china

The US is accusing China of plagiarizing Anthropic models. We have three problems with that accusation.

by

Michael Kratsios, assistant to the president of the United States, did not bite his tongue this week when claimed that “We have information indicating that Moonshoot AI distilled Fable from Anthropic for the development of its K3 model.” Or in other words, they accuse China of plagiarizing the American company’s advanced AI model. There are several fundamental problems with that accusation. Let’s see them. Fear of Kimi K3. This model has just burst onto the AI ​​scene with overwhelming force. Benchmarks show that Kimi K3 is one of the best AI models in the worldwith performance close to Fable 5 and GPT-5.6, the best public models from Anthropic and OpenAI. This milestone has triggered its popularity among users and companies, and also the alarms in Washington. The immediate consequence has in fact been political, because since the Trump Administration there is now open talk of sanctions for “theft of technology” if the accusation is confirmed. The accusation. In his text, Kratsios explains the supposed mechanism used by Moonshot AI to copy his model. It states that the Chinese startup would have built an internal platform to make mass queries to US models, changing access methods to avoid detection and then transferring its capabilities to Kimi K3. He also adds that the company has managed to access servers with Nvidia GB300 chips in countries outside China to avoid export restrictions. Where is the evidence? The first problem we have with that speech is that a key piece of that speech is missing from the start: Kratsios does not support this accusation with technical evidence. It does not provide usage records, it does not provide examples of prompts and responses, nor query patterns. Nor does it provide independent analyzes that any researcher could review. What we have right now is an official story and accusation, not a proven “infringement.” The controversy over distillation. The key word in that speech is “distillation.” In AI, distill a model It means training a new one using the responses of another. The “master” model answers thousands or millions of questions, and the “student” model learns to imitate its reasoning patterns at lower cost and size. The industry itself describes distillation as a legitimate and standard technique when applied to permitted or proprietary models. The red line appears when it is applied to third-party models, on a large scale and without permission, which according to the US Government is theft of intellectual property. But crime, what is called crime, is not. The second problem is that the statement and the accusation are not supported by any clear and defined legal framework. Neither the US nor China currently have a specific law that says under what conditions distilling a rival model is an intellectual property crime. You may violate an API’s terms of service or even end up obtaining sensitive information from companies, but there is no jurisprudence about it. In fact, until recently the debate in the US was about regulating US frontier models like Mythos, which were considered too dangerous. Suddenly the debate is now on sanctioning the distillation of models. Everything is moving too quickly, once again, to adapt the legislation. Hypocrisy made in USA. The third big problem we have with that accusation is that the big American models—including those at Anthropic—have been trained on massive data sets that mix websites, books, code, newspaper articles, and other materials of which a notable portion is protected by copyright. own Anthropic reached an agreement with justice these days for that reason, Meta has already discovered it stealing terabytes of copyrighted books to train your models. Double yardstick. In the US, AI companies defend themselves by arguing that That falls under “fair use” of the content, but the authors neither gave their permission nor charged for it. That same industry and that same Government that have made it normal to train models on other people’s content complain that a Chinese startup uses the outputs of its models to train its own. They are not identical practices, but the double standard is difficult to ignore. AI, once again, as a weapon. The Trump administration is taking this issue to another scale. By accusing Moonshot AI of using restricted chips and developing Kimi K3 by distilling Fable 5, he places distillation in the same box as industrial espionage or the theft of military secrets. AI is no longer a software product and becomes a strategic technological weapon. And there it is not so important to provide evidence or be technically precise: it is important that what China has done is an attack on its (former?) technological hegemony. If this type of accusation sounds familiar to you, you have a good memory. In Xataka | An AI model did not have access to the internet. So he thought it was better to have it and decided to hack something along the way

