problem

The Chuwi Unibook is the $450 Windows laptop that aims to take down the MacBook Neo. The problem is not the specifications

by

The Chinese manufacturer Chuwi has given the surprise with the presentation of its Chuwi Unibook, a mid-range laptop that surprises with its price of $449 and that has undoubtedly been created to compete with the new rival to beat: the MacBook Neo from Apple. The truth is that on paper the proposal seems really attractive, but the problem is precisely that: that this computer, like all those that will soon appear based on Windows with similar specifications, will have to comply with what is important. The user experience will be everything. The MacBook Neo still has no response. The PC industry was used to not having too many concerns in the mid-range. The manufacturers had accommodated themselves and proposed proposals without much ambition, modest but functional. Then came the MacBook Neo from Apple and revolutionized the sector: For the first time it was possible to access the Cupertino laptop ecosystem and its experience for a much more affordable price. There are sacrifices to the MacBook Neo, of course, but the device’s appeal is evident to many users. Apple has the A18 Pro, Intel has Wildcat Lake. The striking thing about the MacBook Neo is that Apple demonstrated that the iPhone chip was more than enough for a mid-range laptop. To compete with it, Intel has launched a new family of low-cost processors called Wildcat Lake. These chips, made with Intel 18A photolithography, are promising, and according to some benchmarks one of their variants It is 21% more powerful than the Apple A18 Pro of the MacBook Neo. The spec sheet rocks. If we look at the pure specifications of the Chuwi Unibook, the difference is notable. The equipment is not only cheaper, but it surpasses the Apple model in almost everything. For example, it has a theoretically more powerful processor, keyboard backlighting, better connectivity and more battery. The sacrifices required by the MacBook Neo are fewer sacrifices in this equipment. On paper, the Chuwi Unibook is really promising. On paper. Source: VideoCardz Project Firefly. Intel’s Chinese division recently announced this initiative. With it, they hope to help manufacturers reduce manufacturing complexity by offering reference designs that reduce production costs. Intel has already done things like this in the past (I’m sure many of you will remember both the Centrino branding and its Ultrabook program), and the idea here is precisely to provide certain tools to manufacturers to develop more competitive models in a market. shaken by the Apple model. Manufacturers wait their turn. The launch of Intel processors from the Wildcat Lake family has caused several manufacturers to begin announcing laptops based on these chips. Lenovo is already preparing some models IdeaPad Slimand so much Asus as HP They also prepare their plays. The Chuwi Unibook seems to be just another variant of those proposals, and in all of them the specifications, although modest, seem to surpass those of the MacBook Neo. Lots of advertising, little real product. Almost all major manufacturers have shown their intention to develop mid-range laptops that compete with the MacBook Neo in that price range. The announcements have been varied, but none of them have communicated the price or availability date of these devices, probably because everyone is waiting to see how the memory crisis evolves. It is reasonable to think that the imminent Computex fair is the perfect occasion to definitively present all these proposals. But. The problem with the Chuwi Unibook, like that of other manufacturers waiting their turn, is not the specifications. The problem will be the benefits and above all the real experience that these teams offer. Windows PC manufacturers have not done well with cutting features in the past, and if that experience is not good we could witness a new phenomenon like netbooks: affordable equipment, but too limited and that ended up condemned to oblivion. In Xataka | “We arrived too soon, but we were right”: The MacBook Neo is everything Microsoft dreamed of with the disastrous Windows 8

