sell

China can’t buy the best Nvidia chips. So Alibaba has decided to connect theirs and sell them as if they were one

by

Alibaba does not want its infrastructure artificial intelligence (AI) continues to depend on Nvidia technologies. Little by little, the largest technology companies in China are assuming the request that Xi Jinping’s government made them at the beginning of October 2024: as far as possible They had to use chips produced in China. Ten months later this recommendation became a requirement. And the data centers that belong to the State throughout the country had to use at least 50% Chinese integrated circuits on their servers. This scenario especially favors Huawei, Moore Threads and Cambricon Technologies because they are Top AI GPU Manufacturers from China, but it also works great for Alibaba. In fact, Alibaba Cloud, its cloud computing subsidiary, has taken a very important step forward. A few days ago it presented a new chip for AI, the Zhenwu M890, and made official a very ambitious itinerary that describes what solutions it will develop over the next three years. This GPU has been designed by T-Head, the semiconductor division that Alibaba founded in 2018. It incorporates 144 GB of HBM3 memory and achieves an interconnection transfer speed between chips of up to 800 GB/s. As we are about to discover, this last feature is essential in the strategy that Alibaba has developed to compete in the AI ​​hardware market. Alibaba is going to spend $53 billion on its infrastructure According to Alibaba, the performance of its Zhenwu M890 chip is triple that of its predecessor. Additionally, it has been designed to perform well both during training of cutting-edge AI models and during inference. An important note: inference is broadly the computational process carried out by language models with the purpose of generating responses that correspond to the requests they receive. Alibaba wants to compete face to face with Nvidia in the deployment of infrastructure for data centers However, there is another relevant fact that is worth not overlooking: in medium precision operations (FP16) the Zhenwu M890 chip reaches 0.6 petaflops, a performance comparable to that of Nvidia’s A100 GPU and three times higher than that of the H20 chip. On the other hand, the ICN Switch interconnection chip allows link up to 128 GPUs M890 so that they work in unison. Alibaba assures that this architecture makes these GPUs work as a single chip, which, on paper, will allow it to compete head-to-head with Nvidia in the deployment of infrastructure for data centers. Regarding the itinerary that will follow until 2028, this Chinese company has anticipated that it plans to launch the Zhenwu V900 during the third quarter of 2027. According to Alibaba, it will implement its own significantly improved parallel computing architecture, will have three times the performance of the M890 chip, will be supported by 216 GB of memory and will reach an interconnection transfer speed of 1,200 GB/s. The Zhenwu J900 will arrive during the third quarter of 2028 with another major architectural leap. This roadmap It reflects that Alibaba goes all out. In fact, it has also announced that it will support this plan with an investment in 380 billion yuan (about $53 billion) over the next three years. Is the largest engagement of its kind in history of the company. Additionally, T-Head is planning its IPO to fund a more aggressive infrastructure investment program, which would put it in direct competition with Cambricon Technologies and Huawei’s Ascend line in the domestic AI chip market. Image | Alibaba More information | Alibaba | ChinaDaily In Xataka | Nvidia has to deal with the absolute distrust of several US legislators. Your plan in China is in danger In Xataka | The US wants to end Chinese AI chips sold abroad. And China knows how to defend itself

Spotify has spent months deleting music made with AI. Now he wants to sell it as a premium product