This is how China circumvents 2,000 ASML patents with a laser trick

by

The extreme ultraviolet lithography (EUV) is the tightest bottleneck in the entire semiconductor industry. Only one company in the world, the Dutch ASML, manufactures these machines, essential for producing integrated circuits below 7 nm, and since 2019 is prohibited from selling its most advanced models to China due to US pressure. Each device costs between 200 and 400 million dollars, and so far not a single Chinese client has managed to obtain one. A team from the Chinese Academy of Sciences, based at the Shanghai Institute of Optics and Precision Mechanics, has built China’s first laser-plasma EUV light source platform. In front is Lin Nanformer ASML scientist and technical manager of its light sources for metrology, who returned to China in 2021 as part of a state talent attraction program. His team has chosen a path that deliberately avoids ASML’s method, as well as more than 2,000 of its patents. Qiu Yanfang, Chinese columnist specializing in semiconductors, has summed it up unambiguously: This patent rodeo is not just a legal maneuver; It’s genuine engineering. And it is this assessment, published just two weeks ago, that has once again put the focus on how far China has really come in its race for technological independence in lithography. The longest road to mass production The technical trick is this: while ASML generates EUV light by firing powerful carbon dioxide lasers at moving tin droplets (a method known as LPP), Lin Nan’s team uses a 1-micron solid-state laser that hits a solid tin target. In a published article In December 2024, Lin and his team were already arguing that this type of laser could replace carbon dioxide lasers as a next-generation source thanks to its compact size and higher conversion efficiency. However, Qiu Yanfang herself qualifies her initial enthusiasm. In the short term, he maintains, this solid-state source cannot shake ASML’s dominance, since the entire global semiconductor industry has built its supply chains, patents and engineering expertise around the carbon dioxide laser. Although Chinese technology achieves results in the laboratory, it still has a long way to go until achieve large scale production. The real value of Lin’s achievement, in any case, lies in laying the foundation and forging a research team that covers everything from theory to applied engineering. The Chinese Government aims to produce chips with this technology in 2028, although many analysts consider 2030 a more realistic date. This effort also coexists with another parallel and much more hermetic path. As revealed Reuters As of December 2025, China has assembled a working prototype EUV machine inside a high-security laboratory in Shenzhen, although it has not yet manufactured functional integrated circuits. Huawei coordinates this initiative, which involves thousands of engineers: the Harbin Institute of Technology is in charge of the light source, the Changchun Institute of Optics for the optical systems, and SMEE (Shanghai Micro Electronics Equipment) of the final integration, with YesCarrier (the Chinese company backed by the Shenzhen Government and linked to Huawei) as another of the key actors. The Chinese Government aims to produce chips with this technology in 2028, although many of the analysts consider 2030 a more realistic date. For now, no Chinese prototype has proven to be capable of manufacturing commercial integrated circuits with its own UVE lithography, and doubts about the real speed of this advance have returned to the front line of information for a very different reason. US Secretary of Commerce Howard Lutnick has raised ASML senior managers the possibility that one of its SVU teams has reached China through clandestine channels, something that the Dutch company has denied emphatically: none of its 314 operational EUV machines in the world, nor is any of the 26 already withdrawn located in Chinese territory. China does not have its EUV machine yet. But, thanks to Lin Nan, it now has something that until recently it didn’t have either: its own path to try to build it. Image | ASML More information | Asia Times In Xataka | ASML’s new lithography equipment divides chipmakers. TSMC plans not to use it until 2030

Chinese AI geniuses could be billionaires in the US. The bad thing for the US is that they prefer to return to China