Researchers solve a problem that has been stuck for decades

by

A team from Monash University in Australia has developed an ultrathin membrane able to operate hydrogen fuel cells at 250 °C and, most surprisingly, without the need for water. This is a wall in which technology has been crossing for a long time and the discovery has been published in the journal Science Advances. Below these lines we tell you all the details. Why is it important. Hydrogen cells are one of the great promises to decarbonize transportation, heavy industry and sectors where batteries fall short. They only emit water and heat, they recharge quickly and offer autonomy comparable to gasoline. The problem is that current membranes, such as those based on Nafion (a synthetic resin), they need to be permanently hydrated so that the protons can circulate. And that forces us to operate below 80-100 °C, because at higher temperatures the water evaporates and the entire system collapses. In detail. The team, led by researchers Huanting Wang and Kaiqiang He, has built atomic-thick nanosheets made of graphene and boron nitride. Between those layers they have introduced phosphoric acid in a state that researchers call nanoconfined, where the acid is trapped in tiny spaces from which it cannot escape or evaporate, even at 250 ° C. The result It is a membrane of just 50 micrometers, named GBP, that acts as a dry highway through which protons move at high speed without depending on a single drop of water. How it works. Wang, professor in the Department of Chemical and Biological Engineering at Monash, account that “by combining proton-conducting nanosheets with nanoconfined phosphoric acid, we have developed a membrane that maintains rapid proton transport without water.” The trick is in a mechanism that the authors define as synergistic, in which protons directly pass through the hexagonal rings of graphene and boron nitride and, at the same time, jump along the network of hydrogen bonds that forms the acid confined between layers. On the other hand, He adds that this combination is what gives the membrane high conductivity and stability in dry and high temperature conditions. The figures. In laboratory tests GBP achieved a proton conductivity of 166 mS cm⁻¹ at 250 °C and a power density of 1,011 mW cm⁻² in a hydrogen-oxygen stack, well above industry reference membranes. In addition, the team kept it running for 150 hours straight at that temperature without signs of degradation. Between the lines. Working at 250°C is a game-changer on several fronts. One: The water management and humidification systems are eliminated, which in current hydrogen cars are heavy, bulky and expensive. Two: at that temperature the platinum catalyst tolerates impurities better such as carbon monoxide, which opens the door to using less pure hydrogen and, therefore, cheaper to produce. Three– Cooling the system becomes much easier, allowing for smaller radiators and lighter vehicles. Beyond the car. Although we usually focus on the hydrogen cars When we talk about this type of technology, the truth is that the potential applications go much further. GBP was also tested in direct methanol cells and performed at 502 mW cm⁻² with 16 M concentrated methanol at 250 °C. This suggests that it could be used for portable systems where hydrogen is difficult to store. In addition, the authors point to uses in data centersplanes, trains, factories and hospitals as energy backup, and other electrochemical processes such as the separation of water molecules, the reduction of carbon dioxide or the synthesis of ammonia. And now what. The next step is the usual one. And when a laboratory announces an advance like this, we have to wait until it ends up coming to fruition and its commercialization on an industrial scale is viable. If they succeed, the combination of cheaper batteries, less pure hydrogen and simpler systems could accelerate the arrival of this technology in sectors where electrification with batteries does not quite fit. Cover image | CARMAN and Monash University In Xataka | The world depends on gas to produce food. Paraguay believes it has the definitive solution thanks to the Itaipú dam

In China they want humanoid robots to do household chores. The problem is that a house is not a factory

by

For years we have seen humanoid robots do somersaults, danceppractice martial arts or move through factories with increasingly striking capabilities. The next step seems almost natural: taking them home to do the laundry, prepare a bed or support elder care. The problem is that this transition is not as direct as it seems. A factory is designed to reduce uncertainty; A home, on the other hand, is full of small exceptions. And for a robot, those exceptions can be exactly the difference between a flashy demo and a useful product. The concept. SCMP account That GigaAI has introduced the SeeLight S1 as the country’s first general-purpose home humanoid robot model, developed in collaboration with the Hubei Humanoid Robot Innovation Center and the Hubei Humanoid Robotics Industry Alliance. In images released by the company, he appears performing very recognizable tasks: cutting vegetables, frying eggs, loading a washing machine, hanging clothes, making a bed or opening curtains. The company also plans to test it for free in homes in Wuhan in the first half of 2027. A house is not an assembly line. That is the fundamental difference. In a factory, the robot can work with known references, pieces always placed in the same way and movements that are repeated thousands of times with very few variations. In a home, on the other hand, nothing guarantees that the shirt is where it was yesterday, that the chair has not moved or that a pet does not cross in front of it just when the robot is trying to complete a task. Much movement, little understanding. Xinhua itself collects an idea that helps cool down the epic of the demonstrations and that does not only affect China, but humanoid robotics in general: humanoids have greatly improved in their “cerebellum”, the part linked to control and coordination, but they still have major problems in their “brain”. In other words, they can execute complex movements, but it is difficult for them to understand what a scene means and what function each object has within it. Home is also a data problem. Now, for these systems to work better in real homes, they need to learn from real homes, but the home is precisely one of the places where it is least easy to collect data. We are not just talking about room maps, but about objects, forces, angles, routines and physical decisions that are difficult to simulate. Advances and challenges. According to NSFCthe country expected to exceed 10,000 humanoid units sold in 2025, with a year-on-year growth of 125%, and there were already pilots in industrial manufacturing, delivery, catering and services. The important nuance is that none of this automatically turns this industrial career into a successful deployment within homes: the sector itself locates the path prudently, first industry, then logistics and commercial uses, and only later the home. A future easy to imagine, difficult to materialize. The difficult part is demonstrating that this can be done usefully, safely, and at reasonable cost outside of a prepared demonstration. There is the real border. China and other countries around the world can accelerate prototypes, pilots and production, but a home does not forgive clumsiness in the same way as a controlled stage. To get home, the robot will not have to understand human life better. Images | GigaAI In Xataka | In China there are already “schools” for robots. Its objective is the same as schools for humans: to teach them to work