by

In just a few weeks, Spotify has been changing its position on AI-generated music: months ago it removed more than 75 million fraudulent tracks, launched a distinctive seal so listeners knew when there were human hands on the other side, and tightened its filters against synthetic spam. But the turn came in the talk for investors on May 21, where it became clear that what worries Spotify is not AI, but generating income with it. The precedents. Let’s start with the moves Spotify has made to control the rampant presence of AI on the platform. In September 2025 the company revealed that had removed more than 75 million fraudulent leads of its platform in the previous twelve months. Many of the AI ​​actions were malicious: massive raises designed to steal royaltiesunauthorized voice clones and content which the company’s own executives called “slop.” By then Deezer had detected that it received more than 30,000 AI-generated tracks per day, and that up to 77% of its reproductions were fraudulent. Just a few weeks before the meeting with investors, on April 30, Spotify launched the “Verified by Spotify” seal, a verification mark that distinguishes human artists from the artificial ones, which are increasingly proliferating on the platform. To achieve it, musicians must demonstrate authentic activity, have linked social media accounts and concerts on the agenda (something that, as we have said over the last few months, does not guarantee anything, given the latest successes of AI-generated music, which have their following on networks and their continuous stream of releases). Deals with Universal. The main news before shareholders is a licensing agreement with Universal Music Group, the largest record label in the world, which will allow Spotify Premium subscribers create covers and remixes with generative AI of songs from the artists participating in the agreement. The tool will arrive as a paid add-on to the usual subscription. It was already known that Spotify was considering charging up to an additional $5.99 per month for a “Music Pro” tier with superfan features. Co-CEO Alex Norström said that with this tool, “one song would become 10,000 songs.” The agreement contemplates a revenue sharing model with participating artists, and it was made clear that participation will be completely voluntary by the musicians. This announcement is no surprise: we already knew that Spotify was working on AI products with Universal, Sony, Warner, Merlin and Believe, but without a closed legal framework. Universal had previously licensed its catalog to smaller AI platforms, such as Udio, Klay Vision and Stability AI, but here it is already we enter in the 761 million monthly active users and 293 million paying subscribers. Long live AI. In an interviewNorström made it clear that, faced with multiple tools that allow songs to be manipulated without permission, they want to be the “legal” and “controlled” option. Norström affirms that the synthetic music market already exists and that trying to stop it would be useless, so he proposes regulating it from within, with agreements between labels and platforms, and turning it into a source of income for all actors. To combat AI content that “makes you feel good in the moment” but ultimately leaves the user feeling like they’ve “wasted their time,” Spotify offers verified authors and artists who charge for it. High tension. The announcement comes at a time when many powerful players are beginning to understand the extent of what they are risking. On May 13, a week before the investor meeting, famous producer Jack Antonoff (he has worked with Taylor Swift, Lorde and Lana Del Rey) posted on Instagram against those who use AI to make music. Norström acknowledged in the interview that there is “some negativity out there” regarding AI and called it “reasonable,” although he added that it is due to “poorly aligned AI.” swerve I mean, potify has spent months arguing that the problem with AI in music was fraud, spam, and impersonation. Now it announces that the same synthetic content, controlled and profitable, may be desirable. As we said in our analysis of the algorithmic model that Spotify has built for years linked above, the platform has been encouraging listening that prioritizes the state of mind over the identity of the artist for some time. That is, the ideal breeding ground for synthetic music. All that was left was monetization. In Xataka | We put Spotify, Apple Music and YouTube Music to the test: music streaming has changed and there is no longer an obvious winner