by

It is normal for someone to take six years to complete a doctorate at Carnegie Mellon University (CMU). A young man named Yang Zhilin got it in four, and immediately what had to happen happened: the big technology companies raffled him off. They wanted to sign him to Apple, Google or Meta, and both MIT and Stanford opened their doors to him. He could have stayed in the US and probably become a billionaire, but no. He returned to China, founded Moonshot AI, created the Kimi family of AI models and became one of the great new AI gurus of the Asian giant. His story is a good example of a big problem they are having in the US. A genius in the making. Yang graduated from the prestigious Chinese university of Tsinghua, and during his years at CMU he ended up collaborating on another of the “iconic papers” that were the seed of ChatGPT, the so-called “Transformer-XL“. That was a turning point in his career, because after that job he considered what he wanted to do with his life. Start up yes, but not in the US. His supervisor at CMU, Russ Salakhutdinov, I remembered in the Financial Times what he discussed with Yang, who made it clear to him that “if he didn’t at least try to create his own company, he would regret it for the rest of his life.” The curious thing is that he did not do it in the US, where he would surely have obtained plenty of financing and economic support, but rather he went to China, where he founded Moonshot AI. There he created the mentioned models of AI, of which the latest version, Kimi K3, has turned the market upside down: an open model that competes with the best of Anthropic and OpenAI and that has made Yang a current protagonist in the sector. Working for others, what’s up?. There is a narrative in Silicon Valley that talks about how immigration regulation in the US is designed to get young talents to join large companies, and not so much to found their own startups. That was much more complicated, so the Chinese talents who had just trained in the US realized that they had a better option: return to their country of origin and create their startups there. Source: Hoover Institution. Chinese talent returns to China. A study by the Hoover Institution at Stanford analyzed the profile of the 356 researchers who worked on DeepSeek and revealed something surprising: 53.5% never studied or worked outside of China. Of those who did have an experience of this type and had studied or worked in the US, 70% ended up returning to their country. The Asian giant is no longer an exporter of brainiacs: it now has a self-sufficient quarry that boosts the country’s AI industry. Less money, better execution. The difference between the US and China is not only financial, but operational. A Chinese entrepreneur in the sector explained in that FT report that although setting up his startup in the US would guarantee a valuation up to 10 times higher, in China he is much more likely to launch real solutions and production in less than five years. It doesn’t matter the sanctions, the risk of censorship or a much more conservative investment market: the density of engineers and the familiarity of the environment compensate. Source: Stanford University. The US does not make it easy. The phenomenon of the “return home” of Chinese engineers not only affects recent graduates, but also veteran professionals who were already installed in the country’s academic and engineering system. a survey conducted by Stanford University in 2024 among 1,304 scientists of Chinese origin who lived and worked in the United States revealed curious data. 73% did not feel safe working as a researcher in the country, and 65% pointed directly to a fear of harassment and racist violence. Political uncertainty and institutional distrust have only accelerated the exits. Silicon Valley has a problem. The trajectory of Moonshot AI, which in just three years and with a team of 300 people has launched a fantastic model, shows that there is life beyond Silicon Valley. With Chinese talent returning en masse to China, the question is whether the US Government will modify its requirements and proposals for obtaining a visa. If they don’t, the “brain drain” can be absolute. Image | Tommao Wang In Xataka | Four decades ago, China decided to invest in training millions of engineers. Today that plan gives it an advantage in the race for AI

China had been chasing “black gold” for years. Now it has started to turn it into something much bigger

