In 2024 a package bomb arrived on a plane. It was the beginning of the great threat to Europe: that of a “ghost” crossing the red lines

Europe lives a strategic transformation that few had imagined possible in such a short time. What began as a series of “flats” (intermittent blackouts, suspicious fires, minor incursions) has become a coherent pattern: a campaign of directed hybrid war that is no longer limited to destabilizing, but rather deliberately explore the thresholds of what it can inflict without provoking a direct military response. It all started a year ago. The silent climb. The plot is explained more clearly from July 2024when several DHL packages exploded in centers logistics from the United Kingdom, Poland and Germany, devices powerful enough to shoot down a plane if they had detonated in mid-flight. The episode, an infiltrated bomb at the heart of the European air system, marked a before and after, because it showed to what extent Moscow was willing to strain continental security and because it exposed the fragility of an Old Continent trapped between an increasingly aggressive Russia and a United States whose commitment has stopped being reliableand. Since then, Europe no longer sees hybrid warfare as a peripheral nuisance, but as a structural threat which targets critical infrastructures, social cohesion and the European institutional framework itself. In Xataka Mercadona has found a vein to grow beyond its white label and prepared food: tourism The Russian laboratory. I counted this week the financial times that the Russian campaign has been refined in breadth and depth. European intelligence services have disabled plots to derail trains full of passengers, set fire to shopping malls, damage dams or contaminate water in urban areas. The attacks are not isolated improvisations: they respond to a “gig economy” model of sabotage in which young recruited by Telegramlocal criminals or foreigners with residence permits act as expendable pawns for unknown objectives. Plus: they are difficult to detect, impossible to anticipate and legally ambiguous, since they rarely there is a direct connection with Russian intelligence that allows them to be accused of espionage. The case of frustrated railway sabotage in Poland (an explosive planted on the Warsaw-Lublin line that came within seconds of causing a massacre) exposed that pattern in its clearest form: unimpeded entry and exit, cryptocurrency financingfalse identities issued by Moscow and a diffuse chain of command that leads to intermediaries as Mikhail Mirgorodsky or even networks managed by former Wagner members. And there is more. Yes, because each cell discovered suggests others not yet detected, and what is worrying is not the errors of saboteurs (sometimes incapable to delete videos of its own attacks) but the scale that this model offers to a Russia resentful of decades of diplomatic expulsions and doctrinally rearmed to a pre-war period. The doctrine that returns. The ISS analysts They recently reported that the archives of the KGB and the StB (Czechoslovak intelligence) reveal parallels disturbing differences between the sabotage manuals of the Cold War and what Europe witnesses today. The objectives listed decades ago (military bases, energy infrastructures, dams, communication systems, transportation) match almost exactly with the whites of the last two years. Equally revealing is the doctrinal sequencing: during times of peace, minor attacks with the appearance of accidents, in pre-war phases, massive sabotage, increased risk tolerated and increasing willingness to cause civilian casualties, and in open war, total activation of clandestine networks for lethal operations. The prelude to something more fat. It we count very recently. If you will, Europe seems to have entered fully into a intermediate stage: a pre-war phase where each incident also functions as offensive reconnaissance, a permanent exercise by razvedka boyem to measure Western reaction capacity, locate vulnerabilities and exploit any weaknesses. The episode of the unidentified drones airports and military bases European operations illustrate this dynamic: cheap raids, of uncertain origin, that revealed systemic failures in the continental air defense and that, due to their replicator effect (copies, jokes, hysteria, false alarms) multiply the psychological and financial wear and tear. A continent without a network. I remembered the new york times This morning an added problem for Europe: that if the Russian threat escalates, the other half of the problem is the growing disconnection with the United States. For the first time since 1945, Europe perceives that Washington is not unequivocally on your side in a matter of war and peace. The Trump administration is not only pressuring kyiv to accept an agreement In Moscow’s terms, it also redefines Europe as a suspicious actor, criticizes the democratic integrity of its governments and promises to openly support the European extreme right. The result is an unprecedented scenario: a Russia that intensifies its hybrid campaign, a Ukraine that depends almost entirely on continental support and a Europe that must finance your own safety while compensating for the withdrawal of US capabilities (satellites, long-range missiles, command and control) that it cannot replace before 2029the year that NATO considers the limit to have a credible deterrent. European leaders also face depleted budgets, electorates hostile to increased military spending, and a rising far-right that Moscow sees as a strategic multiplier. {“videoId”:”x8j6422″,”autoplay”:false,”title”:”Declassified video of the clash between Russian fighters and the American drone”, “tag”:”united states”, “duration”:”42″} The battle of money. The internal European debate on how to finance the resistance Ukrainian reflects the magnitude of the challenge. To support kyiv for the next two years, about $200 billion is needed, an unaffordable figure without activating the 210,000 million euros on Russian assets frozen in Europe. The problem? Right now it takes the name of Belgiumwhich guards the majority through Euroclear, and which fears retaliation from Moscow and the possible erosion of the credibility of the euro as a safe haven. Washington, despite its strategic ambiguity, is also pressing for these funds to be don’t touch each othersince its eventual return is part of the US scheme for a peace agreement favorable to Russia. One more thing. And yet, without that money, Europe would have to coordinate (outside the EU framework) a colossal loan and politically explosive. The crossroads are so profound that in Berlin and Paris they are … Read more

