The memory chipmaker CXMT soared 466% in its Shanghai Stock Exchange debut and became, in a single session, the most valuable company listed in mainland China. The scene repeats a familiar pattern: every time the United States raises a technological wall, Beijing responds by building its own door.
The Chinese semiconductor company ChangXin Memory Technologies (CXMT) made its debut this Monday, July 27, 2026, on the STAR Market of the Shanghai Stock Exchange with a 466% surge on its first day of trading. Its shares, offered at 8.66 yuan, closed near 49; the company’s market capitalization transitorily surpassed 3.3 trillion yuan —some 487 billion dollars— and made it, in a single day, the most valuable company listed on mainland China’s exchanges, ahead of the state banking giant ICBC. Its offering is, moreover, the largest in all of Asia so far in 2026 and the largest listing by a semiconductor firm in the history of the Chinese stock market.
It was not an isolated speculative anomaly. It was the public coronation of a project that had been advancing quietly for nearly a decade, and a political signal hard to ignore amid the full-blown technological war between Washington and Beijing: China-made memory had just received massive backing from Chinese capital itself.
CXMT, born in 2016 in Hefei, capital of Anhui province, is today the fourth-largest DRAM memory maker on the planet, with roughly 7.7% of the world market at the close of 2025. DRAM —the random-access memory that powers everything from a phone to an artificial-intelligence server— had for decades been the near-exclusive preserve of three players: South Korea’s Samsung, its compatriot SK hynix, and America’s Micron. That oligopoly together controls close to nine out of every ten memory chips in the world. The emergence of a fourth player under the Chinese flag disrupts a board that Washington believed it had firmly sealed.
Sizing up the record
It is worth pausing on the figures, because their magnitude is what gives the episode its political meaning.
In absolute terms, the offering raised 57.92 billion yuan —about 8.6 billion dollars—: it shattered the sector’s previous record in China, SMIC’s debut in 2020, and marked the largest listing recorded on the mainland since the Agricultural Bank of China went public in 2010.
Demand overflowed every forecast. Retail investors oversubscribed the offering more than 200 times; the institutional tranche, more than 500. Millions of individual orders competed for a slice of shares that, by their scarcity, became the object of a stock-market frenzy of historic proportions.
That valuation of 3.3 trillion yuan —the 487 billion dollars of its debut capitalization— is hard to grasp without points of comparison. For a Latin American reader, the order of magnitude is best appreciated this way: in a single session, a memory factory founded ten years ago reached a market valuation greater than the annual gross domestic product of countries such as Chile or Finland. And the appetite was not merely nominal: the volume traded that day, some 141 billion yuan, marked the largest movement in a single session ever recorded by any stock in the history of the mainland Chinese market.
Such a market phenomenon had, nonetheless, its flip side. The offering was so colossal that it siphoned liquidity from the rest of the market: to finance the purchase of CXMT shares, numerous investors sold off other positions, and the STAR 50 index fell nearly 20% from its July peak. The shock crossed borders. Shares of Micron and SK hynix retreated on fears that the new Chinese capacity will end up causing a global oversupply of memory and eroding the leaders’ margins. A single debut, in short, moved the entire board of a global industry: within China, draining capital from its peers; beyond it, sowing the disquiet of a price war to come.
The wall that built the giant
None of this happened in a vacuum. It happened while the United States maintains, and periodically expands, a regime of export controls designed precisely to curb the rise of China’s semiconductor industry: restrictions on advanced chips, on design software, and above all on extreme ultraviolet (EUV) lithography machines, the equipment without which —according to industry orthodoxy— it is impossible to print the finest circuits of last-generation memory.
That regime is neither improvised nor static: it is an architecture that has been tightening step by step. It began in October 2022 with the ban on exporting to China the most advanced computing chips and the equipment to make them; it was reinforced a year later to plug leaks; in December 2024 it was expressly extended to high-bandwidth memory; and in January 2026 the Trump administration added a 25% tariff on imports of advanced semiconductors. The theory that sustains the edifice has a name in the international-relations literature: weaponized interdependence —the conversion of the nodes of the global network, those points through which critical technology must necessarily pass, into instruments of coercion—. Washington does not dominate the chip chain because it manufactures it entirely, but because it controls, or pressures whoever controls, its bottlenecks.