The problem is that anyone could look

by

Many homes with small children have installed surveillance cameras or baby monitors. They allow you to monitor the child while it sleeps and, if it cries, you can go to see what is happening in an instant. All advantages, except when the camera in question has a security flaw that gives access to anyone who can look. what has happened. Security expert and independent researcher Sammy Azdoufal discovered that Meari brand cameras (including electronic peepholes and baby monitors) were completely unprotected. By simply analyzing the Android app, Azdoufal was able to extract a unique key that gave him access to more than a million cameras spread across 118 countries. All your research is detailed in its GitHub repository. What is Meari?. Surely the brand Meari Technologies does not sound familiar to you, it is normal because it is a white label manufacturer, that is, it manufactures for other brands. As explained in The VergeMeari cameras are sold on Amazon under various brands, such as Arenti, Anran, Boifun, ieGeek, Wyze, Petcube, COCOCAM, PetTec, SV3C, Joystek, Luvion and Vimar. Doing a search on Amazon Spain we found cameras from many of these brands, some of them with thousands of positive reviews. Non-existent security. The problem was not in a specific model, but in the entire architecture; Many of these brands shared servers and sometimes even credentials, so the system did not have any type of isolation: any of these brands could access the cameras of another. Additionally, the MQTT system (a machine-to-machine messaging protocol that ran on the EMQX IoT platform) did not have adequate protections, allowing real-time viewing of what was happening in thousands of homes. He also discovered that many cameras were still using default passwords such as “admin” or “public” and what is even worse, the alert images that these cameras save (for example when they detect movement) were stored on Alibaba servers without any type of protection, accessible simply through a URL. And this was not all, he also found an unprotected internal server where he found passwords for Meari and the list of 678 employees, including their emails and telephone numbers. He didn’t need to hack anything, just know where he had to look. The answer. According to Azdoufal, they did not take him seriously until they saw that their own employees’ data was being leaked, then they began to respond to his emails and solved the main failure, cutting off access to their cameras. In a statement sent to The Verge, the company admits that “Under certain technical conditions, attackers can intercept all messages transmitted through the EMQX IoT platform without user authorization.” However, it did not answer key questions such as how many camera models were affected, whether different brands have warned users or whether the vulnerability had been previously exploited. Tensions. Azdoufal was paid more than $24,000 for finding the bugs, but it was after several weeks of negotiation in which Meari did some pretty shady things. According to the researcher, the company sent him messages with veiled threats such as that the access he had made to their servers was illegal, that they were ready to “protect his interests” and that they knew his address. The company also tried to pretend that they knew about the bugs before, publishing security bulletins with altered dates. What to do if you have a Meari camera. According to the researcher, Meari manufactures for more than 300 brands and on the official website we have not found any official list, so it is difficult to know which brands are affected. If you suspect that your camera is one of these, Azdoufal recommends unplugging it whenever you are not using it, because the problem is in the cloud and is not something you can fix yourself. Also, please note that some images may still be accessible, and if you live in the EU you can lodge a complaint with your data protection authority. Image | Xataka with Gemini In Xataka | NASA has had its space systems exposed to hackers for three years: an AI discovered it in four days

