chip

MediaMarkt offers this Google Pixel with a 6.3-inch screen and Google Tensor G4 chip at a minimum price

by

If you are looking for a compact mobile phone, with a good level camera and the cleanest Android experience on the market, there is no need to go to the high range. MediaMarkt has just brought down the price of Google Pixel 10a. It is reduced to 449 euros, but you can also get an extra discount of 44.90 euros if you are from miMediaMarkt. Thanks to this privilege, the mobile phone remains available for 404.10 euros. The price could vary. We earn commission from these links A good, pretty and cheap mobile This Google Pixel 10a that MediaMarkt has on sale is a pure Android phone with a 6.3 inch pOLED screen with Full HD+ resolution. This screen reaches a peak brightness of 3,000 nits and has a refresh rate of 120 Hz. As we have already indicated, it comes with the latest version of Android and its brain is the chip Google Tensor G4 (it is not the most powerful on the market, but it is more than enough for everyday tasks), accompanied by 8 GB of RAM and 128 GB internal storage. If there is something notable about this device, it is that it promises up to seven years of guaranteed updates. As for its photographic section, it incorporates a double rear lens and, finally, it should be mentioned that its battery has a capacity of 5,100 mAh and supports fast charging by 30 W cable. ⚡ IN BRIEF: offer for the google pixel 10a today ✅ THE BEST Good mid-range camera: The combination of Google image processing with the Tensor G4 engine achieves impeccable photographs in any scene, especially in portraits and night conditions. Update Support: offers 7 full years of operating system updates and security patches, guaranteeing the longest useful life in its segment. ❌ THE WORST Limited raw/gaming power… Although the Tensor G4 moves the system and AI impeccably, it is not designed for extreme gaming demands with ultra graphics during long sessions. No dedicated telephoto lens… Go for a main + wide-angle sensor, depending on the software digital zoom for long distances. 💡 BUY IT IF… You are looking for one of the best cameras under 500-600 euros. It’s the ultimate purchase if your number one priority is taking great photos and videos without the hassle. ⛔ DON’T BUY IT IF… You need to fill the battery from 0 to 100% in 15 or 20 minutes before leaving home, this model will be slow for you. Some accessories for this mobile that may interest you JETech Magnetic Case for Google Pixel 10a The price could vary. We earn commission from these links Google Pixel Buds 2a – Wireless Earbuds with Active Noise Cancellation The price could vary. We earn commission from these links Some of the links in this article are affiliated and may provide a benefit to Xataka. In case of non-availability, offers may vary. Images | Iván Linares (Xataka) and Google In Xataka | Best mobile phones in quality price. Which one to buy based on use and ten recommended models In Xataka | The Google Pixel 10a, face to face with its Pixel 10 family: this is the mobile I would choose

DeepSeek CEO explains how he has transformed the chip shortage into his strategic alibi