It already has permission to sell its H200 GPU to 10 Chinese companies

by

Alibaba, Tencent, ByteDance and JD.com are four of the ten Chinese companies that already have access to the GPU for artificial intelligence (AI) NVIDIA H200. According to Reutersthe US Department of Commerce, which is the institution that grants or denies export licenses, has authorized at least ten Chinese companies and several distributors, including Lenovo and Foxconn, to acquire Nvidia’s second most powerful AI chip. This news comes almost two months after the US Government confirmed which was going to allow the company led by Jensen Huang to deliver its H200 chip to its Chinese customers. Nvidia announced in mid-March during its annual developer conference that the US and Chinese Administrations had unlocked the sale of this GPU in the nation led by Xi Jinping. However, so far not a single delivery has been made. In practice, the blockade continues despite the March announcement. In all likelihood this is why Jensen Huang has joined the White House delegation participating in a summit with Chinese President Xi Jinping this week. Nvidia is caught between the opposing interests of the US and China, and Huang is going to try to recover a market, the Chinese one, valued at 50 billion dollars in 2026 and which has come to represent 13% of its income. Now the problem is the Chinese Government Earlier this May, Jensen Huang confirmed that he is currently Its market share in China is 0%. Nvidia has not sold its AI chips in this country for several months because US regulations require Chinese buyers to demonstrate that they have implemented sufficient security procedures and that they will not use the GPUs for military purposes. In addition, Nvidia must also certify that it has sufficient inventory in the US. And all this bureaucracy is not being resolved quickly at all. Currently the greatest reluctance to sell Nvidia chips in China comes from Beijing However, currently the greatest reluctance to sell Nvidia chips in China comes from Beijing. The Chinese Government wants to promote developing your own GPUs for AI at any price, which in October 2024 led him to send a recommendation to Chinese AI companies in which he asked them to use chips produced in China as much as possible. Ten months later this recommendation became a requirement. And the Chinese Government is already forcing state-owned data centers throughout the country to use at least 50% Chinese integrated circuits in their servers. The Administration led by Xi Jinping has made this decision because it can afford it. And it is that It already has three very clear alternatives to Nvidia: Cambricon Technologies, Huawei and Moore Threads. On the other hand, in the US there is also a pressure group that opposes the sale of advanced US AI chips in China. Chris McGuire, senior fellow on China and emerging technologies at the Council on Foreign Relations, holds that “any deal that allows Nvidia to sell more chips to China means fewer Nvidia chips for US companies and a minor US advantage over China in AI“. Besides, McGuire argues that “it is surprising that President Trump continues to allow himself to be convinced to put Nvidia’s interests before those of America.” Image | Nvidia More information | Reuters In Xataka | The US remains committed to stopping China. Now it has targeted the second largest Chinese chip manufacturer

A company wants to sell sunlight on demand using gigantic mirrors in space. We have questions

by

A Californian startup wants to sell solar light at night and, although it has not yet started, many scientists are already putting their hands together. They find it difficult to do it correctly for technical reasons, but they consider that it would be even more serious if these difficulties are resolved. The consequences for people’s health, the environment and the work of astronomers can be devastating. The longest day. The goal of Reflect Orbital is to launch into space a swarm of 4,000 satellites loaded with giant mirrors. These would capture sunlight from the illuminated side of the Earth and reflect it in dark areas. Thus, the solar panels could work 24 hours a day, not only when sunlight naturally falls on them. First steps. For now, the objectives of this startup have been developed only on paper. They already have their first satellite ready, which they have named Eärendil-1, in honor of a JRR Tolkien character. However, They are still waiting for the Federal Communications Commission (FCC) to of the United States gives the green light for its launch. In principle It is scheduled to take place throughout this month of Aprilbut there is no definitive date. Once in low Earth orbit, this satellite will deploy an 18-meter-wide mirror, which would be capable of illuminating a 5-kilometer patch on Earth. If all goes well, a swarm of 4,000 mirrors could be launched by 2030. The background is not good. There was already a project similar to this developed in Russia in the 1990s. The goal of the project, named Znamya, was to illuminate Siberia in the dark winter months. And they got it. However, the resulting light was so dim and the satellite so difficult to control that the mission was never completed. More than technical difficulties. Fionagh Thomson, researcher in spatial ethics at Durham University, explained in statements to Live Science who does not believe that the project is viable today, since the engineering involved is very complex. They already verified it in Russia. But that’s not all. Both this and other experts warn that a large amount of light pollution would be generated, which could affect the circadian rhythms of living beings in the illuminated environment. It could also dazzle aircraft pilots and make the work of astronomers difficult. Even astronomy enthusiasts trying to look at the sky with binoculars or a telescope could suffer vision damage if they encounter the light reflected from these satellites. After all, the population would not be notified before changing the direction of the mirrors. Worse than Starlink. starlink, Elon Musk’s telecommunications companyhas been receiving criticism for many years for the artificial way in which they illuminate the night sky. However, this company’s satellites accidentally illuminate the Earth. In this case it would be something deliberate and, therefore, even more intense and serious. It’s not worth it. All these risks are not worth it when you consider the results. And many other experts assure that the light that would be obtained would be too dim. The solar radiation that would reach the solar panels, for example, would be a minimum fraction of that which arrives during the day. In order to obtain a sufficient amount of light, an exorbitant number of satellites would have to be launched into space and that would be expensive and even more dangerous. Beware of space debris. If the mirror of Eärendil-1 will measure about 18 meters in diameter, the goal of Reflect Orbital is to launch satellites into space with even larger mirrors, up to 54 meters. In general, they would be giant objects; who would therefore be at greater risk of impacting with meteorites or space junk fragments. The more exposed surface, the more risk. This would not only mean the uncontrolled release of fragments resulting from the impact, it would also cause damage to the structure of the mirrors themselves. A leaky mirror would be even more difficult to control and its harmful effects could worsen. Therefore, although the goal of selling sunlight at night seems feasible on paper, in reality it is complicated and dangerous. We’ll see where all this goes. Image | Reflect Orbital In Xataka | Solar thermal plants are in the doldrums, so now they have two jobs: generating energy by day and hunting asteroids by night