We lost a robot under Antarctica for eight months. He’s back with disturbing facts about the Denman Glacier

On planet Earth there are still many mysteries that we do not know, especially those that hide in the deep sea where It is very difficult for us humans to reach due to the high pressures that exist. That is why science is committed to send robots to explore this area, although the last one we sent stopped sending signals for months, something that would undoubtedly indicate the worst. But the reality has been very different. The exploration. The mission led for the Australian Antarctic Program Partnership Its objective was to study the ice shelves of Denman and Shackleton of which we really knew little. To do this, they used a float Argo standard that modify their buoyancy to sink, take measurements and rise to the surface again. But in this case there was a big problem: under the ice platform there is no surface, only a ceiling of ice hundreds of meters thick that prevented this robot from rising again, remaining adrift, dragged by the waves. sea ​​currents. We expected the worst. This robot stopped emitting signals about what it was doing in the depths, and was already considered ‘dead’ by the researchers. But now it has resurfaced after eight months of being in this situation. And the good news is that even if it didn’t emit signals, the robot continued working and created 200 profiles, going up and down every five days, collecting data that no human had ever seen. Sailing blindly. The question is obligatory: how do we know where the robot was if it did not have GPS under the ice? To do this, the researchers used an ingenious reverse engineering technique described in the study: every time the robot tried to go up to the surface to transmit, collided with the surface and recorded the pressure at the moment of impact. Something that can be related to the depth at which it was found. When recovering data after its reappearance in open waters, scientists compared those impact points with satellite maps indicating the thickness of the ice in the area. In this way, it was like solving a 3D maze: if the robot crashed 300 meters deep, it had to be at a point where the ice is 300 meters thick. Thus they reconstructed their erratic trajectory of almost 300 kilometers. What we know. Until now we had the idea that East Antarctica was always a ‘stable’ and cold area because, unlike the West, it thaws very slowly. But now we have data that makes us doubt this a lot. In the case of the Shackleton platform we know that it is a cold fortress with a cavity under the ice that is filled with very cold water that protects it from melting from below. If we talk about Denman Glacier We can now relate it to a danger zone, as an intrusion of ‘warm’ water was found flowing towards the base of the glacier. Why it is important. Denman Glacier It’s not just any harmless thing.but it has enough ice to raise global sea level by 1.5 meters. In this way, if the water is warm it will be able to cause this glacier to end up melting and this will undoubtedly be a big problem for the coasts of our planet. Now all that remains is to monitor this area, which can now be categorized as dangerous due to the risks that the planet may finally have if this is something that ends up being confirmed. Images | henrique setim AOML In Xataka | We have been trying to figure out what Antarctica would be like without ice since 1950. We just discovered it