by

One of the many battles that China and the United States are fighting is technological, specifically one related to artificial intelligence. US business spending on AI It’s astronomicalbut in China the ambition is no less and the Government wants it to be one of the pillars for the country to be the first world power in the short term. One of the conversations surrounding Chinese AI compared to American AI has to do with efficiency. Very capable models, higher in some cases to the Americans, but with a much lower price and, above all, trained in record time. There are controversies about it, such as theft accusations by giants like Anthropic and OpenAI, but the reality is that if you don’t want to use the AI ​​of an American company, they are there alternatives like DeepSeekKimi or those from Zhipu AI. But, speaking of Deepseek, the company’s CEO is clear that there is a Huge difference between Chinese AI Big Tech and American counterparts: computing capacity. He gives a fact to illustrate it: four Huawei chips are equivalent to one from Nvidia, and that opens the door for the US to have even more arguments to continue putting pressure on the Chinese technology industry. The bottleneck is power, nothing more Liang Wenfeng He is the CEO of DeepSeek and, like so many in his position in recent weeks, he has had the presentation before investors. These types of events are very interesting because they allow us to learn more about both the companies and their plans, and in this case it has been revealing to listen to the head of one of the most relevant AI companies today. They are not official statements, since they are the leak of the almost four-hour meeting published by Tencent Tech, but among its many phrases, we can extract very interesting ideas. The first is what we mentioned: for the boss, the main gap is money and resources. “There is no gap in personnel, since it is basically the same group of people. Talent is not the bottleneck: resources are the biggest bottleneck,” says Liang. The CEO continued to point out evidence, stating that “more cards are always better” and that, at a reasonable price, they buy as many as they can because “converting money into Nvidia cards is definitely better than leaving that money in the bank.” And the Nvidia cards stand out for a fact that is also very interesting: how many Huawei cards are equivalent to one Nvidia card. “Our price offers a reasonable profit: we buy a batch of equipment and they pay for themselves in about ten months” – Liang Wenfeng Assures that the relationship is four to one, or what is the same: if they wanted to train a model with 800,000 million active parameters like the main American AI models, they would need 50,000 Nvidia GB300 cards or 200,000 Huawei 950 cards. That It doesn’t seem to leave Huawei in a very good place.which are those that the Chinese industry and the Government itself is trying to promotebut Liang affirms that Nvidia is digging its own grave in the Chinese market, that the CUDA ecosystem is eroding quickly and that Huawei supernode 950 It can completely replace Nvidia’s GB200 and GB300 in both performance and price. Although the 4:1 ratio between the chips seems alarming for the domestic company, Liang did not specify the workload or whether it measures raw performance, training performance or inference. The US looks closely One detail that Liang points out is that this computing power gap will end up closing as Huawei launches solutions and that, although he does not dare to guess when it will happen, it is something that will end up arriving. But of course, when the competition between the US and China is, in part, in that power of AI, the United States is not in the least interested in closing the gap. In recent months we have seen how pressure has continued to prevent Chinese companies from accessing Western hardware that is key to the development of the semiconductor industry. ASML Extreme Lithography Machines are an exampleand more recently the White House accused Chinese startup Moonshot AI from acquiring servers equipped with Nvidia chips that they should not be able to access. The case of Moonshot AI has been quite popular because a few days ago we told you about that amazing Kimi K3 model that, according to the Americans, would have been trained in Thailand with Nvidia GB300 chips to avoid United States export regulations. And they go further, arguing that Kimi K3 had ‘drunk‘the model Anthropic Fable. Deepseek Strategy Beyond politics, and returning to Liang, although right now it seems that China is not on par in terms of computing power, DeepSeek has a long-term strategy that has less to do with having the most powerful model currently. As we read in SCMPLiang commented that they are setting prices “to make only a reasonable profit, not to maximize revenue.” And that is the key that the company apparently pursues, since they are keeping prices low for users with the aim of their service being increasingly used in the hope of increasing the possibility of achieving artificial general intelligence. So while other companies prioritize immediate profits and capturing market share, Liang is focusing on “increasing the probability of achieving AGI“because it will be that AI that will have great commercial value. And that is what they do seek to exploit. In Xataka | While most oppose AI data centers, there is one group enthusiastic about them: merchandise thieves

ASML sells the most important machines in the chip industry. Now everything indicates that they will be more expensive

by

ASML, the only company in the world capable of manufacturing lithography equipment extreme ultraviolet (EUV) needed to produce the most advanced chips, is preparing for a rise in costs in its machinery. And the first to stand up has been, precisely, its most important client: TSMC. And according to has revealed The Information, the Taiwanese manufacturer is beginning to resist negotiations. Why is it important. ASML is the bottleneck of the entire semiconductor industry. Without their machines, neither Nvidia, nor Apple, nor TSMC, nor Samsung would be able to manufacture the chips that fuel the artificial intelligence boom. That it decides to raise prices, something The Information says it has rarely done in its history, is a sign of the extent to which demand for AI has given the company bargaining power it didn’t have before. In detail. ASML’s plans affect its two large families of machines. On the one hand, according to has confirmed The company’s own financial director, Roger Dassen, during the presentation of second quarter results, there is a way to increase the price of low numerical aperture (Low-NA) EUV systems, its most in-demand technology. On the other hand, ASML has already told several customers, including Chinese manufacturers, that their DUV (deep ultraviolet, somewhat less advanced but still essential) equipment will cost 10% more, according to share the middle. Some Chinese clients have already agreed to pay that extra cost; TSMC, on the other hand, has not wanted to jump through hoops for the moment. Each new generation of ASML machines processes more wafers per hour and improves etching precision, which means more and higher quality chips for its customers. It is what the company itself calls “value-based pricing”, since the logic that follows is that if its equipment generates more economic benefit for the factories that use it, ASML wants to keep a part of that additional benefit, as Dassen explained in the conference with analysts. Between the lines. TSMC’s anger has its explanation. And the Taiwanese manufacturer has been defending for years that You do not need to jump to the very expensive High-NA equipment (which exceed 350 million euros per unit) and that can continue to squeeze out its cheaper Low-NA machines through design improvements and computational photolithography techniques. In fact, according to they count From Tom’s Hardware, its entire manufacturing roadmap until 2030 is based on that premise. If ASML makes precisely those Low-NA machines more expensive as their performance improves, the cost advantage that TSMC had built with that strategy is reduced. Additionally, TSMC needs to buy dozens of these machines for its new plants in Taiwan, the United States and Japan, so even a moderate increase can add billions of dollars to its investment. It is also worth noting that ASML has practically sold its production capacity until 2027 and a good part of 2028, with prices already agreed in those contracts. That means that, in practice, the company could only apply the new rates to orders delivered from the second half of 2028. TSMC will not notice the blow immediately, but it knows that what is negotiated now will mark the price of hundreds of machines in the coming years. And now what. The pulse comes at a time of historic results for ASML. And on Wednesday the company raised its sales forecast for 2026 to a range of between 43,000 and 45,000 million euros, well above what the market expected, according to collect Bloomberg. Gross margin will also improve to 56%, up from 53% previously forecast. To meet demand, ASML plans to increase its manufacturing capacity of its most important EUV machines by 30% this year, and is studying another similar increase for 2028. Cover image | ASML In Xataka | The war in Iran and the Chinese veto redraw the world map of helium for semiconductors