What is SMIC, China’s big chip manufacturer, doing right now? According to the US, sell them to Iran for the war

by

The war in Iran continues. On the one hand it is said that it is almost finished, but on the other we have the shipment of thousands of American paratroopersmore calls for support and one sided offensives and from another. But in almost any conflict, not only those in the countries involved come into play, but also the allies. And the United States has leveled a pretty serious accusation against China: SMIC is selling chips to Iran. Well, “almost certainly.” SMIC in the spotlight. Semiconductor Manufacturing International Corp the great Chinese semiconductor foundry. Included in the blacklist of the United States government along with Huaweihas managed to develop advanced chips in record time. They have not only challenged everything the US thought they could dobut that association with Huawei and the country’s push for the technology industry have made it one of the spearheads of China’s technological sovereignty. That SMIC has been able to manufacture advanced chips when it was denied access to cutting-edge technology is something that upsets the US government, which reiterated the sanction and keeping the company on the blacklist for alleged ties to the Chinese government. And the latest accusations are not going to relax the tension. ANDUSA says yes. SMIC makes chips and obviously sells chips. And the United States claims that they are supplying technology to Iran. a few days ago, Reuters published an article in which it included two statements by “two senior officials in the Trump Administration” that suggested that Beijing, perhaps, is not staying as far away from the Iran war as they would have us believe. In the article they state that SMIC has been sending chip manufacturing tools to the Iranian army. This raised questions about Beijing’s stance in the conflict, with officials noting – on condition of anonymity – that the company began shipping the tools about a year ago and that they have “no reason to believe shipments have stopped.” A year ago, the United States was not at war with Iran, and China has long maintained a normal trade situation with Iran. US officials note that, in addition, “they have almost certainly also technically trained Iran on semiconductor technology.” And let’s remember that these chips are in everything: from routers to missiles. China says no. The Reuters article does not give any further information or details on whether Iranian tools that included US technology have been confiscated –something that does occur in other conflicts– and neither the Chinese embassy in Washington, SMIC or an Iranian spokesperson at the UN responded to requests for comment. Who has left Lin Jian, the spokesperson for China’s Ministry of Foreign Affairs, spoke out and did not hesitate to classify the report as “false information.” He accused certain media outlets of launching self-serving news and then classifying all reports as “false information.” On this issue, China has been caught between two waters, first condemning the assassination of Ayatollah Ali Khameini by the Israeli and US forcesbut also expressing his rejection of the Iran attacks on Gulf states that house US bases. Back in focus. Beyond Iran, the United States accusations are part of an operation that began a few years ago. The veto of Huawei marked the beginning of the current trade war between China and the United States, but it also marked China’s ‘awakening’ in technological matters, quest for sovereignty and a technological war that branched into chips, robotics, energy, communications, artificial intelligence and in the military arm. SMIC is the large Chinese manufacturer that defied US vetoes by managing to manufacture the chip of the Huawei Mate 60 Pro before whom The US authorities could not believe and, if they manage to demonstrate that they are involved in supporting Israel when China is not actively participating in the conflict, they will have more reasons to intensify the vetoes and sanctions. And all this is framed in a current situation in which Trump and Xi Jinping will meet in a few days to discuss international relations and where the purchase of American technology by China is expected to be one of the points of the day, with NVIDIA very interested in biting a piece of the $50 billion pie that the Asian giant represents. Images | Ballistic Missile, ASML In Xataka | While the US bombs Iran, something unusual has happened: drones attacking the nuclear bases in North Dakota