Microsoft knows that something has gone wrong, and promises these changes

For years, if you wanted to play on a computer, the answer was almost automatic: you needed Windows. Linux experience was limited and macOS did not offer a competitive catalog. That landscape seemed immovable until Valve decided to really bet on the game on Linux and showed that there was room to shake the board. Steam Deck It came as an experiment that many did not see coming and ended up reconfiguring expectations, to the point that more and more players are talking naturally about switching to Linux. This change of mood has put Windows under a magnifying glass that it did not have before. Windows’ historical strength in gaming was explained by something very simple: it offered the largest catalog, the most mature tools and a fluid relationship with developers. That basis is still real, but its perception has changed. The end of Windows 10 support along with the strict requirements of Windows 11has put teams that were still performing well on the ropes, unless their owners agree to run out of patches or use the extended update program. At the same time, the integration of functions that many interpret as unnecessary additions has generated some wear and tear. Microsoft tries to retain its throne in PC gaming Valve has been preparing the ground for years so that gaming on Linux stops being an experiment and becomes a viable option. Proton has allowed thousands of games designed for Windows work on SteamOS with a level of compatibility that was previously unthinkable, and the Steam Deck has served as a showcase for that progress. The recent announcement of a new Steam Machine for the show consolidates that movement, placing Valve in a position that challenges the idea that Windows is the only natural destination for PC gaming. It is not a frontal assault, but it is increasing strategic pressure. In parallel, far from presenting a laptop with an Xbox seal, Microsoft has opted for a more flexible path: supporting manufacturers that already dominate this segment. Together with Asus and AMD, he has shaped the ROG Xbox Ally and Xbox Ally Xdevices that run Windows 11 but boot into a full-screen interface designed for controller use. This experience unifies games from different launchers and reduces distractions, bringing the console feel closer without giving up the PC nature. It’s a way to compete in portable devices without having to design and maintain new hardware of your own. Besides, Microsoft presented Several internal improvements in Windows 11 are the result of work with the ROG Xbox Ally, which today benefit a large part of the Windows PC ecosystem. They include more efficient power settings, more stable memory management on Ryzen APUs and lower CPU load on tasks that previously affected performance. Still, the company insists that there is still room to cover. “We are committed to making Windows the best place to play, and we will continue to refine the system behaviors that matter most in gaming: background load management, power and scheduling improvements, graphics stack optimizations, and updated drivers.” Several of the technical improvements announced by Microsoft have already reached the desktop. DirectX Raytracing 1.2 is available and provides tools to process complex scenes more efficiently as long as the GPU and drivers are compatible. Advanced Shader Delivery works on select titles and speeds up initial loading when precompiling shaders during installation. Work on neural rendering is advancing cautiously and is only available in preliminary mode for studies. In parallel, Windows 11 has expanded support for LE Audio, which reduces latency and improves the experience in games that depend on sound. The push for Windows on ARM has become another relevant front to expand the reach of the ecosystem. During 2025, devices enrolled in the Insider program have been able to install compatible games from the Xbox PC app, allowing many titles to be played locally. The Prism emulator has added support for AVX and AVX2 instructions, and several anti cheat vendors, such as Easy Anti Cheat and BattlEye, have added specific support for Windows on ARM. From a gamer’s perspective, Windows retains obvious advantages, such as its catalog and the guarantee that almost everything will work without additional tweaks. Even so, the experience in Linux has improved Enough so that some see a more limited system as attractive, with fewer background processes and more predictable behavior. SteamOS solves many historical obstacles, although its popularity does not reach that of Windows, which continues to concentrate around 95% of Steam users compared to Linux still close to 3%. Windows’ journey in gaming has been long and dominant, but its role is no longer automatically sustained. Microsoft’s recent decisions show that the company is aware of this and wants to correct the wear and tear with technical improvements, a clearer roadmap to the future. Even so, Valve’s push has changed expectations and introduced a competitor that did not exist before. What remains to be resolved is whether these movements will be enough for Windows to retain the preferred place that no one discussed for years. Images | Microsoft | Xataka In Xataka | We knew that Valve was betting on Linux, but it was hiding something bigger: a years-long plan to bring Steam to all devices

AVLO’s departure from Madrid-Barcelona seemed like another problem for Renfe. He has left us an unexpected winner

August 2025. The summer had started out average for Renfe and ended even worse. After an exchange of information, Renfe accepted that some Talgo AVRIL trains had suffered cracks in their structures and that they were being taken out of circulation, with the consequent suspension of service. And, of course, that has had consequences for travelers. Goodbye AVLO, goodbye. Everything ended up precipitating in the last days of August but the origin must be sought a few days before. At the end of July, Renfe paralyzed by surprise the sale of tickets for the AVLO service. The news was given from The Economist: AVLO trains had suffered cracks in their cars on the Madrid-Barcelona line. A few days later, Renfe confirmed this fact and began to apply a temporary solution. The idea is that the trains would continue running but reducing traffic speed there where it was believed that the trains had cracked. Some leaked photos Already in August they demonstrated the seriousness of the events. Renfe decided withdraw AVRIL trains of the service. And days later he ended up suspending the AVLOs. September. It was a strange month for Renfe. The company maintained the AVLO service in the early stages, but in the absence of finding a solution, it ended up canceling. relocate passengers of the service low cost of the company on the AVE. That is, a kind of upgrade to travelers who already had their ticket for beyond September 8. That put the company in trouble. Or at least that was the first reading. Since then Renfe does not compete directly against Ouigo and Iryo. Without a cheaper service, the Spanish company was left without the possibility of competing directly against foreign companies. However, it does not appear to have directly affected their results. How has it affected Renfe? If we take into account the latest data from the CNMCwhich refer to the months of July, August and September, we could say nothing or very little. Renfe Viajeros (which adds data from AVLO and AVE) has increased its occupancy by 4%, the number of travelers has increased by 0.6% and they have increased the use of rolling stock by 1.6%. Regarding prices, Renfe has also won. And the company has managed to increase the number of travelers despite an obvious increase in ticket prices. The AVE cost 70.58 euros on average, 13.3% more than in the same period last year and AVLO went to 51.35 euros, repeating the same growth. Retail. But what interests us most about this period is how prices behaved when AVLO was not available. In the month of September, AVLO prices dropped but let’s remember that they were only available for very few days. Instead, Renfe only offered AVE tickets. And they were shot. In September, service prices premium increased to 75.11 euros. It is 11.4% more than the previous month. But above all, it is a price 17.4% higher than that of the same month in 2024 when there were four companies available on the market. Beyond Renfe. Curiously, the one who performed the most in this case has not been Renfe. Iryo is the company that has increased prices the most in this period. With an average ticket price of 63.82 euros, the company seemed to have positioned itself between Ouigo and Renfe, offering an alternative halfway between both services. However, the absence of AVLO in the month of September must have triggered the demand for Ouigo and Iryo. This is the only way to explain why the Italian company raised the prices of its routes to 74.13 euros in that month, just one euro less than the AVE. Year-on-year growth that month reached 83.5%. For its part, Ouigo also raised prices but remained on a somewhat more contained line. In the quarter, the average ticket price was 51.86 euros, which already represents a growth of 20% compared to the previous year. In September, however, prices remained at 52.20 euros, slightly below the month of August but, yes, 30% more expensive than in September 2024. The complete photograph. Expanding on everything that happened, as expected, AVLO’s departure from the Madrid-Barcelona corridor has only increased prices. It is something that we were already beginning to suspect and that was logical if we take into account that it is the corridor with the highest occupation and use of the line. In fact, the latter exceeds 84% ​​and remains around 10% above any other high-speed corridor with liberalized services. It is the perfect environment for passengers dynamic prices suffer. Photo | Xataka and Logan Armstrong In Xataka | Renfe is selling its AVLO for 7 euros in Andalusia: it is the new battlefield in the price war against Ouigo and Iryo