“This is the first time China has proposed a new principle for the chip industry.” Not everyone is convinced

by

Huawei’s path in recent years has been most curious. After being ostracized by the West and being the target of the trade war between the US of Trump’s first term and China, has become the company that thank the United States because of how vetoes have boosted Chinese technology. Because Huawei has become one of the pillars of all the technology companies in the country and they are not satisfied with making chips: they want to lead a new paradigm with a Tau Law that challenges the Moore’s Law. But behind the promises, there are those who can only see one thing: marketing. The Tau Law. My partner Laura exemplified it perfectly. If the chip is a city, the transistors are the buildings, and the cables are the roads, Moore’s Law says: “make buildings smaller to fit more into the same space.” What Huawei proposes, however, goes in another direction: “buildings can no longer be much smaller, so instead let’s make cars travel faster on the roads, and redesign the urban layout so that they travel less distance.” Huawei’s bet is that cars take much less time to go from one building to another, reducing times tremendously so that the chip, without the need to work on smaller lithographs, is much faster. Inside that Tau LawHuawei proposes the LogicFolding architecture, focused on shortening the wiring of the critical paths to increase the density of the transistors. Qionists. And the company is tremendously proud of this to the point that He Tingbo, Huawei’s chip director, affirms that “it is the first time that China proposes a new principle for the global chip industry.” Because, until now, Moore’s Law was the one that prevailed in lithography leaps, but if instead of concentrating on achieving the very complicated objective of making everything smaller, companies focus on shortening the “routes”, the density objective can be achieved in other ways. In fact, Huawei maintains that, before LogicFound, it took three years to go from 126 to 155 million transistors per mm2. In 2026, and with the new architecture, they jumped directly to 238 per mm2. Equivalence, not parity. And if the question is why Huawei is pushing this when there was already another viable way, the answer is that that proven way was not as viable for Chinese companies. Due to the technological veto, they cannot access the most sophisticated machines in Europe ASMLso they can’t “print” chips in advanced lithography as easily as TSMC can, for example. Thus, they have looked for another way to achieve that density, but something important is the terms. Because there is talk that Huawei would reach a transistor density “equivalent” to that of 1.4 nm processes, by 2031, and that is the key. They would be “equivalent” in density and there would be no lithographic parity. With raised eyebrow. In any case, Tingbo is proud of the logo and believes that it can benefit the entire Chinese technology ecosystem, but it makes an appeal. The directive states that “there remain many open questions that no organization can address alone. The tool chain, standards and economic models require contributions from more than one company.” For its part, and as we see in The Registerthere are those who do not see that revolution in the Tau Law. Although they see what the Chinese company is proposing as interesting, they affirm that it is more of a hybrid movement between reality and marketing because “Huawei had to get creative” to overcome the limitations and “they are going in the right direction, but this density is created through packaging and not by making smaller transistors“Indeed, this is what we already had, but they point out that the result will not be equivalent to a true 1.4 nm node from TSMC or Intel in issues such as energy efficiency and temperature. Future. In fact, as our colleagues point out Xataka Mobileit is in thermal management where Huawei is encountering some major problems. But well, the company’s intention is that the results of this Tau Law allow massive AI clusters that they behave as a single chip by 2035 and that the upcoming Kirin chips for consumption also benefit from it. Image | IBM (edited) In Xataka | While the industry obsesses over power, TSMC is clear about where the future of chips lies: in efficiency