sell us a non-car for 6,000 euros

by

The cars have gone up. I don’t know if you’ve noticed but I’d bet you have. In 2023 we will tell you in Xataka that, once inflation had been discounted between 2014 and 2022 (around 18%), the prices of the best-selling cars had risen by 50%. A good part of that increase was concentrated in cars between 20,000 and 30,000 euros. There, the density of cars whose price had increased by more than 20% was greater. And there was another worrying focus. Cars between 15,000 and 20,000 euros they became more expensive on average very close to 20%. An increase that, in this case, affects the customer’s pocket even more because they were entry-level cars, which many people buy as their first car or because they cannot buy another type of vehicle. There, the Dacia Logan was key. According to the data collected, it was the cheapest car on the market. In 2014 it cost 9,150 euros. By 2023 it had become almost 30% more expensive (discounting inflation). Today the Logan is not even sold. Dacia’s cheapest car, discounting Springis the Sandero and costs 13,520 euros as long as we adjust to its financing to obtain the corresponding discounts. The low range of cars has seen how the European Union has hit it with a series of security obligations which has irremediably raised the price of cars. Today, every car sold on our continent has to have rear view cameras, emergency braking or speeding alerts, among other obligations. This increase in price has called into question Dacia’s commitment. The Romanian company has had to leave behind that “it does the same as any other car but for only 6,000 euros.” Denis Le Vot, who was CEO of Dacia, criticized these decisions of the European Union by pointing out that the majority of drivers deactivate them, in statements collected by Coach. The truth is that these They put the company on the ropes so we have seen some change in its image. Dacia is no longer the company that offers you the cheapest car on the market (it is difficult to compete with MG and a good string of Chinese brands), now he is the one who tries to give you more for less. Dacia has made greater efforts to take care of its image. The new ones Dacia Duster and Bigster They are good proof of this. But Dacia wants to keep one foot in the cheap car, in the most affordable option. And that is why they have thought of a “no-car”. The Citroën AMI will have a Dacia rival. For 6,000 euros, a “non-car” It has been rumored for a long time but our colleagues from L’Automobile They exclusively say that they have confirmation from people who work within Dacia that the company will launch a small electric car to rival the Citroen AMI. That is, Dacia will enter the game of light and/or heavy quadricycles. And he will do it by hitting the table: “our goal is 5,990 euros,” the sources consulted have told the French media. This would mean putting an electric vehicle on the market about 3,000 euros cheaper than the Citroën option. The vehicle would not yet have a name but it would be a heavy quadricycle limited to 45 km/h and 80 kilometers of autonomy. It is a solution that, as Citroën has explained to us, is used by parents who want their children over 14 years old to be motorized but also by many older people who are looking for a car to get around an urbanization or to run errands and short trips. The movement makes some sense. Renault and Dacia have been two of the brands that are putting the most pressure to put electric cars on the market in a new category that the European Union would be studying. The one known as eCar would be a small electric car, with limited power and speed and for which it is expected that it will not be necessary to inflate it with the driving aids already mentioned. a sort of kei car europeeither. The project, however, seems to have been in neutral. The idea is that the car would be manufactured in Europe and perhaps that is why Dacia can convert its proposal (they already presented a prototype of this eCar known as Hipster) to a small quadricycle that is very cheap to manufacture. Of course, two doubts arise. The first is how to put a car on the market cheaper than the Citroen AMI or the Fiat Topolino (much cheaper, about 3,000 euros) if these two Stellantis options are manufactured in Morocco and are little more than the sum of two parts. And the front and rear are the same piece (only the color of the headlights changes) and it only uses a door that is mounted traditionally on the right and suicide style on the left to save costs. The second is whether there is a really juicy market for this type of vehicle. In fact, doubts arise as to whether there is a market niche for the eCar project since there is talk that it would be limited to 90 or 100 km/h. If you apply pure logic, you would think that European cities are perfect for this. But the truth is that we are not Japan and, here, the logic is not as overwhelming as in Japan. Photo | Dacia In Xataka | The best-selling car in Spain is the Dacia Sandero. And that says nothing about our supposed poverty