A mathematical problem had been resisting experts for more than 80 years. An AI has surpassed them all

by

In 1946 the Hungarian mathematician Paul Erdős asked a seemingly very simple question: if you place n points in the plane, how many pairs of points can be exactly at a distance 1 from each other? This dilemma is known as unit distance problem in the planeand has maintained many mathematicians who research in the field of geometry, immersed in its resolution for no less than eighty years. The classic strategy proposed by many of them to try to solve it was to resort to a square grid. They soon realized that the number of pairs at unit distance grows at least as n to the power of (1 + C/loglog(n)), where C is a positive constant that quantifies how much a particular construction can be better than a basic square grid. It’s a complicated idea, it’s true, but we can try to approach it in a slightly more intuitive way. A standard square grid produces approximately 2n pairs of points at unit distance. If we rescale it in an ingenious way by choosing the scale factor as a number that has many divisors (in number theory this property is known as a number with many small prime factors), you get more pairs of points to fall exactly at distance 1. The value of C measures precisely the efficiency of that choice. This is the key. An AI from OpenAI has achieved the first major breakthrough in 80 years As we are seeing, the question Erdős asked is very easy to state, but extraordinarily difficult to resolve. If we develop the classical approach a little further we will realize that since loglog(n) grows very slowly, the exponent approaches 0. This means that the square grid grows only slightly faster than n, but not enough to exceed n at a fixed rate. This milestone was achieved by a general-purpose inference model that OpenAI was testing internally. This is why for decades mathematicians predicted that the upper bound would be approximately n^(1+o(1)), that is, just slightly larger than n. Now we know that they were wrong, and the person who refuted this conjecture was not a particularly skilled current mathematician; this milestone has pointed it out a general purpose inference model which OpenAI was testing internally. and not one artificial intelligence (AI) specialized in mathematics. This model has provided an infinite family of examples that produce polynomial improvement. In fact, he has shown that it is possible to construct configurations of points with at least n^(1+δ) pairs at unit distance, where δ is a fixed value greater than 0 that does not disappear as n grows. When the AI ​​delivered this result, OpenAI researchers asked a group of Princeton mathematicians to review it. And his conclusion was blunt. The AI ​​was right. This is the first progress on the lower bound of the problem posed by Erdős in 80 years. And, curiously, the OpenAI model has achieved this by using advanced engineering tools. algebraic number theory for an apparently elementary geometry problem. Several renowned mathematicians, such as Fields Medal winner Tim Gowers or number theory expert Arul Shankar, have declared that the result that AI has delivered is an extraordinary achievement that could provide mathematicians with a bridge to explore other problems in the future. Image | Jeswin Thomas More information | OpenAI In Xataka | These two problems have baffled mathematicians for decades. A genius has solved them with a stroke of the pen

Giving seven times more vitamin D during pregnancy improves children’s memory at 10 years old. The problem is in the fine print

by

During pregnancy, the recommendations of supplementation They are an area where science advances with lead feet, since the most important thing is always to guarantee safety. One of these supplements that is heard the most is vitamin Dtraditionally known for its role in calcium absorption and bone health, but which has been in the spotlight for years for its possible impact on neurodevelopment. A new study of Danish origin has put its objective on this statement to be able to clarify what happens when a mother supplements with vitamin D during pregnancy. Through its publication in JAMAtells how, to achieve good results, almost 500 children were analyzed for several years until finally being able to see if they had cognitive improvement during their childhood. What were they based on? To understand this discovery we have to go back in time to a randomized clinical trial titled as COPSAC2010whose initial results were published in 2016. This trial sought to evaluate whether vitamin D prevented the risk of suffering from asthma or persistent wheezing in babies, and to verify this the researchers divided the mothers into two groups from the 24th week of gestation: One group would receive the standard recommended dose of vitamin D of 400 IU per day. The other group had a “megadose” of vitamin D of 2,800 IU daily. The discovery. Taking advantage of this valuable group of 498 children, the research team decided to get more out of it, since when these children reached 10 years of age they were subjected to rigorous cognitive tests to see if the fact of having given vitamin D to their mother during pregnancy had left its mark on their brain. In this way, two objectives were covered with a single investigation. Here the results revealed that children in the high supplementation group showed a modest but significant improvement in verbal and visual memory compared to the children of mothers who took the standard dose of vitamin D. Although something important to note is that it puts to rest any idea that this supplementation is a machine to “create geniuses”, because there were no differences in IQ and they only saw that the ability to retain information was improved. The small print. Given such a finding, it is tempting to think that all pregnant women should multiply their vitamin D intake to give their children an advantage over others. But here we must pay attention to different problems, such as that the original trial was designed to measure respiratory problems and not neurological development. This means that drawing conclusions from here reduces the statistical robustness of the discovery. But this is not the only problem, since we have seen that the effect is “modest” without seeming to give children a great advantage. And furthermore, the study is based on women who already had normal vitamin D levels before the study, so it is not clear how this dose would act in populations that truly have some type of chronic deficiency of the vitamin. Will there be changes? At the moment, these studies do not justify the need to recommend that all pregnant women supplement their diet with vitamin D, as is the case with other supplements such as folic acid. The real value of this research is not to give us an immediate new prescription, but to open the door to future clinical trials specifically designed to unravel how what happens in the womb continues to shape our brains a decade later. Images | amylla battani In Xataka | We have been sending pregnant women to bed for decades as a precaution. Science has just proven that it is a big mistake