There is now a standard to charge companies to take down the website

When we use Gemini, ChatGPT either Grokit is easy to think that this ability to produce results in a few seconds borders on the extraordinary, even with its common flaws. But there’s no mystery: they depend on models trained with massive amounts of information. This process has ignited an increasingly intense debate about how all that content is used and the extent of control of those who generate it. In this climate a proposal appears that attempts to bring some order. Mass extraction of content. The accelerated growth of AI has exposed the aforementioned phenomenon. Companies use proprietary trackers and third-party data sets that aggregate material from thousands of websites. For publishers, the problem is not just scale, but a lack of transparency about what is collected, how it is used, and who profits. The clash between these interests has fueled demands and debates about the balance between innovation and copyright. What is RSL 1.0. Now it comes RSL 1.0an open standard designed to let publishers express, in machine-readable form, how their content should be used in the age of AI. The initiative arises from the RSL Collective and the RSL Technical Steering Committee, where internet companies, media and standards organizations such as Yahoo, Ziff Davis and O’Reilly Media participate. The objective is for the media to be able to define transparent rules of use and licensing that AI systems must respect. An operating standard. Here the robots.txt file appears on the scene, which has been the fundamental tool to guide web crawlers, allowing or denying access to certain routes on a site. That simplicity was useful for years, although it did not contemplate specific uses such as training AI models. RSL 1.0 goes one step further and describes differentiated permissions through categories such as “ai-input”, designed for training, or “ai-index”, linked to classic indexing. The “ai-all” category allows you to block any use related to AI. The idea is that with this system editors can define specific limits without losing visibility in search engines. The rules are still simple, but now much more informative. Resolving a key limitation. Until now, according to the promoters of the initiative, a publisher who wants to avoid this use must accept that their content will also stop appearing in traditional search, because Google does not offer an individual option to separate both areas. For the co-founders of the RSL Collective, “RSL provides exactly that layer that was missing,” by allowing independent control between both uses. The contribution model. One of the most notable new features of RSL 1.0 is the “contribution” system, designed so that creators and non-profit organizations can demand contributions from the AI ​​systems that use their material. The initiative has been developed together with Creative Commons and seeks to reinforce the sustainability of the digital commons, which brings together billions of open resources on the web. Its executive director, Anna Tumadóttir, points out that “it is essential that there are fair sharing options beyond commercial licenses, in order to continue supporting the commons and protect access to knowledge in the age of AI.” Wide adoption. The release of RSL 1.0 has generated notable support among publishers, platforms, and technical bodies, as well as support from infrastructure providers such as Cloudflare, Akamai, and Fastly. Their involvement is relevant because these services can directly apply the rules that the editors define. Now, although RSL 1.0 introduces a clearer framework for expressing usage rules, it does not solve all the problems posed by training AI models. The standard relies on trackers to follow it and infrastructure providers to enforce it, so companies that ignore these signals could continue to collect content without permission. It is also unclear how it will affect small publishers who lack the resources to negotiate with large platforms. The advancement of AI has changed the way we interact with information, although we often forget that behind those quick results is content created by millions of people. We have to wait to see if RSL 1.0 will balance the rules of the game. Images | Xataka with Gemini 3 Pro | Solen Feyissa In Xataka | McDonald’s has not learned from Coca-Cola and has presented a Christmas advertisement made with AI. The reactions have been even worse

Elon Musk has been refusing to take SpaceX public for 20 years. His new obsession has changed his mind