by

An almost invisible fiber can end up supporting part of a gigantic wind blade, forming part of the main loading structures of a metro train or protecting several satellites during the ascent of a rocket. That is what we have seen with the carbon fiber: a material that seems modest when we look at it as a simple filament, but that takes on another dimension when thousands of strands become a piece capable of reducing weight without giving up high resistance. Behind these products there is not only advanced chemistry, but a technology that China has been trying to industrialize for decades with its own means. The change is not that the country has now discovered this material, but that it is managing to take it far beyond the laboratory. For years, some of its research centers managed to obtain advanced fibers, but China continued to encounter difficulties to industrialize certain grades with the required stability, uniformity and scale. That barrier began to break with new production lines and today the expansion covers the entire chain, from the precursor to the composite materials and finished parts. That is the real leap: converting a technology pursued for decades into an industry capable of feeding other sectors. To understand why all this matters, it’s worth first looking at what’s inside that black coil. Carbon fiber is not a sheet or a solid piece, but a set of extremely fine filaments grouped in bundles. When we read 3K, 12K or 48K, we talk about of 3,000, 12,000 or 48,000 filaments per bundle. Afterwards, these threads are normally combined with a resin to form a composite: the fiber provides strength and rigidity in the chosen directions, while the matrix holds the whole together, protects it and transmits the forces. From filament to the development of a key industry The advantage appears when we relate its performance to weight. A structure made of a carbon fiber composite can offer high resistance and rigidity using less mass than certain metallic solutions, something that can translate into lower consumption, more autonomy, greater payload or larger components. The Airbus A350 incorporates CFRP in 53% of its structurewhile composite materials represent around 50% of the primary structure of the Boeing 787measured by weight. But it is not a universal solution: the material is still expensive and requires complex processes, while composite structures can suffer internal damage that is difficult to detect and repair. To find the origin of this advance we have to go back several decades. China was researching carbon fiber since the sixties and seventiesbut the obstacle was not only to produce it once, but to repeat the result within a factory. In 2005, the Shanxi Institute of Coal Chemistry received a national mission to industrialize a fiber equivalent to aerospace T300. Three years later, on June 30, 2008, a line in Yangzhou produced the first stable roll of aerospace fiber with performance equivalent to the T300. Since that advance, China has been developing higher performance fibers, first with references equivalent to T700 and T800 and later with T1000. In Datong, a line with about 200 tons per year capacity had completed in November 2025 verification of continuous operation. Numbers, however, can fool us if we read them as a simple ladder. T300, T700, T800 or T1000 come from commercial names developed by Toray and are used as references for certain performances, but they do not form a universal international classification nor do they alone summarize the quality, reliability or final use of the material. The magnitude of the change can also be measured in tons. According to the ATA world reportChina concentrated in 2025 an operational capacity of 171,080 tons annuallys, 52.5% of the 326,080 counted worldwide. The data places the country in the lead by scale, but it does not mean that it produced exactly that percentage or that it dominated all segments. In other words, available capacity does not equate to effective manufacturing, and an industrial fiber intended for wind blades does not require the same controls, intermediate materials and certifications as a product prepared to be incorporated into an airplane. In 2025, China had an operating capacity of 171,080 tons per year, 52.5% of those recorded worldwide. What is truly relevant begins when we stop counting reels and observe everything that happens before and after. To obtain the fiber, the precursor must be prepared, oxidized and carbonized; Then come the fabrics, the prepregs, the compounds and the shaping of the pieces. In Jilin, this sequence already coexists within the same industrial hub. Provincial authorities speak of 190,000 tons of precursor capacity, 70,000 of fiber and 50,000 of composites, in addition to pultrusion, machining and product manufacturing lines. That chain already ends in objects that we can identify. In Jilin, 230 lines make structural plates for wind blades, and the company says its products are used in 95% of the blades on the Chinese market. The CETROVO train entered commercial service in Qingdao in January 2025 with carbon compounds in main load elementswhile the fairings manufactured by Tianjin Aisida had contributed, according to local authorities, to 46 commercial launches that reached orbit until June 2026. On a global scale, the material is also used in hydrogen tanks and systems intended to reinforce bridges. China already leads the way in added capacity, and domestic manufacturers such as Jilin Chemical Fiber and Zhongfu Shenying are expanding their capacity and offering higher-performance fibers. Compared to them, the Japanese company Toray maintains a leading position due to its catalogue, its international presence and its industrial integration; The American Hexcel stands out especially in the aerospace market, while the Japanese Teijin and Mitsubishi Chemical continue to be relevant technological players. What China is building is not just more fiber: it is the ecosystem that allows it to become an industry. Images | Lawless Capture | Vong Vathanak In Xataka | “We chose the wrong technology”: Jensen Huang returns to Japan to acknowledge his debt to Sega 30 years … Read more

China has just launched its new heavy ion accelerator. He hasn’t built it just to study atoms

by

There are machines that are built to observe what we already know and others to push matter to places where we barely know what we are going to find. In Huizhou, China has raised a complex whose beam line runs two kilometers and which is capable of accelerating heavy ions and direct them against atomic nuclei with an intensity that is difficult to achieve in other laboratories. It looks like an infrastructure designed for nuclear physicists. That description, however, falls short. The decisive step came on Tuesday, July 21. The High Intensity Heavy Ion Accelerator Facility, known as HIAF, passed the acceptance reviewn that certifies the fulfillment of its construction objectives and entered into test operation for scientific research. The project, built by Institute of Modern Physics of the Chinese Academy of Scienceshad begun to rise in December 2018 and produced its first beam in October 2025. The work is finished; Now is the time to do science. Now, the HIAF does not function as a single track through which particles enter and exit at full speed. Its architecture distributes the work between three major systems: a superconducting linear accelerator that delivers intense beams in continuous or pulsed mode, a synchrotron that accumulates the ions and takes them to higher energies, and a storage ring that allows them to be preserved while researchers carry out precision measurements. It all starts with ions, atoms from which one or more electrons have been removed and which can be accelerated and guided by electric and magnetic fields. Those in charge of the project claim that it is the first advanced heavy ion research facility that brings together these three technologies. A facility designed to study much more than atoms We still don’t fully understand how many of the heavy elements we find around us formed in the universe. To reconstruct that history, researchers will use the HIAF to produce unstable nuclei, measure how they behave and explore the frontiers at which nuclear matter stops holding together. That work also opens the door to trying to create new superheavy elements. The goal is not only to expand what we know about the periodic table, but to better understand the cosmic processes that ended up forming elements such as uranium. The same beams that are used to investigate nuclei can also be used to subject chips, electronic components and materials to intense radiation. In space, the impact of an energetic particle can corrupt data, crash a circuit, or cause permanent damage to a component; Reproducing part of these effects on land allows them to be studied before incorporating the technology into a satellite or ship. The HIAF was also born with that function: to offer a controlled environment in which to check what resists and what fails when a device must operate under the radiation of space. When these ions reach tissues or materials, the experiment changes scale, but not principle. Its energy can be used in the medical sector and, at the same time, to accelerate the accumulation of radiation damage in materials intended for advanced nuclear systems or to future fusion environments. We are not facing a hospital or a reactor intended to produce electricity. What the HIAF offers is something prior and decisive: a place to verify, under controlled parameters, what works, what degrades and what needs to be redesigned. The start-up phase has already left some figures with which to measure its capacity. According to those responsible for the projectthe team completed beam commissioning along the two-kilometer line in 16 hours, a record among comparable facilities. It also maintains that the pulsed intensities of the oxygen and bismuth beams exceeded previous international references by factors of three and 7.5, respectively. During the tests, work has already been carried out on nuclear masses, production of radioactive beams, nuclear structure and irradiation of materials. These are initial results: the HIAF is still being tested and must now demonstrate to what extent it can transform that capacity into real discoveries and applications. Images | Institute of Modern Physics of the Chinese Academy of Sciences In Xataka | In the 70s, a Nobel Prize winner in physics mathematically solved the tourist’s great dilemma: which restaurant to choose