A Japanese store begs its customers to sell it their used equipment.

by

Building a computer today is more complicated and more expensive than it was just a few months ago. It is not just that some components are in short supply, but that the market balance has shifted and is directly affecting availability and prices. Memory is one of the best examples of that pressure, and what is happening in that market It’s starting to have knock-on effects.. What seemed like a problem limited to certain user profiles has been gaining scope and can no longer be understood as something isolated. The scene that explains everything. In Akihabara, one of the great electronics and computing centers in Tokyo, a store has decided to do something unusual: ask its own customers to sell it their used computers. Sofmap Gaming posted a message on his X account in which he openly acknowledged the situation: “Gaming PCs, even second-hand, are really out of stock right now.” Next, they launched a direct request: “Please, if you are going to buy a new one, sell us your gaming PC…”. The scene was completed with practically empty shelves and another revealing detail: the store itself assured that it is buying back quite expensively and that it buys practically any PC, whether gaming or not. Click to see the original message in X It is not an isolated case. We are facing the visible consequence of a tension that has been accumulating in the hardware market for some time. First it affected those who assembled their own equipment, with increasing difficulties in finding certain components or assuming their price. Then it began to be noticed in manufacturers and assemblerswho have had to adjust configurations and rely on previous stock to keep up. Now, that pressure has ended up being transferred to the point of sale, where it is no longer just about selling, but also about getting the product. What’s behind the shortage? To understand what we are seeing we have to look at a clear change in the industry’s priority. The explosion of AI has skyrocketed the demand for memory for chips and systems intended for that businessespecially in the data center environment, and that is altering how production is distributed. Part of the problem arises in the most advanced memory used for AI, but its impact ends up spreading to the rest of the market. From Micron they summarized it this way in statements to CNBC: “We have seen a very strong and significant increase in demand for memory, and it has far exceeded our ability to supply it.” Consequences. The pressure on memory ends up trickling down to the devices we buy, whether in the form of higher prices or less ambitious configurations. As we have said, it has already put in the computer industry, but it is also threatening the smartphone sector and the consoles. Without going any further, Sony recently announced an increase of 100 euros on the PlayStation 5. And everything seems to indicate thatCars are not going to escape this crisis either.. Old hardware that is revalued. In this context, what until recently we considered old hardware begins to have a different value. Not because its performance has changed, but because the market around it has. What we have seen in Akihabara is not an isolated anecdote, but a sign of the extent to which availability has become a real problem. When a store asks its customers to sell it their own equipment, what it is showing is that something does not fit into the usual supply chain. Images | Andrey Matveev In Xataka | The price of RAM has skyrocketed and the best example to see the debacle is a 100 euro PC: the Raspberry Pi

Doomsday’ are released at the same time and Hollywood already wants to sell us a new Barbenheimer. But it’s not the same