the biggest bet in its history to national chip manufacturers

by

China has defined a series of steps to become the first world power. Technological sovereignty falls within this roadmap and, to achieve the objective, a greater investment in local technology is necessary. The current battlefield is that of AI chips, a segment that Nvidia dominates with an iron fist, but if there is a way to change the trend it is with technology and money, a lot of money. And China has found its three winning horses. In short. China is a giant market for Nvidia. With Western Big Tech in his pocket (although Anthropic and Meta are looking favorably at Samsung for inference chips for the age of agentic AI), Nvidia was missing the other big pie. China is a market valued in 50,000 million dollars for whoever manages to position itself as a leader in the AI ​​chip segment, but the problem for Nvidia is that it is not easy at all. Due to bureaucratic problems on the part of the US and China itself, Nvidia is a bit in no man’s land. Meanwhile, national companies have been positioning themselves, presenting their alternatives and GPUs for training this technology. And with the order to bet on the national product, Chinese companies are clear about it. As we read in Business TimesChinese Big Tech currently invests 30% annually of its budget in national AI accelerators, but that spending will increase to 46% over the next twelve months. The horses. Some of these Big Techs are creating their own chips, or already have them. This is the case of Alibaba or Tencent, which would be a kind of Meta or Google: they offer a product to the user, to other companies and, in addition, they have their own data centers with their home-made chips. However, there are other companies that are more focused on creating the hardware to power the internal data centers of those other large companies. This is the case of Hygon Information Technology or Cambricon Technologies, but also of Huawei. All three offer AI platforms that are becoming increasingly important in a context in which Nvidia H20 chips designed specifically for China They are increasingly difficult to find due to that national order to prioritize homemade hardware. The advantage of these companies is that it is no longer about raw power, which is one of the advantages that Nvidia had (although Nvidia is much more, since it also has a complete software suite and so on), but about speed and low latency. For agentic AI, high bandwidth for fast data transfer is essential, and that seems to be a point that Chinese hardware companies are focusing on. The Ascend. The interesting thing is that this strategy is not a future plan and is already being put into practice. Zhipu AI, for example, is a company that has already trained its GLM-5 model on more than 740 billion parameters entirely with the Huawei Ascend 910C. In addition, the Chinese company is working on the development of the Ascend 950 and 960 chips that promise to achieve parity with Nvidia’s most modern architectures. And the supernodes. On the other hand, Huawei continues to push the Atlas 950 supernode. Presented a few months ago, it is a cluster of up to 9,192 Ascend 950DT NPUs per system with up to 1,152 TB of unified memory. It is, according to the company, a direct competitor to Nvidia’s NVL144 rack, achieving 6.7 times more computing power than that of the American company. It is designed for both training and inference and, although it raises doubts about the energy efficiency per computing unit compared to that of Nvidia, what is clear is that it is a product that Chinese Big Tech They will be able to ‘shoot’ for their future models. Huawei is taking the platform around the world and its next appointment is at the World Artificial Intelligence Conference 2026 in Shanghai. Whether Nvidia solutions are used or not, what is clear is that the intention is to prioritize national hardware. China is going to invest nearly $300 billion to build data centers over the next five years to boost AI across all sectors of society. For its part, at least 80% of the basic technologywhere the chips come in, will be supplied by national companies. As several have already pointed out (Jensen Huang among them), the US veto of China It was the worst move they could make. because they only achieved one thing: give wings to their accelerated technological development. In Xataka | Huawei is very close to a great technological milestone. And he is very clear about what has allowed him to do so: the US veto

Anthropic is already preparing just the only thing it was missing to reinforce its AI leadership: its own chip