China has just launched three giant carbon fiber lines: bad news for the US and Japan

by

China has taken a new step in its race to dominate the advanced materials sector. And the state giant CNBM (China National Building Material Group) has launched three new high-performance carbon fiber production lines at its Lianyungang plant in Jiangsu province, operated by its subsidiary Zhongfu Shenying Carbon Fiber. The new facilities cover the three main categories of the material (general use, high resistance and high modulus) and add to the advances that the country has been announcing in recent months in this field. Why is it important. High-performance carbon fiber has been around for decades being a strategic material controlled by very few actors. Japan and the United States have historically dominated its production, and for years China has had restricted access to the most advanced variants through international export control agreements. With these three lines, Beijing reinforces its capacity to produce at home a material that is key for sectors such as aeronautics, energy, robotics or consumer electronics. In detail. The three lines put into operation are, according to has reported the Chinese state agency Xinhua, the largest in the world in their respective categories. The first is a high-capacity line for 5,000 tons per year, which uses the subsidiary’s own spinning technology and whose material is mainly used for wind energy and electric vehicles. The second, with a capacity for 1,000 tons per year, produces T1100 grade fiber, designed for aerospace applications that require very light structures with high load capacity. And the third, of 600 tons per year, manufactures SYM40 grade high modulus fiber, aimed at precision applications in the seabed, deep mining, space or competitive sports equipment, according to they collect from Global Times. A relevant fact about this announcement is that, depending on the mediumthe new lines have achieved a localization rate of more than 95% in their main equipment. That is, almost all of the machinery used to manufacture the fiber is domestically manufactured, which confirms that China not only produces the material, but also much of the technology necessary to do so on a large scale. The context. This announcement comes in the midst of a technological pulse with the United States and Japan. According to South China Morning PostWashington and Tokyo have recently tightened controls on the export of equipment and technology to manufacture high-strength carbon fiber, precisely to make it difficult for Chinese companies to produce it. Ma Jie, an analyst at Guoyuan Securities, explained in a report that these restrictions have pushed the Chinese industry to pursue technological self-sufficiency and control its entire supply chain, and recalled that the sector has gone from basic T300 fiber to today reaching standards that are among the most advanced in the world. Furthermore, just as we counted A few months ago, CNBM had presented at the JEC World fair in Paris the world’s first mass production of T1200 fiber, the strongest variant of the scale, with more than 8 gigapascals of tensile strength. Added to this news, at the beginning of June, was the announcement by Shanghai Petrochemical, a subsidiary of Sinopec, about the large-scale production of T1000 fiber using its own process. Supply. Zhou Yuxian, president of CNBM, pointed out Xinhua that these new lines help solve two structural problems in the sector: the shortage of high-end carbon fiber products and the lack of a complete offer that covers all categories of the material. The manager added that the production “will reinforce the security and resilience of the Chinese supply of strategic materials.” It is worth remembering that, until now, the world market has been dominated by a handful of manufacturers. Three Japanese companies (Toray Industries, Teijin and Mitsubishi Chemical) together controlled just over half of the global market in 2025, according to a report from Nikkei Asia. The American Hexcel is another of the major players in the sector, especially in the aerospace and military fields. morenda. China’s push for carbon fiber responds to a demand that continues to grow. And according tothe report According to Guoyuan Securities, global demand for carbon fiber exceeded 220,000 tons in 2025, with a growth of 43.8% compared to the previous year, while in China this growth was even higher, 57%, driven by the manufacturing of the C919 commercial aircraft, humanoid robots and low-altitude vehicles. Cover image | CGTN In Xataka | “It’s not cheating, it’s engineering”: this is how China avoids 2,000 ASML patents with a laser trick