If there is something that Elon Musk has been repeating since before Starship was called Starship, it is that SpaceX would not go public until the gigantic Martian rocket was flying regularly. The excuse was that Wall Street likes short-term profitability plans more than multi-generational plans to colonize Mars. But the script has changed: SpaceX is preparing its jump onto the stock market, and not to pay for the trip to the red planet. He does this because he needs a lot of capital for “something more” than Starship and Starlink. The largest IPO in the United States. As revealed BloombergSpaceX plans to launch a Public Offering in late 2026 or early 2027. The company is seeking a valuation of $1.5 trillion (trillion, on an American scale) and more than $30 billion in cash, dizzying figures that would be the largest IPO in the history of the United States, close to the global record set by Saudi Aramco in 2019. Musk has been leaving breadcrumbs in X for days about this change in strategy. When the first rumors leaked about a financing round that valued the company at 800,000 million, the tycoon denied itclarifying that “the valuation increases are based on the progress of Starship, Starlink… and one more thing, which is possibly the most significant by far.” What is that thing that makes another round of investment insufficient? Orbital computing. What is clear from Musk’s latest tweets is that SpaceX wants to raise a lot of cash with its IPO for more than just Starship and Starlink: to develop space data centers. The logic, that Musk himself considers validis the same one that other companies like Google are following, but with the advantage of being the largest rocket launcher in the world. On Earth, AI data centers have two major bottlenecks: power and cooling. In space, satellites can receive sunlight 24 hours a day without atmospheric interference and with the possibility of dissipating heat on the dark side of the satellite, eliminating complex water systems and air conditioning of the Earth. Beyond Starlink. SpaceX already has a constellation of 9,000 satellites in orbit, many of them interconnected by laser links. The plan would be to take advantage of all the knowledge and technology that the company has to create a new constellation of localized AI: in Musk’s words, the cheapest way to generate AI bitstreams in less than three years. Their roadmap is hard science fiction: scale up to adding 100 GW of capacity per year using high-bandwidth lasers connected to the Starlink constellation itselfwhich is already highly profitable. And from there we move on to factories on the Moon and the use of electromagnetic rails to launch these AI satellites without the need for rockets. The umpteenth gold rush. Figures like Sam Altman, Eric Schmidt either Jeff Bezos They are already moving to have their piece of the pie in the orbital data center business. Google created the Suncatcher project and Nvidia collaborates with Starcloudwhile smaller startups like Aetherflux have announced projects like “Galactic Brain” planned for 2027. The difference is that SpaceX has the launch experience and is building the largest rocket in the world, with the peculiarity that it aspires to be completely reusable. It’s just the beginning. If 1.5 trillion is already a historic valuation, a recent report by ARK Invest projects that by 2030, SpaceX’s enterprise value could be around $2.5 trillion in a base case scenario, driven almost entirely by recurring revenue from Starlink and declining launch costs thanks to Starship reusability. Going public in 2026 would not just be a financial operation: it would give SpaceX the capital it needs to become the backbone of AI computing infrastructure, turning an internet service like Starlink into something that Musk himself considers “much more significant.” Images | SpaceX In Xataka | Building data centers in space was the new hot business. Elon Musk just broke it with a tweet

They charged him for “leg room” and “priority seating”

Michael O’Leary, CEO of Ryanair, went to dinner at a restaurant in Ireland and ended up receiving a karma master lesson. The restaurant owners had added a surcharge to their bill. things so common like sitting in a chair or having leg room under the table. Checkmate. A quiet dinner…until the bill arrived. As and how I collected the british The Independentthe CEO of Ryanair went out to dinner on a random Friday at a restaurant in Navan, a city in County Meath, Ireland. Dinner at the Luvida restaurant went smoothly while the millionaire executive I tasted some breaded prawns, mushroom toast and sea bass, everything washed down with a good wine. Once the tasting of the dishes was finished and he was satisfied with the course of the dinner, the manager asked for the bill without knowing what they had in store for him from the other side of the counter. The bill: 104.45 euros with many “extras”. When he received the bill, O’Leary found some concepts that caught his attention. The dishes and drinks they had consumed amounted to a subtotal of 104.45 euros. However, to this amount were added some additional charges which added 37.85 euros more to the amount of the account. The restaurant shared the bill for O’Leary’s drinks that night in your profile from Facebook. As can be read on the ticket, the breakdown of those 37.85 euros was made up of 7.95 euros for a supposed “extra leg room” at your table, or 9.95 euros for a “priority seat.” In addition, there was an additional charge of 19.95 euros for a “quiet area reservation.” To make things even more interesting, in the corner of the note you can read the indication “Terminal 1”, reinforcing the parallelism with an airport experience like the one that Ryanair provides to its users. The total bill: 142.30 euros that the manager had to pay after dinner. An Irish “trolling” with a British backlash. Obviously, these are not common concepts in a restaurant since it is assumed that sitting in a moderately comfortable chair while dining is part of the service. However, and displaying sharp British irony, the restaurant staff put O’Leary in the mirror for the way his company treats its customers. In fact, the airline has already been sanctioned on several occasions for its policy of charging an additional fee for hand luggageto which is added the collection of other charges and penalties such as the selection of seats and even make the boarding at the last minute. O’Leary took it with humor. According to what they say local mediathe manager took the joke from the restaurant service with great humor. Which has not transcended It is whether O’Leary finally paid the surcharges or they remained in the final bill as happens with the surcharges of the users of the airline he runs. What the millionaire did do is pose with a friendly attitude with the restaurant staff, who appreciated the visit and the executive’s good nature. “Thank you to Michael O’Leary for choosing to dine with us tonight. It was a pleasure having you. I hope you don’t mind if we add some additional charges to your bill for extra legroom, priority cabin seating and quiet area reservation,” the restaurant published on its profile on social networks. In Xataka | Now we know why Ryanair charges its passengers for everything: it is the key to having a profit of 2,540 million euros Image | Wikimedia Commons (Polish presidency of the Council of the EU 2025), Luvida