by

Warner Bros. and Disney are going to release their two big bets of the year (‘Dune: Part Three’ and ‘Avengers: Doomsday‘) on the same date: December 18, 2026. The industry has not seen something like this for years. In fact, we haven’t seen it since the Barbenheimer phenomenon, and the question is whether its effect at the box office can be repeated… when the two films share exactly the same audience. Dunesday, or whatever you want to call it. The two most powerful Hollywood studios, two of the most popular franchises in recent years, a single date. The industry already has a nickname for the event: “Dunesday“. But the fact that both coincide on that date does not respond to an agreed strategy. ‘Dune: Part Three’ arrived first in the pre-Christmas gap. ‘Avengers: Doomsday’ was initially scheduled for May 2025, but postponements derived from the strikes of actors and scriptwriters pushed it to where it is now, from where it is unlikely to move since on websites, networks and others there is a countdown in motion until the inevitable premiere. The Barbenheimer precedent. Summer 2023 was left for the annals thanks to the coincidence of two very different films, ‘Barbie’ and ‘Oppenheimer’, on the same weekend in July. The combined weekend exceeded $300 million in the US alonebecoming the fourth best in historyonly behind the three opening weekends of ‘Avengers: Endgame’, ‘Infinity War’ and ‘Star Wars: The Force Awakens’. Different audiences, even different motivations for going to the movies, but a curious, almost miraculous real possibility of doing a very rare double program on the same day. Why Dunesday is different. ‘Dune: Part Three’ and ‘Avengers: Doomsday’ don’t work like that. Both target similar viewers: devotees of science fiction and big-budget action, with a predominantly male base (this is important). Significantly or coincidentally, both feature Florence Pugh in their casts, which illustrates the extent to which they are not identical films, but do have their roots in a shared territory of blockbuster Hollywood. The idea of ​​Dunesday, as Variety says, is not to propose two different and complementary plans, but to fill the respectable with a giant dose of similar things. Why don’t they wiggle. That neither film wants to move from that date has a concrete explanation, prior to the invention of Dunesday: there are not many good gaps in that part of the year. Sony releases ‘Jumanji 3’ on December 11, a week early. Further back, the first weekend of December is usually a black hole at the box office (families are shopping, not in theaters), and Thanksgiving is traditionally occupied by Disney’s animated bet, because it is a more family-friendly box office weekend. December 18 is the best date available for two films that need a strong start. What if it works? If the experiment works, the combined box office could be among the best weekends in history. In any case, here we find the usual problem: that July 2023 generated a social phenomenon (costumes, double sessions planned weeks before, debates on networks, friendly rivalry between the actors of the films, who encouraged people to go to both) that was organized outside the industry. “Dunesday” arrives much more in advance and with a slight air of a commercial montage. That’s not going to make the films click, of course, but the truth is that we are dealing with phenomena with nuances that go in opposite directions. In Xataka | The reality of Spanish cinema: ‘Torrente’ has brought more people to theaters this weekend than any film since 2019

whoever sells the shovels also wants to sell the map

by

NVIDIA prepares the launch of NemoClawits own open source platform for enterprise AI agents. The official announcement is expected for GTC 2026, the company’s annual conference, which starts on March 15 in San José. Why is it important. NVIDIA has built its dominance by being the neutral infrastructure provider: it sells shovels to anyone who wants to dig. NemoClaw changes that position. By entering the agent software layer, you begin to compete directly with Anthropic, Microsoft, Salesforce and the community itself open source which until now considered NVIDIA an ally, not a rival. The context. The obvious trigger has been OpenClawan open source AI agent that allows complex tasks to be executed locally without human intervention and that OpenAI acquired a few days ago. Its success showed that there is a huge demand for freelance agents, but it also exposed its risks: Meta even banned its use on business devices following an incident in which an agent accessed a computer without instructions and deleted emails en masse. Companies needed something more controlled and NVIDIA has seen the opportunity there. Between the lines. The platform will be hardware agnostic: it will run on chips from AMD, Intel and others, not just NVIDIA GPUs. It is an apparently generous movement that hides a clear expansionist logic. It’s the same move that Meta made with Llama: giving away the software to boost demand for the hardware that runs it. If NemoClaw ends up becoming the standard de facto For business players, NVIDIA will be able to maintain its influence on the ecosystem even if competition in chips intensifies. The big question. NVIDIA has reached out to Salesforce, Cisco, Google, Adobe and CrowdStrike to forge early partnerships, but none have confirmed any agreement. There are reasons for skepticism: Salesforce has Einstein, Google has Vertex AI Agent Builderand both have clear incentives not to give ground at the application layer. The fact that they contribute to NemoClaw’s open source does not prevent them from continuing to develop their own platforms in parallel. NVIDIA’s success will depend on whether NemoClaw brings something that no one else can, or if it is just another framework gathering dust on GitHub. Yes, but. Gartner estimates that more than four in ten agentic AI projects will have failed by 2027. The business agent market is promising, but still more promise than reality. Furthermore, NVIDIA is entering an area where its competitive advantage (raw silicon power) matters less than the ability to orchestrate complex workflows, manage agent memory, and ensure security in regulated environments. That’s something that chips don’t provide. In Xataka | If the question is how much OpenClaw is taking over, the answer is… in China they are lining up to install it Featured image | Xataka