China wants the world to pay for lithium at its price. And it’s not that easy

by

China has taken a definitive step towards the internationalization of its raw materials markets: this month has opened foreign investors its lithium futures contract on the Guangzhou Futures Exchange (GFEX). This measure comes after similar openings in iron ore and oil, and is part of the Chinese Government’s ambition to strengthen your influence on global prices of raw materials and expand the international use of the renminbi (RMB). Futures are contracts that set today the price at which a commodity will be bought or sold at a future date, regardless of what its market price is at that moment. This contract is not an anecdotal movement. In the last year, the GFEX has traded 120 million lots of lithium carbonate, an essential material in lithium-iron-phosphate batteries (LFP) that equip a good part of current electric vehicles. China processes around 60% of the world’s lithium and it is by far the largest battery market on the planet. In this context, its physical weight in the supply chain is undeniable. However, converting that industrial weight into real financial power is much more complicated. Chinese capital controls mean that the vast majority of futures trading, despite its enormous liquidity, is domestic in nature. It is primarily domestic speculators and not global commercial participants who set these prices. A recurring speculative fever Volatility is the first obstacle. Last year the closure of a CATL mine was enough to unleash a real speculative fever in the GFEX contract. In fact, on July 6, 2025, the daily volume fell to 174,787 lots, just 35% of the annual average, an example of how quickly interest deflates once the initial frenzy passes. This stock exchange has had to intervene on several occasions to stop similar episodes. In July 2025 he imposed a daily limit of 3,000 lots in new positions for those who were not members of a futures society. In November he went further: he tightened commissions and position limits, and prices plummeted immediately. The clearest precedent dates back to 2016, when Chinese retail money poured so heavily into steel rod futures that the volume traded in a single day exceeded the negotiated total on the Shanghai Stock Exchange. China opened its crude oil contract on the Shanghai International Energy Exchange using this same rhetoric in 2018 Another important note: adding to this regulatory volatility is the problem of financial plumbing. Foreign operators can now deposit margins in dollars, but with a 5% discountso only $95 out of every $100 counts as collateral. All trading and settlement is still done in RMB, and repatriating those profits is not easy. China opened its crude oil contract in the Shanghai International Energy Exchange (INE, for its acronym in English) using this same rhetoric in 2018. And although achieved considerable foreign participationcompanies still have to simultaneously manage price risk, RMB/USD exchange rate risk and capital repatriation risk; three exposures that on the New York Mercantile Exchange (NYMEX) or the Intercontinental Exchange (ICE) would be reduced to one. The result is that as long as these futures remain subject to political intervention and capital controls, they will predictably continue to function as an essential reference for the market, but not as the authentic benchmark against which the rest of the world can manage its risk. Image | freepik More information | Volt Insight In Xataka | ASML has had an exclusive monopoly on the very complex photolithography machines for years. China is close to breaking it