by

Tech giants have learned something the hard way: depending on third parties is a weakness. Anthropic seems to have realized the same thing, and has been mulling over an important idea for months: making its own AI chip. As indicated in The Informationthe company is in negotiations with Samsung for a potential collaboration, although at the moment everything is unknown. Surprise, none. Already in April Reuters Indian that Anthropic was considering the idea of ​​developing its own AI chips in order to respond to the chip shortage. It seems that this proposal is really gaining traction, because what was once a possibility has now translated into concrete conversations with one of the largest semiconductor manufacturers in the world. The shadow of OpenAI. The move comes just a week after OpenAI unveiled its own custom inference chip, baptized as “Jalapeño” and developed in collaboration with Broadcom. According to company officials, that chip offers better performance per watt than other inference chips, and that could leave Anthropic behind in this race. The reaction of the firm led by Dario Amodei therefore seems logical. They want to diversify, not replace. The strategy, however, is not to completely eliminate its current partners, but rather to have a more diversified strategy for the future. In statements to TechCrunch Anthropic officials have indicated that its “diversified hardware stack, which includes chips from Google, Amazon and Nvidia, will continue to be crucial to its computing strategy.” Everyone wants their own chip. The announcements and news surrounding OpenAI and Anthropic are not, as we said, any surprise. In the last two years we have seen how more and more technology companies joined the trend of having their own AI chips, when before they delegated that aspect to specialized companies like Nvidia. Thus, we have: Samsung, a perfect partner. The South Korean company has been deeply involved in the AI ​​industry for some time. It is a key partner of Nvidia, as it manufactures some of the chips it needs to train and run AI models with its GPUs. Not only that: Samsung is improving your OPC lithography of chip manufacturing thanks in part to the “AI Factory“of both. In Xataka | Anthropic already had Claude writing code. Now he has put it in the laboratories

The first sub-1nm chip is here. It was manufactured by IBM and it is spectacular

by

It has been a long time since we witnessed a milestone like this. Innovations in the field of semiconductor manufacturing they happen constantlybut what IBM has just announced is a monumental achievement: it has managed to produce the world’s first chip with subnanometer technology. This simply means that it has been manufactured on a 0.7 nm (or 7 angstroms) node, which has allowed this company’s engineers to pack almost 100 billion transistors into a surface the size of a fingernail. Crossing the nanometer barrier is not just a matter of numbers. For decades, the integrated circuit industry has evolved under the logic of Moore’s Law. The problem is that this principle has been losing force as transistors approach the dimensions of the atoms themselves. Quantum physics is relentless: each further reduction is an almost unsolvable problem. Reaching 0.7 nm means that IBM has found a way out of that alley. And it has done so not by further miniaturizing transistors according to conventional designs, but by completely reinventing how they are built. This new chip offers up to 50% more performance. Or 70% more energy efficiency if we compare it with 2nm integrated circuits from IBM itself. These two metrics represent the extremes of a spectrum that designers can adjust depending on the application these semiconductors are intended for. For workloads of artificial intelligence (generative AI), cloud infrastructure or next-generation devices, this flexibility is not a minor detail: it is exactly what differentiates a viable chip from a disruptive one. Stack to scale The most important innovation of IBM’s 0.7nm integrated circuit is the technology nanostackwhich we can translate into Spanish as ‘nanostacking’. This is the industry’s first three-dimensional architecture based on stacked nanosheets, and has been developed entirely by IBM. To understand what it means, we are interested in remembering that the previous generation of frontier technology, nanosheets, represented a very important conceptual leap compared to FinFET transistors: instead of a transistor with a vertical fin, nanosheets have several horizontal sheets of silicon stacked and wrapped around the control gate, which improves electrical performance in a smaller space. Nanostack goes one step further: it stacks and staggers entire transistors in three dimensions, thus taking advantage of 3D sequential integration to insert more logic in less surface. What differentiates this architecture from a mere exercise in miniaturization is that each stacked layer can incorporate combinations of different materials, allowing the performance and energy efficiency of each transistor to be optimized independently. Not all transistors on a chip need to behave the same. Some prioritize speed, and others prioritize energy savings Or put another way: not all transistors on a chip need to behave the same. Some prioritize speed, and others prioritize energy savings. Nanostack makes it possible to fine-tune that balance layer by layer, something that planar architectures (or even conventional nanosheets) do not allow with the same granularity. IBM also presented results at the VLSI 2026 conference that demonstrate a 40% improvement in SRAM scaling thanks to this architecture, enabling the manufacturing of semiconductors capable of handling the bandwidth demands of the most demanding AI workloads. Experimental validation of the architecture nanostack It is based on three essential pillars: the ultra-thin dielectric link in CMOS integration, the demonstration of dual-channel engineering capability and the functional operation of a CMOS inverter with expected switching performance. This last point is especially important: a functional CMOS inverter is, in practice, the most basic logic unit of any digital circuit. That nanostack Running it with the expected metrics confirms that this architecture is not just a promising lab result; It is a technology that can be physically built and translated into real computing. IBM and its partners (Lam Research, Tokyo Electron and SCREEN Semiconductor Solutions) have long been working on manufacturing tools and processes with High NA extreme ultraviolet lithography of ASML at its facilities in Albany (New York). Anderon is a quantum chip manufacturing company independent of IBM However, traveling the path that separates the laboratory from the factories requires a lot of time. IBM estimates a production horizon of between three and five years for the first commercial adoption of the technology nanostack in the sub-nanometer node, with a schedule that projects at least a decade of additional scaling. On the other hand, this company has just announced Anderon, a quantum chip manufacturing company independent of IBM that will combine its experience in quantum computing and semiconductors to manufacture quantum wafers on an industrial scale. Be that as it may, with the 7 angstrom node IBM not only demonstrates that the era of subnanometer scaling is physically possible: It also claims its role as a reference laboratory in an industry that has been searching for a way out of the limits of silicon for years. Image | IBM In Xataka | An unexpected salvation for the end user emerges from the memory market debacle: Chinese chips In Xataka | China needs to develop a new type of chips immune to US sanctions. And your scientists have just achieved it