China finally has a competitive desktop processor. Its problem is that it is six years behind Intel

by

China has your own alternative to processors for PCs, servers and data centers made by Intel, AMD and other companies. Loongson is one of the few Chinese companies that can manufacture advanced microprocessors. We have been following it for several years because in the current climate of geopolitical tension it has acquired more relevance than ever, and there is no doubt that its cruising speed is high. At the end of December 2022 this company launched its CPU 3A5000 32 corea general-purpose processor with LoongArch microarchitecture implemented by this company on the MIPS architecture. And in February 2024 it presented its LS3C6000 server processor, a CPU with DragonChain technology that could be scaled up to 64 cores. Its latest milestone is not the presentation of a new chip. The reason why we have decided to talk to you again about this Chinese company is that just a week ago it confirmed that it has distributed more than one million units of its flagship desktop processor, which represents a milestone in China’s efforts to build a self-sufficient semiconductor industry. The 3A6000 CPU has been designed and manufactured entirely in China Loongson implements its processors on the MIPS architecture, but the microarchitecture of these chips has been expressly designed by engineers at the Chinese Academy of Sciences. By not using x86-64 or ARM architectures, this company has been able to continue refining its designs without being conditioned by US sanctions. Be that as it may, Loongson is dedicated to the design of microprocessors, but does not have the capacity to manufacture them itself. China recently had no alternative to US-made CPUs SMIC takes care of this (Semiconductor Manufacturing International Corp), what is the largest semiconductor manufacturer from China, in the same way that TSMC produces the integrated circuits designed by AMD, Apple, NVIDIA or Qualcomm. According to the publication Fast Technology, the third generation of Loongson chipsto which the 3A6600, 3B6600 and 3C6600 CPUs belong, has a performance comparable to that of the 12th Intel Core and 13th generation. Curiously, according to Fast Technology, the 3B6600 model in particular is the one that rivals these Intel CPUs and comparable AMD proposals. In fact, according to SCMP Loongson herself has acknowledged that the performance of her desktop processors is comparable to that of Intel chips launched around 2020. Six years is a long time in this sector, but it is important that we do not overlook that China recently did not have any alternative to US-made CPUs. This achievement by Loongson is part of Beijing’s effort to channel resources to reduce China’s dependence on foreign semiconductor technology. However, this strategy has been accelerated in response to restrictive export controls Americans who limit China’s access to advanced chips, integrated circuit design software and next-generation semiconductor manufacturing services. It will be interesting to see if Loongson finally catches up with Intel and AMD. Image | TSMC More information | SCMP In Xataka | China takes off in quantum computers: it already has the first dual-core and 200 qubits on the planet ready

Lead has its days numbered in hunting. The problem is that no one really knows how to replace it.