China has a weak point under the sea. Your new anti-submarine plane points right there

by

Above the surface, a fleet of ships can be counted, photographed and tracked in almost real time. Under water, however, the rules change: the target disappears from sight, the sound is deformed and each contact can be a valid clue or a false alarm. For this reason, anti-submarine warfare remains one of the most demanding tasks of any Navy. China knows this well and is expanding its response to monitor a space where the size of its forces alone is not enough. A plane enters the scene. The response that China has begun to show from the air is an improved variant of its Y-9 anti-submarine aircraft, now seen for the first time in maneuvers. The aircraft had appeared publicly at the parade on September 3, 2025, but the new official images allow external modifications to be seen in its detection systems. According to military analyst Zhang Junshe, quoted by Global Timesits incorporation into these exercises indicates the beginning of operational training and the acquisition of initial search and attack capabilities. The difficult part is below. An official evaluation by the US Department of Defense allows us to place the progress of the Y-9 within a gap that China would still be trying to close. Its 2024 report stated that the Chinese Navy was improving its anti-submarine means, but still lacked a robust deep-water capability. In March 2026, scholar Andrew Erickson described an unequal situation, although he made it clear that his conclusions were personal. According to the expert, China is developing a layered detection architecture within the first island chain, but its capabilities remain more limited beyond it. There, the smallest number of deployed sensors, limited maritime patrol aviation and logistical restrictions weigh in. That does not make these evaluations neutral diagnoses: they come from the United States, the power that tries to preserve its submarine advantage against China. The next obstacle. Zhang points out that the South China Sea reaches depths of more than 1,000 meters in some areas and has more complex conditions than the East China and Yellow Seas, which makes it difficult to locate submarines. In addition, temperature, salinity, currents and underwater relief modify the propagation of sound. This requires knowing the environment precisely to correctly interpret the information collected by the acoustic systems. A network, not a single plane. Beijing is not entrusting this entire task to Y-9. A Reuters investigation published earlier this year documented an extensive operation supported by oceanographic vessels, hundreds of sensors, buoys and arrays installed underwater. Separately, Erickson describes a Chinese architecture that integrates space, air, coastal, surface and underwater assets, as well as unmanned vehicles. The new plane represents the aerial and mobile component of that broader effort to understand and monitor what happens under the sea. The most visible change appears under the nose. According to Zhang, the images show what appears to be a new active phased array radar, with broader coverage, a longer sweep and a greater detection distance than the version presented in 2025. That does not mean directly finding a submerged submarine from the air, but it does mean expanding maritime surveillance and delimiting areas of interest. The more the search area is reduced, the easier it is to then target systems specifically designed to locate underwater targets. Listen and confirm. Once the area is delimited, the Y-9 can deploy up to a hundred sonobuoys of different types, housed on both sides of the fuselage. These form a listening network capable of covering tens of kilometers and collecting signals from submerged targets. In the tail, what appears to be a new magnetic anomaly detector looks for small alterations associated with a large metallic mass. Its smaller size would reduce the magnetic interference of the structure itself, facilitate maintenance and extend the detection distance. Beyond what it detects. The Y-9 aims to improve compared to the Y-8 in range, combat radius, takeoff weight and autonomy. In the maneuvers, in addition, the crews operated without a script or fixed tactics, with electromagnetic interference, simulated breakdowns and coordination with other units, within training aimed at improving their search and attack capacity day and night. None of this makes the ocean transparent, but it does show where Beijing is concentrating its efforts. Images | PLA In Xataka | In 1989, the Soviet submarine Komsomolets sank with one reactor and two nuclear warheads. To this day it continues to radiate

China is going to do everything it can to win the AI ​​race. Even if that means penalizing Huawei and Alibaba

