenue).blitzkrieg316 said:They are about 3.6 times bigger than the US. The US GDP is also 50% higher. This means China is subsidizing multiple times more per person but from a much smaller tax pool thus plunging their artificially supported economy further into oblivion.
$14 billion a year that has developed some success stories like YMTC, and may eventually deliver them EUV and other technologies that they are denied. -
nookoool Replyblitzkrieg316 said:They are about 3.6 times bigger than the US. The US GDP is also 50% higher. This means China is subsidizing multiple times more per person but from a much smaller tax pool thus plunging their artificially supported economy further into oblivion.
This site is very pro communist China... disturbing. Follow the money I guess...
About 10$ per a chinese citizen a year directed to semiconductor subsidy will lead to their economy into oblivion? -
zsydeepsky "disruptive failure" is objectively correct, the only problem is...China never planned to achieve that in 2026.Reply
in China's roadmap, it was never considered feasible to own domestic EUV tech by the year 2026; instead, the most optimistic target was to own domestic DUV tech by the year 2025, thus it could secure 7nm-level chips supply to support the basic needs for IT/AI requirements. The policy maker's target was always, to be 2-3 generations behind the West, until China completes its own EUV.
...so far, it has achieved perfectly, even beating its most optimistic plans. By the looks of Huawei products, China has achieved domestic DUV & 6nm chips in year 2025.
and along the way, the entire semiconductor ecosystem around the mature nodes, this report actually mentioned it yet considered it as a somewhat "failure":
advanced “legacy” chips—28 nm and higher—where their market share has risen substantially. While legacy chips occupy a large share of the overall semiconductor industry, a lead there still leaves Chinese companies behind their foreign counterparts and industry leaders.
in the report, it quoted a 2024 article from The Wire, so I gonna just also quote the graph in that article as well:
Which country is projected to expand their chips capacity the most over the next decade?
Or, a more recent 2026 article from Reuters:
China's manufacturing capacity for chips made on mature 22nm to 40nm process nodes - used in cars, smartphones and electronics - is projected to reach 42% of global output by 2028, up from 37% in 2026
So the data is there, but it's up to YOU to consider if it's a failure or a success. -
Thunder64 Replyzsydeepsky said:"disruptive failure" is objectively correct, the only problem is...China never planned to achieve that in 2026.
in China's roadmap, it was never considered feasible to own domestic EUV tech by the year 2026; instead, the most optimistic target was to own domestic DUV tech by the year 2025, thus it could secure 7nm-level chips supply to support the basic needs for IT/AI requirements. The policy maker's target was always, to be 2-3 generations behind the West, until China completes its own EUV.
...so far, it has achieved perfectly, even beating its most optimistic plans. By the looks of Huawei products, China has achieved domestic DUV & 6nm chips in year 2025.
and along the way, the entire semiconductor ecosystem around the mature nodes, this report actually mentioned it yet considered it as a somewhat "failure":
advanced “legacy” chips—28 nm and higher—where their market share has risen substantially. While legacy chips occupy a large share of the overall semiconductor industry, a lead there still leaves Chinese companies behind their foreign counterparts and industry leaders.
in the report, it quoted a 2024 article from The Wire, so I gonna just also quote the graph in that article as well:
Which country is projected to expand their chips capacity the most over the next decade?
Or, a more recent 2026 article from Reuters:
China's manufacturing capacity for chips made on mature 22nm to 40nm process nodes - used in cars, smartphones and electronics - is projected to reach 42% of global output by 2028, up from 37% in 2026
So the data is there, but it's up to YOU to consider if it's a failure or a success.
How has giving up on leading edge nodes worked out for GloFo? They too tried the "fast follower" strategy until it was no longer feasible for them. That said, they did not have the huge cash influx and a government hell bent on them becoming a dominant player behind them. -
Beltrano "Samsung, and TSMC press on with 2nm processes and achieve yields as high as 90%" Yeah and I'm Santa Claus!Reply
Everyone know that Samsung is fighting hard to get good yieald on its 2nm node! -
zsydeepsky Reply
The biggest difference is that Global Fundries does not have the US government to protect its market through sanctions.Thunder64 said:How has giving up on leading edge nodes worked out for GloFo? They too tried the "fast follower" strategy until it was no longer feasible for them. That said, they did not have the huge cash influx and a government hell bent on them becoming a dominant player behind them.
the cash flow SIMC or other Chinese fab got from the Chinese government is actually small, like the article said, $142B through 10 years (and not only for fabs). During the same time period, the Chinese domestic semiconductor market value increased from ~$500B to ~$2000B annually, and the cumulative worth of the market is around $1.3T. This is the market GloFo doesn't have for its old nodes, and which made all the Chinese government investments sustainable.
In fact, if the US hadn't launched the tech war, the Chinese fabs would still be weak by now; the US sanctions are literally the best gift they could ever get.