The new Chinese gem of semiconductors is called Enflame. This is the new member of “the four chip dragons” of China

by

The name Enflame may not ring a bell yet. But it is very likely that in the coming months it will end up giving us a lot to talk about. And this Chinese AI chip company just got the go-ahead to go public on the STAR market in Shanghai, the preferred market for the country’s large technology companies. After this, we see how the scheme of large chip manufacturers begins to take shape. Enflame enters the select group of the four big technology companies that are dedicated to AI chipsand that are already listed or are about to do so on the public markets. Who is Enflame and where does it come from? The company was founded in Shanghai in 2018 by Zhao Lidong, an engineer who came from AMD, where he led the development of high-performance processors at the American company’s R&D center. Together with his co-founder Zhang Yalin, Zhao set out to replicate that knowledge in Chinese territory and build a domestic alternative to Nvidia. In seven years has developed five AI chips distributed across four generations of architecture, and has built a catalog that includes processors, accelerator cards, computing clusters and software platforms. Its most recent chip, the L600 module, has passed silicon verification testing, although it has not yet entered large-scale commercial production. Why this IPO matters. Enflame plans to raise up to 6 billion yuan (about 888 million dollars) selling between 10% and 15% of its shares. The money, as could not be otherwise in these times, will be used to accelerate the development of its next generation of AI chips in the cloud and build the software that surrounds them. However, the operation also has a certain symbolic character, since it is the fourth and final addition to the group known as the “four little dragons” of Chinese chips. The other three (Moore Threads, Biren Technology and MetaX) have already debuted on the STAR market, and have been received enthusiastically by investors. In fact, Moore Threads, nicknamed “the Chinese Nvidia”, rose 425% on its first day of trading in December of last year, according to Bloomberg. Restrictions. The reason China is betting so big on these manufacturers is that the United States has been applying restrictions on chip exports for years advanced towards the Asian giant. Nvidia’s most powerful models are blocked, which has created a real shortage in the Chinese market and a strategic urgency to develop its own alternatives. Beijing has responded with public moneyincluding a relaxation of STAR board rules to allow loss-making companies to list, and a $295 billion plan to build data centers that do not depend on American chips. In this framework, Enflame and its groupmates become part of an infrastructure of technological sovereignty. What does it look like? Tencent. Enflame’s greatest asset is also its greatest vulnerability. Tencent owns about 20% of the company and in 2025 it represented 84% of its income, compared to 38% the previous year. That is, almost everything that Enflame sells is bought by Tencent. The Chinese tech giant uses its chips to power large-scale data centers, recommendation systems, chatbots and generative AI infrastructure. The company itself acknowledged in its IPO prospectus that “Tencent’s demand has far exceeded its supply capacity.” That’s good in the short term, as it guarantees income. But how they point out In The Next Web, a chip maker that relies on a single customer for the majority of its sales ends up being exposed if that customer changes priorities. The numbers. Enflame is growing at breakneck speed, as revenues have multiplied a compound rate greater than 80% between 2023 and 2025, but still in losses. Net losses were reduced to 1.2 billion yuan in 2025, compared to 1.5 billion the previous year, and the company plans to close the first half of 2026 with losses of about 600 million yuan. For the same period, it expects its revenue to grow more than three times compared to the previous year, reaching between 10.6 billion and 11.5 billion yuan. On the other hand, investment in R&D has exceeded 100% of sales over the last three years, which says a lot about the phase the company is in (still building, not harvesting). Before the IPO, the Hurun Index valued the company at around $2.8 billion. Where Enflame fits in. Not all dragons are the same. Within China, Enflame competes in a market where Huawei and Cambricon They continue to be the benchmarks in the sector and are already profitable. Enflame, Moore Threads, Biren and Iluvatar CoreX make up a second, younger layer that is trying to break through. Technically, Enflame has opted for application-specific integrated circuits (ASICs), a more specialized architecture, rather than the general-purpose GPUs used by Moore Threads or Biren. Xu Dawei, of Jintong Private Fund Management in Beijing, points out Bloomberg that Enflame “benefits from solid comparatives,” given that its Chinese competitors are already listed on the stock market with valuations well above what their revenues would justify. Companies like ByteDance are actively looking for domestic alternatives to Nvidiaand second-tier manufacturers, including Enflame, are on the radar. Cover image | Enflame In Xataka | TSMC is on the ropes and its biggest problem is not competition: it is water