by

The Chinese Ministry of Commerce has begun to consult its main technology companies on a package of export restrictions that, paradoxically, points inward: towards the ecosystem of artificial intelligence (AI) that Beijing has been trying for years shield in front of Washington. As advanced Financial Timesthe proposal contemplates prohibiting Chinese chip designers from using foreign semiconductor companies, such as TSMC or Samsung, to manufacture their integrated circuits. This is not an isolated measure. And this same regulatory package would also include limits on the export of advanced AI models, restrictions on the training data that companies can take out of the country and stricter controls on foreign acquisitions of technology companies considered strategic. The Chinese Government ultimately aspires to retain both the hardware and software of the AI ​​race within its borders; as well as the chips, models and data with which they are trained. A blow to their own companies The most impressive part of this proposal is the one that affects the semiconductor manufacturing. Huawei, Alibaba or ByteDance design chips that today depend, to varying degrees, on semiconductor factories such as TSMC (Taiwan) or Samsung (South Korea) for their final production. Prohibiting that access would leave these companies tied to domestic manufacturing capacity, still behind in leading nodes, or would force SMIC and other Chinese IC factories to take on demand for which they are not yet prepared. The regulatory package would also include tightening rules on foreign acquisitions of strategic Chinese technology. The Ministry of Commerce has also conveyed to Alibaba, ByteDance and Zhipu its intention to limit the output of training data outside the country, as well as to restrict the downloading of their model weights by foreign users. This measure would mark a turn regarding the strategy followed until now by companies like DeepSeek, which have opted to openly publish the weights of their models as part of their low-cost offensive against Western AI. Another note: the regulatory package would also include a tightening of the rules on foreign acquisitions of strategic Chinese technology, especially in the field of agentic AI. This review seeks to close what the Chinese government interprets as a legal loophole, the same one that allowed Meta to acquire the emerging company Manus for $2 billion before the Chinese authorities ordered to undo the operation. The restrictions, if confirmed, would foreseeably be incorporated into the next review of the Chinese catalog of prohibited or restricted technologies for export, the same instrument that already regulates rare earths and lithium. For now all the proposals are in the consultation phase and none of the companies mentioned, nor the Ministry of Commerce itself, has confirmed its final scope. The result would be a regulatory paradox: China restricting its own companies’ access to the most advanced chip manufacturing in the name of technological sovereignty that, along the way, could slow them down. Image | TSMC More information | Tom’s Hardware In Xataka | TSMC raises its bet in the US: there are already 265,000 million dollars for 2 nm

In China, taking a taxi is already cheaper than using your own vehicle for one reason: electric cars

by

China is experiencing a peculiar situation: on its streets, it is now cheaper to get around by taxi than by private car. According to the latest data from the Ministry of Transport collected by the Reuters agency, lTaxi trips grew by 6% since the outbreak of the Hormuz crisisa phenomenon that coincides with the increase in fuel prices and the enormous electrification of urban fleets. This percentage translates into more than 3,050 million taxi trips in the month of May alone, demonstrating that the international geopolitical situation has a great impact on people’s daily lives. The context. The war in Iran and the situation in the Strait of Hormuz have caused Chinese crude oil purchases to reach historic lows. Last June, oil imports had a collapse of 41.3% year-on-yearwhich placed them at their lowest level in a decade since October 2016, with only 29.27 million tons (about 7.12 million barrels per day). Added to this is also the low production that the national refineries are having due to a fairly low internal demand. All of this has created the perfect cocktail to generate a shortage situation that has skyrocketed prices, among other things, of gasoline. The secret is in the electrical. This is where the taxi comes into play. Although fuel prices rise every day, the rates for this type of transport in China are becoming cheaper. This is because, as Reuters indicates, almost half of the country’s taxi fleet is electric (a condition that in cities covers almost 100% of the car supply) and, therefore, is not as dependent on volatile oil prices, making it an effective shield from China. in front of shock of Hormuz. But not only that. In recent times, the increasingly easier access to an electric vehicle and the slowdown of the Asian country’s economy has pushed many Chinese citizens to take the leap to be independent professionals and look for work in the mobility sector as drivers with their own car. According to Reuters, companies like DiDi, the main Chinese MaaS application, have grown thanks to this phenomenon. In this case, the company has incorporated more than 2 millions of hybrid or electric vehicles in the last year, which represents a total of more than 8 million fossil fuel-free models in its fleet. Thus, there is more supply and more competition and the rates, consequently, are much cheaper. The user wins. This new trend has already flooded the Asian country’s social networks and, in addition to the lower rates and savings on gasoline, many Chinese citizens point out the comfort of the trip and the ease of not having to look for parking as other benefits of opting for a taxi, as expressed by Yang, owner of a gasoline car, to Reuters. However, not everything is risk-free, since these economics professions in demand -such as those of a professional driver, riders and couriers – are in high demand and could end up saturating the market. A change forever? Since the crisis, it is no secret that the world is getting used to living with a 9% less oil which previously passed through Hormuz. Like many other countries, China is now seeking to be less vulnerable to international volatility and rely less on fossil fuels.. It does so in a very clear commitment to leading decarbonization by dominating the global production of solar panels, batteries and, of course, electric vehicles. The big question is whether this change responds solely to the rise in oil prices or whether it will end up consolidating a change in habits in the mentality of Chinese citizens even when the energy market stabilizes again. Cover image | Teresa Wang In Xataka | China has so many electric cars running on its streets that it is going to use them to generate energy for homes

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.