In London more and more people lose money when they sell their house. The question is whether it is the canary in Europe’s mine

by

Located north of the Thames, Tower Hamlets is one of the districts most emblematic from London. In fact, it covers a large part of the East End, the historic center of the capital. For years (like most of the city) it also represented something else: a juicy market for those who wanted to invest in housing and achieve high returns. Not anymore. In 2025 about 30% Of the owners who got rid of their homes in that neighborhood (mostly apartments) had to do so for less money than they paid at the time. And it’s not just something that happens in Tower Hamlets. What has happened? That in London housing is no longer an infallible business. This is suggested at least by the latest study published by Hamptons, which reveals that in 2025 Londoners were the Britons most likely to lose money from the sale of their properties. Even more than its neighbors in the northeast of the United Kingdom, who have spent years leading the ranking. “Rising London house prices are no longer the safe bet they once seemed,” concludes the report, which is supported by the Property Registry. What do the figures say? that last year 14.8% of people Those who sold their home in London did so for less money than they originally paid. It may seem like a modest percentage, but it is striking for several reasons. To begin with because it is the largest in the entire United Kingdom. The national average is 8.7% and there are British regions where this indicator is much lower, such as Wales (6.2%), East Midlands (6.7%) or West Midlands (6.9%). London has effectively ousted Nort Easth, which had dominated the sales ranking with losses for the last decade. Is Tower Hamlets a unique case? No. Tower Hamlets is the London district where the trend is best appreciated, but is not the only one in which a significant proportion of homeowners (28.2%) have lost money by getting rid of their homes. In the City, 26.2% of sellers closed transactions in “red numbers”, in Kensington & Chelsea 22.4%, in Westminster 22.1% and in Hammersmith & Fulham 20.8%. Curiously, in the cheapest district of London, Barking & Dagenham, only that indicator is much lower: 5.3%. “In some cases, even homeowners who bought a decade ago risk getting back less than they paid, something almost unthinkable in 2015. And for many the sums are small,” the study insists. “In the coming years it is likely that more sellers will have missed out on the price boom that London experienced between 2012 and 2016, as they bought at the peak of the market.” Is there more data? Yes. The Hamptons report raises some interesting ideas. For example, most of the sales with losses (close to 90%) were carried out by apartments. If we talk about houses, the photo is somewhat different. Hamptons technicians recognize that in 2025 the average seller in London pocketed 172,500 pounds more than what they originally paid when purchasing their home, but they insist on the increase in sales at a loss: if in 2019 they represented 5.9%, in 2025 “red” operations already represented 14.8%. Is it the only report? No. Over recent months, more analyzes have been published showing that the London property market is not going through its best moment. There is talk of a price drop of 5.1% at the end of 2025 (which takes the market even further away from the 2022 data) and even from a sluggish prime housing market that will not rise until at least 2028. “In London, the growth of house prices is no longer a safe bet,” he explains to Financial Times Aneisha Beveridge, Hamptons manager. There is studies which show that prices are declining in half of London’s neighborhoods, leaving a “two-speed” market: that of the most expensive (and volatile) areas and the cheapest, which has demonstrated greater resilience. In December Bloomberg warned that homes worth more than two million run the risk of depreciating, losing almost 5% of their value in one year. What is the reason? The big question. When explaining the London trend the analysts they point out several factors. One of the main ones is the regulatory change, marked by the end of discounts to the purchase of housing and a greater penalty for the purchase of second homes and houses as investments. The authorities have also focused on the prime segment, rethinking the status nom-dom for large foreign fortunes and raising local taxes for the most expensive properties. Added to the above is the influence of Brexit, the exorbitant prices that London reached in 2022 or how difficult it is for families to access the market, partly because the cost of rent neutralizes the ability to save. The question that some are already made is whether London is an isolated case or should be understood as a canary in the mine for other European capitals. Image | Benjamin Davies (Unsplash) In Xataka | Housing is getting so expensive that in the United Kingdom there are already people opting for plan B: living on boats

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.