And here the story takes on an almost novelistic edge. In April 2026, the U.S. Congress advanced in committee the so-called MATCH Act —described by lawmakers themselves as the largest package of semiconductor export controls in the history of Congress—, a law conceived to seal off those bottlenecks comprehensively and to compel the Netherlands and Japan to align within 150 days, under threat of applying the Foreign Direct Product Rule, which extends U.S. jurisdiction to any piece of equipment that incorporates a single component or line of software of American origin. Among the facilities the bill expressly names as a target to be strangled figures, in so many words, ChangXin Memory Technologies. Put another way: the very company that this Monday broke every record on the Chinese stock exchange appears written, by name, in the legislation the superpower is designing to asphyxiate it. The debut and the sentence coexist in the same file.
Here lies the most revealing paradox of the whole episode. The United States forbids China a machine that the United States does not even manufacture. EUV lithography is not produced in Silicon Valley: it is made by a single company on the entire planet, the Dutch firm ASML, headquartered in the small city of Veldhoven. One hundred percent of the machines capable of printing the world’s most advanced chips comes out of that single European company, which took three decades and billions of dollars in research to master a technology no competitor managed to match. Washington does not block Beijing with a tool of its own: it blocks it with someone else’s monopoly, pressuring the Dutch government —and Japan along with it— since 2018 so that ASML does not sell a single unit to China. To this day, not one of those machines has crossed the Chinese border by legal means. The superpower’s lock depends, ultimately, on the obedience of an ally and of a company that does not belong to it.
And even so, the ant is already digging beneath the wall. In December 2025, the Reuters news agency revealed that a laboratory in Shenzhen had assembled its own EUV lithography prototype, developed by a team made up largely of engineers who had previously worked at ASML itself. It is in the testing phase and years away from serial production; no sensible person would yet announce the breaking of the monopoly. But the gesture is unmistakable: China did not accept the ceiling as a natural limit, but as the next objective to demolish.
It is worth, all the same, not falling into triumphalism. CXMT built its entire roadmap without access to that EUV lithography, and still lags several years behind its South Korean rivals in the most coveted segment of the moment: the high-bandwidth memory (HBM) that feeds the graphics processors of the artificial-intelligence fever. The technological ceiling exists and is real: China’s self-sufficiency in chip-manufacturing equipment barely hovered around 13.6% in 2024, and in the most delicate links —lithography, metrology, ion implantation— dependence on foreign suppliers remains almost total. But the essential point is not that China has already won the chip race. The essential point is another, more uncomfortable for whoever imposed the blockade: the blockade did not stop China. It pushed it.
The aggregate numbers confirm it. In the first half of 2026, Chinese industry manufactured close to 279.8 billion integrated circuits, 23.1% more than a year earlier, and profits in the integrated-circuit manufacturing sector soared dizzyingly, according to data from the National Bureau of Statistics. The situation helped —the global DRAM shortage, caused when the big three shifted their lines toward AI memory, left a gap that CXMT filled with discipline—, but a situation is only seized if the capacity to do so was built beforehand. And that capacity was built, brick by brick, during the years of low profile.
The pattern: as with the space station, so with the chips
Anyone who has followed China’s trajectory over the past two decades will recognize the figure immediately. It already happened, with almost identical choreography, in space.
In 2011, the U.S. Congress passed the so-called Wolf Amendment, which barred NASA from cooperating substantively with China’s space agencies. The practical effect was to leave China out of the International Space Station, the great orbital laboratory operated by the United States, Russia, Europe, Japan, and Canada. No Chinese astronaut has ever visited it. Facing the closed door, Beijing did not protest: it built its own. It launched the core module Tianhe in 2021, docked to it the Wentian and Mengtian laboratories in 2022, and placed in orbit a station of its own, permanently crewed, called Tiangong —”heavenly palace”—. If the International Space Station retires around 2030 without a ready successor, the only human home in Earth orbit could end up being the Chinese one.
The parallel is not a decorative metaphor: it is a method. The exclusion, conceived as a tool of containment, functioned as a catalyst for self-sufficiency. What Washington read as a lock, Beijing read as a deadline. And in both cases —the orbital and the silicon— the response was the same: to work like an ant. Without fanfare, without premature triumphal announcements, accumulating infrastructure, patents, wafers, and technical cadres until one day the result becomes impossible to ignore. Monday’s market debut was precisely that day: the moment when the underground labor surfaced and the market, with 9.4 million retail orders, put a price on it.