by

The practice of hunting is a ‘hobby’ that has been much discussed among different sectors of society in recent years, but beyond its ethical aspects, the European Agency for Chemical Substances and Mixtures has decided to intervene by pose the ban on lead in hunting ammunition. And this has raised a great debate between the hunting sector and environmentalists and researchers that require immediate measures, although at the moment it is in pause. The siege of Europe. This lead fence is not new, but since 2023 the European Union prohibits its use in wetlands for hunting waterfowl, a historic measure to prevent lead poisoning in ducks and other species that ate the pellets at the bottom of the lagoons. But now they want to go one step further, since ECHA has recommended to the European Commission a total veto by prohibiting the marketing and use of cartridges and fishing tackle that contain more than 1% lead. And to soften the blow, the European drafts have proposed transition periods that range between 18 months for large hunting and up to 5 years for small hunting. The role of science. For the Spanish scientific community, these grace periods are a luxury that ecosystems cannot afford. Specifically, 130 researchers from different institutions have signed a manifesto urging the Government to support the European restriction without any type of concession. That is, the ban applies immediately overnight. The arguments they offer focus mainly on the great toxicity that lead generates when it is left abandoned in the countryside, and above all they argue that there is no safe exposure threshold. That is, the only security we can have is when there is zero lead in the environment. Its impact. On the one hand, we have the environmental impact, since tons of lead end up scattered in the countryside every year due to hunting practices, poisoning fauna, especially scavenger birds that consume prey with pellets. On the other hand, we have a public health problem. In this case, there are several reports that exist warning about the nutritional risks of consuming game meat shot with this type of ammunition, recommending that children and pregnant women avoid its consumption due to the neurological data associated with lead. The hunters. Faced with the scientific urgency, they ask to hit the brakes on the application of these measures. Entities such as the Royal Spanish Hunting Federation (RFEC) and the Andalusian Hunting Federation (FAC) they argue that a sudden ban would be a death sentence for the sector and, by extension, for the economy of many rural areas. The problem that arises is that the alternatives to lead are not completely convincing, since, although there are options such as steel or bismuth, the hunting sector denounces that they are not validated at a toxicological level, they are much more expensive and, above all, that there is no large-scale production to cover the demand. Furthermore, the use of steel shot requires in many cases to adapt or change older shotguns, since they can damage the barrels due to the hardness of the material. This is why they ask for at least 10 years to adapt. A political battle. Right now the Government supports aligning itself with the hard line that comes from Europe, but the opposition parties, such as the Popular Party, ask for a fight in Brussels over get those decades of margin and funds for safe, alternative ballistics research. And right now the ball is literally in the court of Brussels and the REACH committee in charge of regulating chemical substances in the EU. Right now the only thing left is to open a space for debate that is not easy at all. Images | freepik In Xataka | Hunting has been printed on the Spanish national ID card for centuries. Now you have a problem: there is no relief

Gemini Intelligence promises to be Google’s AI revolution. The problem is that almost no one will be able to use it.

by

Updates have always been Android’s Achilles heel, but for several years we have seen how manufacturers are pushing to offer up to seven years of updates on your mobiles. It is good news for users and regulators. The problem is that AI threatens to introduce a new form of fragmentation: having an updated mobile phone no longer guarantees access to the most important functions, even if it is high-end. What has happened? A few days ago, during the Android Show, Google announced the new star feature coming to Android: Gemini Intelligence. We are no longer talking about specific functions, but rather about a layer of AI that covers everything, making the mobile phone act autonomously within the system and the apps. It sounds great, what doesn’t sound so good is the list of requirements. Hardware requirements. Google has detailed the minimum requirements for a mobile phone to run Gemini Intelligence and it is quite not encouraging. We are talking about devices with 12 GB of RAM and that mount recent “flagship” processors. These requirements directly leave out the majority of the current vehicle fleet, but also at the current time with the memory crisis raising pricesthe high-end is going to become even more unattainable. The real problem. If the RAM and the processor already leave out many mobile phones, the software requirements are even worse. This is where Google makes the real difference since, to run Gemini Intelligence, compatibility with Gemini Nano V3, the local language model for mobile phones, is necessary. If we look at the current compatibility list, it is no longer that it affects cheap phones, it is that it also leaves out phones like the Samsung Galaxy Z Fold7 which was launched in the summer of 2025 and cost 2,109 euros, or the Xiaomi 17 Ultra which has just been launched for almost 1,500 euros. It is not clear that the list is definitive, since it is possible that there are changes because they allow it to be updated later, but for now the outlook is bleak: The list of devices compatible with Gemini Nano v2 and Gemini Nano v3. Image: Xatakamovil There is still more. The software requirements don’t end here. Google has also put several additional conditions for a device to have Gemini Intelligence. The device must receive at least five years of operating system updates and six years of security patches, in addition to meeting a series of quality requirements regarding stability, failure rate and multimedia, among others. The privilege of AI. “The best of Gemini in our most advanced devices” is the phrase we find in the Gemini Intelligence official websiteso Google already warns us from the beginning. That updates or more advanced functions reach the most expensive phones is something we are used to, but with AI we are seeing the bar rise even higher. Furthermore, it is not a specific function as it was ARCorewe are talking about the central axis of the proposal, a new way of interacting with the mobile that only a small percentage of users will be able to test, including those who have a Google Pixel. Cover image | Google In Xataka | Android 17 news: list with a summary of everything that will arrive in this version of the operating system