The last link that Huawei was missing to do without the West in chip design has appeared

by

There is an indispensable component to the semiconductor industry that often goes unnoticed: the software used to Design cutting-edge integrated circuitsknown as EDA by its English name (Electronic Design Automation or automation of electronic design). It is currently in the hands almost exclusively of US controlled companies and its allies, so China needs to have its own software tools specialized in chip design. And little by little he is having them. One of the Chinese companies that are already working in this area is SEIDAand, curiously, its leader knows the American idiosyncrasy very well. Liguo “Recoo” Zhang is Chinese, but he has lived in the US for several decades and has worked at Siemens EDA, the US subsidiary of this German company that dominates the chip design software market in China. SEIDA promised to have its OPC software ready (Optical Proximity Correction or optical proximity correction) by early 2024, but has since disappeared from the news radar. OPC software is very important because it corrects in advance the optical distortions that occur during the photolithography process. When ultraviolet light is shined onto a silicon wafer to “print” the chip design, the light diffracts and the resulting shapes are not exactly as designed. Edges are rounded, corners are deformed and fine lines are narrowed. OPC software anticipates and compensates for these distortions by modifying the original design before it reaches the lithography machine. In this way, the final result on the wafer conforms to the intended design. The EDA that changes the rules In October 2025 Qiyunfang, a subsidiary company of YesCarrier and Huawei, advertisement that your EDA tools They were already being used by more than 20,000 engineers in China. This data has not been independently verified, so it is most prudent to collect it with some reservations. In any case, SEIDA and Qiyunfang are not the only assets that China has in the field of integrated circuit design software. LogicFolding architecture folds transistor-level logic within a single chip into multiple vertical layers And a group of researchers from Peking University has presented a prototype of an EDA tool that is compatible with Huawei’s LogicFolding architecture. The goal of the latter company is to produce chips by 2031 capable of matching the performance of 1.4nm integration technology from TSMC, Intel or Samsung, but without depending at any time on Western chip manufacturing tools subject to US export restrictions. The LogicFolding architecture folds transistor-level logic within a single chip into multiple vertical layers. This optimization requires the use of location and routing tools capable of working on the entire vertical structure simultaneously, instead of working on separate layers. Peking University addresses this problem precisely because its prototype treats the multi-layer structure as a unified design space from the beginning, as opposed to conventional designs, in which each layer is optimized separately and then stacked. During initial testing with industrial-grade open source integrated circuits, this EDA tool has achieved, according to its designersreduce the total length of internal wiring by 30%. Besides, has introduced performance improvements and thermal management versus conventional EDA workflows. It doesn’t look bad, but we will have to wait until Huawei places its first commercial chips with LogicFolding architecture on the market to assess whether this technology is really up to the task. This company has anticipated that its next generation of Kirin chips, arriving this fall, will be the first to incorporate these innovations. Image | YesCarrier More information | SCMP In Xataka | The condemnation that afflicts China: after decades of manufacturing a competitive desktop processor, it is six years behind