And they are not two isolated cases: they are a habit. Every time the West decreed a ceiling for it, China turned it into a floor. It was told it would never master high-speed rail, and today it operates the most extensive network on the planet. It was cut off from telecommunications technology, and Huawei ended up leading the global rollout of 5G. Bets were placed that it would never compete in the automotive industry, and BYD dislodged the traditional manufacturers in the electric car. It was written off in solar energy, and today most of the world’s panels come out of its factories. The list is uncomfortable because it repeats with a regularity that by now should surprise no one —and yet always surprises—. The memory case adds a precise geopolitical nuance: the wall CXMT is climbing is this time not American, but South Korean. Samsung and SK hynix, Seoul’s two industrial jewels, together control close to two-thirds of the world DRAM market. To reach them means measuring oneself against the technological heart of South Korea, the country that made memory its export hallmark. China has already broken the dishes against the United States and against Europe; now it breaks them against Seoul.
Self-sufficiency as a doctrine of state
None of this is accidental or spontaneous. China’s response to the encirclement obeys a doctrine of state with deep roots. In Mao Zedong’s time it was called zili gengsheng, “regeneration through one’s own forces”: the principle of being self-sufficient in the face of a hostile world. Xi Jinping updated it and elevated it to the center of economic planning under the formula keji zili ziqiang —”self-reliance and self-strengthening in science and technology”—, enshrined as the axis of the fourteenth five-year plan and geared to the “dual circulation” strategy, which reorients the economy toward the domestic market without renouncing the external one. Technological independence ceased to be a rhetorical aspiration and became a state policy endowed with budget, targets, and calendar.
The very design of the STAR Market, inaugurated in 2019, is part of that scaffolding: a stock board created on purpose to channel national savings toward the companies the state deems strategic, without depending on the capital or the exchanges of the West. That CXMT raised 8.6 billion dollars from Chinese investors, on a Chinese exchange, to manufacture in China what the West denies it, is no accident of the market: it is exactly the result the doctrine set out to produce. And there is a final twist, almost cruel for whoever imposed the encirclement: much of the investor appetite is explained by the blockade itself. By cutting off China’s access to high-end foreign memory, Washington guaranteed CXMT a captive domestic market of hundreds of millions of devices. The sanction manufactured the demand that today finances the sanctioned.
Time as strategy
Beneath all these episodes lies a trait that Chinese political culture cultivated over centuries and ended up turning into doctrine: knowing how to wait. Where the Western political cycle measures success in quarters and elections, Beijing reasons in decades. Deng Xiaoping condensed it into a formula that guided the country’s foreign policy for a whole generation —to hide one’s own strength and bide the moment—, and that patience transformed, over time, into industrial method. ASML took thirty years to tame EUV lithography; China seems willing to invest as many replicating it, without the timeframe intimidating it. CXMT raised plants and accumulated patents for years, with a deliberately low profile, while the world looked the other way, until the global shortage opened a crack and it stepped through without hesitation. The ant is in no hurry because it does not need to be: in a civilization accustomed to thinking in centuries, time is rarely an adversary. It is usually an ally.
What the debut reveals
There is, finally, a semiotic reading not worth overlooking. A stock-market debut is not only a financial event: it is an act of signification. What the Shanghai market staged on Monday was not merely the valuation of a company, but the proclamation of a thesis —that China can produce, finance, and consecrate its own strategic technology without anyone’s permission—, addressed equally to its domestic audience and to its external adversaries. The 466% figure operates here as a sign before it operates as a datum: it communicates confidence, will, and above all the promise that the encirclement will not be eternal. In the grammar of power, a record like this reads as a declaration. And its statement is unmistakable: technological sovereignty and the sovereignty of capital advance, in China, hand in hand.
Open questions remain, and it would be dishonest to close them today. No one knows whether CXMT will manage to cross the EUV lithography barrier, nor how long it will take to catch Samsung and SK hynix in high-bandwidth memory, nor whether the euphoria of its debut will withstand the scrutiny of future balance sheets. The technical distance is still there. But the direction of movement is unmistakable, and that is the fact Washington ought to weigh before designing the next wall.
Because the pattern is already written. First it was space. Now it is the chip. Every time the blockade seeks to lock China in, it ends up manufacturing for it a reason —and a captive market of hundreds of millions of consumers— to make at home what was denied to it abroad. The ant does not argue with the boot. It carries on in silence, and waits. It knows —because it has been learning it for centuries— that no wall is eternal, and that everything one country manages to manufacture, sooner or later another country learns to manufacture too.