Now their biggest challenge is to convince Beijing to let them use them

China is experiencing an unexpected situation in the midst of the race for artificial intelligence: the country’s big technology companies want access to the chip NVIDIA H200but this time it is not Washington that sets the pace, but Beijing. The American government has opened the door to its export under clear conditions, although the final permit now depends on China, that has been tightening its policy for months on foreign semiconductors. Alibaba and ByteDance move in this delicate balance, aware that their ability to advance in AI in the immediate future will depend on what their own regulator decides. Two giants with enormous needs: Alibaba and ByteDance are not simple technology companies, but two of the companies with the greatest demand for computing capacity in China. Alibaba maintains an e-commerce network and cloud services that centralizes a good part of the purchases and sales that go through Taobao, Tmall or AliExpress, both in China and abroad. ByteDance operates TikTok and its Chinese version, Douyin, in addition to maintaining Doubao, its own AI chatbot. This combination of platforms with massive loads turns each jump in power into more than just a technical improvement: it conditions their ability to keep up with the pace of the sector. The change of course in Washington: On December 8, President Donald Trump announced that the United States would allow H200 to be exported to “approved customers” in China, a move that marked a turn from previous restrictions. The agreement contemplates that the US government receives 25% of the income from these sales, above the 15% applied to H20. The White House presented the decision as a formula to strengthen domestic manufacturing and sustain high-skilled jobs, while maintaining direct control over the flow of chips to China. Where the H200 fits into the NVIDIA lineup: The H200 belongs to the Hopper architecture, presented in 2022, and occupies an intermediate position between the generations already established in the market and the new Blackwell line, which is NVIDIA’s priority today. Blackwell-based servers can achieve tenfold performance gains on certain models compared to systems using H200, according to recent company data. Still, the H200 remains a relevant product for advanced training, especially in markets where access to newer hardware is restricted by export controls or limited supply capacity. NVIDIA H200 Why the H200 makes such a difference: The distance between the H200 and the H20 is still notable. According to the Institute for Progressthe H200 achieves a total throughput of 15,840 TPP, almost six times more than the 2,368 TPP of the H20. Compared to the most advanced domestic chips, the gap continues. He Huawei Ascend 910C It reaches 12,032 TPP and offers a memory bandwidth of 3.2 TB/s, while the H200 reaches 4.8 TB/s. That combination of power and speed explains why this chip is so coveted for training large-scale models. Alibaba and ByteDance have conveyed to NVIDIA their willingness to acquire large batches of the H200 if they receive approval from Beijing, according to information shared with Reuters by several sources. Chip availability is reduced because some manufacturing capacity is geared toward newer generations, increasing pressure on the purchasing window. In this scenario, both companies are trying to anticipate whether the Chinese regulator will allow a processor of this level to be incorporated into their training systems without additional restrictions. Access conditioned by the Chinese strategy: Authorization to purchase H200 depends not only on company demand, but on how it fits into the self-sufficiency goals set by Beijing. According to sources cited by the aforementioned agency, regulators are likely to demand precise details about the purpose of each order. In all this, it is no secret that China tries to accelerate the development of its own products through manufacturers such as Huawei and Cambricon, and any import of advanced hardware is examined in light of that strategic horizon. The situation leaves a market in which the rules seem inverted: chips like A100 and H100 They remain under export control, while the H200, more powerful and recent, could arrive in China under an exceptional framework. This asymmetry conditions the advancement of the country’s most ambitious models, which need competitive hardware to continue evolving. The outcome will depend on what Beijing decides in the coming days. Images | NVIDIA | Arthur Wang | In Xataka | Media China is talking about a feature of the ZTE Nubia M153. And the most surprising thing is that the phone is already out of stock.