Chinese Big Tech can now buy Nvidia GPUs. The problem for Nvidia is that they don’t need it now

by

The United States and China are immersed in a trade and technological war that has caught the line of fire to the AI ​​giant: Nvidia. The situation is that Nvidia must prioritize AI companies from the United States to guarantee the supremacy of this country, but as a company it would be interested in taking a bite out of the giant Chinese market. And the problem is twofold: it has not been able to do so for a long time due to trade vetoes, but now that it seems that it can sell its famous H200 to China, it turns out that China has turned the page. More or less. green light. Nvidia has gone from having a monopoly on AI GPUs in China to have a 0% quota. These are the words of the CEO, Jensen Huang, and the reason is the aforementioned trade restrictions between the powers that prevented Nvidia from selling its most powerful products to the Asian giant. Huang has spent months insisting on Donald Trump’s government to allow them to sell with a very clear logic: China is going to develop its alternatives and what better way to make a profit until then. The situation is gone relaxing at the end of last year and at the beginning of this to get to the point where we are now. According to Reutersthe US Department of Commerce already allows ten Chinese companies and distributors such as Foxconn and Lenovo acquire that long-awaited H200the company’s second most powerful AI chip. Good news for the company. Or they should be if it weren’t for the fact that the Chinese industry is going its own way looking home. Alibaba, ByteDance, JD.com and Tencent are the Chinese giants that can supposedly already buy H200. Up to 75,000 chips each, to be exact. However, it is noted that they have not yet made any shipments. Here there is a mix between very restrictive bureaucracy and, above all, that emphasis on national development. Tencent, for example, noted in September last year that they had no intention of producing AI chips, but that they were going to invest a lot of money in domestic partners. For example, they are in the process of adapting their infrastructure to be able to connect Huawei’s Ascend platform (particularly the Ascend 950 series) as the main training tool for large models. A few days ago, Tencent’s strategy director already pointed out that that strategy was still in place and that the company expects a significant increase in spending on AI GPUs designed in China. Manufacturing at home. Alibaba and Bytedance have a different approach. If Tencent is focusing on acquiring Huawei platforms, Alibaba and Bytedance are looking to create their own chips. Alibaba seeks to be the most powerful RISC-V chip created to date and it was reported that Bytedance wanted Samsung will manufacture its processor. In the end, whether buying from Huawei or developing the tool internally, the two approaches respond to the great national objective: that at least 50% of the data centers that belong to the State use at least 50% Chinese integrated circuits in their servers. That is one of the great Chinese technological impulses of recent years, one of the crucial points of the Five-Year Plan for the development of the country and, above all, the strategy that Nvidia had been warning the United States about for some time. The age of inference. Because this period of ostracism to which the US condemned China has served for the country to develop three very clear alternatives to Nvidia and encourage companies that are already working with models to develop their own hardware. This is important especially in the new AI framework we are entering, that of inference. Although the AI ​​will continue to train and GPUs will be needed for this, the next step is inference, the agentic era in which the processor or CPU is very important. AMD is moving there, same as Intel or ARMand precisely processors are something that Huawei is good at and in which the Chinese giants can shine as much as their American counterpart by developing chips tailored to their models and needs. Also, as pointed out in CNBChaving your own chips means you don’t have to fight with anyone else in a time when there is scarcity and, of course, if you don’t have to buy from an outsider, there is an improvement in the gross income margin. juicy cake. And this leaves Nvidia in that uncomfortable situation, one in which it wants to participate, but in which it seems that it is no longer needed as much as before. Because China is developing its chips for this new era of AI and Nvidia is running into a final boss called bureaucracy and the pressure groups of the ‘Make America Great Again‘. The first is due to the slowness of the export order processes, something that takes months when orders should be much more agile. The second are the aforementioned pressure groups that hold that any deals Nvidia makes with Chinese companies are less chips for American companies, something that should not be allowed. Meanwhile, Chinese companies are developing their alternatives and Huawei wants to flood the market with 750,000 chips this year, three times more than its shipments in 2025, and Nvidia is falling short of a $50 billion pie. In Xataka | The US has the best AI models. China has something else: AI too cheap to care about

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.