eight laptops that you can buy soon with the new NVIDIA RTX Spark chip

by

The market of Windows laptop processors It seems that he is more alive than ever. After Qualcomm’s arrival with its ARM chips, MediaTek and NVIDIA have just officially announced their counterattack: NVIDIA RTX Spark. This is a new generation of processors designed specifically for slim laptops and compact computers running Windows 11. He Microsoft Surface Ultra It will be one of these laptops that will go on sale soon with this new chip from NVIDIA and MediaTek. It is an improvement over the previous generation that has Snapdragon 1,799 euros. Of the new generation with the new SoC, the price of the next Microsoft laptop is still unknown. Microsoft Surface Laptop | Copilot+ PC | 15” touch screen | Snapdragon® X Elite | 16GB RAM | 1TB SSD | Latest Model, 7th edition | Black The price could vary. We earn commission from these links ARM efficiency and RTX power Until now, ultra-thin laptops with long battery life had to sacrifice performance in demanding video games. The alliance of these two giants seeks precisely to end this problem. MediaTek brings its experience in the design of SoC low power consumption, low latency wireless connectivity and intelligent power management. For its part, NVIDIA puts its graphic architecture on the table RTX and its local Artificial Intelligence ecosystem. The result (according to both companies) promises hyper-realistic visual effects and brutal graphic power on devices that neither heat up nor consume the battery. Although not all the technical details have been revealed, the architecture of this chip already shows its intentions to compete in the high rangethanks to key features such as: Advanced local AI: prepared to run AI agents and relatively heavy workflows directly on the device, which could reduce the constant dependence on cloud services for everyday tasks. Up to 128 GB of unified memory: It is projected with a high-capacity, high-speed unified memory architecture designed by MediaTek, an approach reminiscent of the strategy applied by Apple in its M series processors. Cutting-edge manufacturing: The chip benefits from the collaboration between MediaTek and TSMC, which points to optimized power consumption and even notable power efficiency under demanding workloads. Laptops that have been presented and will integrate this chip With the presentation of RTX Spark we have also been able to discover a new series of laptops that will integrate this SWc. However, it is worth keeping in mind that these teams have not yet landed in stores. If you urgently need to renew your computer and prefer not to wait for the next few months, we have selected some of the most interesting proposals that you can find today among the current generation models after presenting the features that the new models will have. Microsoft Surface Ultra: The upcoming Surface Laptop Ultra aims to become the most powerful device in Microsoft history when it debuts at the end of the year. Its great hardware assets will be a spectacular 2,000-nit Mini-LED screen, a haptic trackpad and total connectivity without sacrificing ports. All of this powered by the new RTX Spark chip, which promises graphics performance on par with a laptop RTX 5070 and the ability to process advanced Artificial Intelligence locally and 100% privately. The previous generation of this Microsoft Surface (which is the current one) comes with a 15-inch touch screen, Snapdragon X Elite as the brain, 16 GB of RAM, 1 TB of SSD storage and Copilot+ PC. Microsoft Surface Laptop | Copilot+ PC | 15” touch screen | Snapdragon® X Elite | 16GB RAM | 1TB SSD | Latest Model, 7th edition | Black The price could vary. We earn commission from these links Asus ProArt P16 and P14: aimed squarely at creative professionals, the new ProArt P16 and P14 will stand out for their 120 Hz OLED touch screens. Inside they will hide a brutal configuration of up to 128 GB of LPDDR5X RAM, 1 TB or 2 TB SSD storage options depending on the size of the chassis and all-terrain connectivity to work without limitations. The current generation of Asus ProArt It can be purchased with different configurations. For example, this one that costs 2,399 euros It comes with a 16-inch screen, AMD Ryzen AI 9 HX 370, 32 GB RAM, 1 TB SSD and Windows 11 Home as the operating system. ASUS ProArt P16 OLED H7606WM-SC056W The price could vary. We earn commission from these links MSI Prestige N16: This 2-in-1 convertible seeks to redefine the concept of premium ultraportable equipment. Although the details about its price, dimensions and other technical components are kept secret, the company has confirmed that it will have a spectacular 16-inch OLED screen with UHD resolution and a peak brightness of more than 1,000 nits. If you don’t want to wait for the new MSI Prestige, you have the current generationalso with a 16-inch screen with 2.8K resolution, Intel Core Ultra 9 386H processor, 32 GB of RAM and 1 TB SSD storage. MSI Prestige 16 AI+ C3MG-013ES 16″ laptop The price could vary. We earn commission from these links Dell XPS 16: The Dell XPS with RTX Spark chip will maintain its iconic 16.3-inch aluminum chassis, establishing itself as a workstation with a continuous design but with renewed power. Although the specific technical details and its starting price remain unknown, the firm has confirmed that it will retain versatile connectivity with three USB-C ports, HDMI, audio jack and SD card reader. For those who do not want to wait for the current generation of the Dell XP, it is also a good option. In your 13 inch version costs 1,679 euros on Amazon and comes with an Intel Core Ultra 7 processor, 16 GB of RAM and 512 GB storage. Laptop DELL The price could vary. We earn commission from these links HP OmniBook and Ultra 16: HP has announced the development of its new OmniBook Although the brand has preferred to keep the technical specifications and prices secret, it has announced that these devices will come optimized … Read more

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.