Jeff Bezos fired the CEO of Blue Origin two years ago. In retrospect, it was the best decision he could have made.

The most surprising fact about Blue Origin is that it was founded before SpaceX. Obsessed with space since childhood, Jeff Bezos saw the potential the aerospace industry would have and began selling thousands of Amazon shares to build a rocket company. He founded Blue Origin in 2000, when his net worth was around $6.1 billion. Two years later, a young Elon Musk obsessed with the conquest of Mars invested $100 million (more than half of what he had from the sale of PayPal) in founding SpaceX. Who would suspect that the company that would end up revolutionizing the sector would be that of the eccentric South African businessman and not that of the CEO of Amazon, who multiplied his assets by 30. The sleeping giant The Blue Origin coat of arms For almost two decades, Blue Origin was the butt of jokes in the sector: a company financed with infinite funds that sold 15-minute suborbital trips to millionaires, but when it came time to reach orbit it only produced powerpoints and legal lawsuits to stop its opponents. Blue Origin was aware of its apparent slowness in the face of SpaceX, to the point of deliberately adopting it as its motto. The company’s coat of arms includes two turtles and a Latin phrase that Jeff Bezos has publicly defended with pride: Gradatim Ferociter“step by step, fiercely.” But although projects such as the powerful BE-4 engines and the reusable New Glenn rocket had been in development for years, the reality is that Blue Origin did not step on the accelerator until the end of 2023, when Bezos said enough and caused a CEO change that has been like night and day. The Dave Limp Effect The first stage of the New Glenn rocket returning to the factory A little context. By 2023, under the leadership of Bob Smith, Blue Origin had become a bottleneck for US national security. The new Vulcan rocket from ULA (the company that had a monopoly on government launches until the arrival of SpaceX) depended on Blue Origin’s BE-4 engines, which kept falling behind schedule. At the end of that year, Jeff Bezos made the decision to remove Bob Smith and entrust the company to the executive who had led Amazon’s devices division during the creation of Alexa or Kindle: Dave Limp. Today, the engine crisis is more than resolved. Blue Origin has celebrated the delivery of the 30th engine to ULA, which will allow its partner to meet its launch obligations for the Space Force. But it has not been the only thing that Dave Limp has managed to channel as the company’s new CEO. Under old management, Blue Origin operated with a crippling risk aversion. He sought perfection on the first try, which translated into eternal development cycles. Limp arrived with the Amazon system under its arm: Blue Origin went from being an R&D company to becoming a real rocket factory willing to take risks. The internal culture had already begun to improve when, in February 2025, Limp laid off 10% of the workforce. “We grew too fast and lost focus,” he explained. But the effect was immediate: Blue Origin has become a company that is agile in decision-making. Instead of having a single rocket that’s scary to break, they’re a real rocket factory. So when the New Glenn finally took off, crashing on the landing attempt, it was not a single prototype: there were other stages of the rocket already on the production line. From New Glenn to Super New Glenn New Glenn vs Saturn V vs New Glenn 9×4 If anyone had doubts about Limp’s management, the events of this last year have dispelled them. Blue Origin has successfully completed two orbital launches that have completely changed the narrative, and which have soon been overshadowed by the company’s roadmap. He maiden flight of the New Glenn It was a partial success. The rocket reached orbit (and there are few rockets that can say that on the first try), but the first stage disintegrated while trying to land. Far from stopping to investigate the failure for a year, Blue Origin analyzed the data, adjusted the software and moved forward with the second attempt, as SpaceX would have done. In November, the second New Glenn successfully launched NASA’s ESCAPADE mission, two probes that were placed at the L2 Lagrange point awaiting gravitational assistance to travel toward Mars. But even a Martian mission can take a backseat when, against all odds, the first stage of the rocket landed on the Jacklyn maritime platform in the Atlantic Ocean. Blue Origin is only the second company to achieve the propulsive landing of a rocket. For the first time, SpaceX has a real competitor capable of recovering orbital-class boosters. One that uses methane for cleaner and cheaper combustion, and that promises to carry up to 45 tons to low Earth orbit. Shortly after the launch, taking advantage of the momentum of success, Blue Origin announced an improved version of the BE-4 engine and a new variant of the rocket: the New Glenn 9×4, which instead of seven engines in the first stage and two in the second, carries nine and four. In addition to a larger 8.7 meter diameter canopy, to launch larger space stations, telescopes and satellites. What does this mean? That Blue Origin is going for the “Super Heavy” category, in which SpaceX competes with the Falcon Heavy and the gigantic Starship, still in development. This variant of the New Glenn will be able to carry 70 tons to low orbit, which with Starship’s permission surpasses almost everything else on the market and, most importantly, with an architecture that has already flown and landed. To conquer the orbit and the Moon With the New Glenn 9×4 scheduled for 2027, Jeff Bezos and Dave Limp’s attention is now focused on scaling the rocket’s manufacturing and reusability capacity to reach 24 launches per year between now and then. SpaceX continues to play in its own league with 160 launches … Read more

The Great Rental Review of 2026 is going to be dramatic for thousands of Spaniards for one reason: 1,700 euros more

The usual thing around these times is that people start talking about New Year’s resolutionsprojects, trips… plans for 2026 that is already around the corner. That’s the usual. In Spain there are thousands of families who face the year with a very different feeling: restlessness. They are tenants, they have been residing in rented houses for years and now they see how their contracts are about to expire in a very different scenario to the one they had when they signed them, back in 2021. Things have changed so much that there are those who estimate that some tenants will have to pay up to 4,600 euros more per year if they don’t want to move. What has happened? For thousands of Spanish families, 2026 will not be the year of North America World Cup nor that of Eurovision without Spain. 2026 will be the year in which they will have to decide whether to move or agree to pay much more for their homes. The reason is a phenomenon that some have baptized as “the big rental review” and in practice it is nothing other than the expiration of the contracts signed between 2020 and 2021. After the five-year extension that marks the lawnow many tenants have to sit down and negotiate with their landlords. But that’s normal, right? Correct. Contracts signed from 2019 onwards last five years if the landlord is an individual or seven if it is a company. During this period they are renewed annually automatically and the normal thing is that the rents are updated in a controlled manner, based on the CPI or the IRAV index. That hasn’t changed. What is special about the rental contracts signed in 2020 and 2021 is that they were agreed in a very specific context, conditioned by the impact of the crisis of COVID-19. It comes with taking a look at the price chart of Madrid prepared by Idealista to understand it. After years of moderate rent increases (or stagnation), in mid-2020 rents began to become cheaper and did not recover until well into 2021, when they gained momentum that continues even today. What does that mean? That if you signed a rental contract in January, February, March… 2021, you did so at an advantageous time that has kept you ‘safe’ these last five years from the price increase that the market has accumulated. Now, once that agreement expiresif your landlord wants to renegotiate the contract, he will do so in a very different context, with rents in maximum values. Has rent become so expensive? Yes. Until now we could get an idea thanks to platforms like Idealista. Now we have a theoretically more precise tool: calculations from the Ministry of Social Rights and Consumption prepared from data from the INE, the Tax Agency and the IEF. The results has advanced them The Country and they show that contracts that must be renewed in 2026 will become more expensive by up to 383 euros per month compared to the time of the original signing, which translates into about 4,600 euros more per year. That would be the forecast for the most extreme cases (not the average), but it is eloquent. Is there more data? Yes. The estimates of advanced consumer The Country show an estimate of how much rents will rise per year for a household with a median income. For Spain as a whole, this calculation shows an increase of 1,735 euros. In the case of the Valencian Community it would reach 2,686, in the Canary Islands 2,267, in Madrid 2,042, in Cantabria 1,869 and in Andalusia 1,952. In the rest of the regions analyzed, the increase in median income ranges between 1,408 and 884-329 euros/year, the latter data corresponding to Ceuta and Melilla. And the calculation of 4,600 euros/year? It comes out of the heaviest estimate, the one that corresponds to the Balearic Islands. There the Consumption data show the increase in rent prices can exceed 4,615 euros per year. As a reference, Idealista indicates that in March 2021 the residential square meter was rented on the islands at 11.2 euros. Today it is above 19. If we take an 80 m2 apartment as a reference, that means that a tenant who five years ago paid 896 euros/month today would have to pay 1,528. That is, 632 more. When managing the advanced table by The Country It is worth keeping several keys in mind. To begin with, it does not include data from the Basque Country or Navarra due to their regional regimes. Nor from Catalonia, since one relevant part of the population resides in declared neighborhoods “stressed market areas”which influences their prices. The increase calculations also seem to have been carried out with respect to the values ​​at the signing of the contract (2021), which leaves the doubt as to whether they have taken into account the updates of recent years. Another fundamental factor is the context: the estimates are based on a portfolio managed by Sumar, which takes time pressuring its government partner to extend hundreds of thousands of rental contracts about to expire. Does it affect many people? The answer is once again positive. At least if we take Consumption as a reference. After examining the data from the Household Panel, Pablo Bustinduy’s department has come to the conclusion that in 2020, 568,500 contracts and in 2021 another 632,300. The first ones have been completing their five years of validity in recent months. The latter will begin to do so from January, affecting 1.6 million people. The communities that will (potentially) be most affected are Madrid, Catalonia, Andalusia and the Valencian Community. The first saw 145,900 contracts signed in 2021, affecting some 404,100 people. In Catalonia, 112,700 and 301,000 were recorded respectively, although there the tenants have the declaration of stressed areas in their favor. In Andalusia there are some 85,500 contracts with 213,700 affected tenants and in the Valencian Community there were 65,500 agreements with 155,000 people involved. Anything else? Yes. … 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.