asterisk

Intel just put 5 billion and an asterisk on the table

by

Europe has been trying for years gain weight on the world map of semiconductors. It is not just about manufacturing more chips, but about reducing dependence on supply chains concentrated outside the continent and regaining ground in the most advanced processes. The United States pursues a similar objective and has reinforced its efforts to attract investmentsfactories and jobs linked to strategic technologies. This industrial race has left a striking scene: one of the largest American companies in the sector has decided to bet billions on expanding its production on European soil. That company is Intel, that has announced an investment of 5,000 million euros to expand and modernize its Leixlip complex, in Ireland. The goal is to increase production of Xeon 6 processors and certain upcoming Xeon products made with Intel 3, the most advanced process the company currently produces in Europe. The move, however, comes after the manufacturer canceled its industrial projects in Germany and Poland. The European Union reinforces its production, but the fine print forces us to clarify how far this victory really goes. More capacity in Ireland, but a European chain still incomplete The core of the plan is not to build a new factory or expand the clean room, but to better equip Fab 34, Upgrade your facilities and extend your automated network that moves the wafers during the numerous stages of the production process. This infrastructure will allow the different modules of the campus to be integrated more fluidly and increase the efficiency of the whole. Intel began rolling out the program in early 2026, although it has not detailed when it will complete the improvements. The expected result is a greater volume manufactured with Intel 3 taking advantage of the existing space. Fab 34 began full-scale production in 2023 and turned Leixlip into Intel’s large European advanced manufacturing center. The installation was born working with Intel 4, used in the first Core Ultra, and later incorporated Intel 3 for the Xeon processors. Both technologies use extreme ultraviolet lithography, known as EUV, to print smaller, more complex structures on wafers. When it began its activity, Fab 34 became the first European factory to use this technique in high-volume production. Main entrance to the Robert N. Noyce Building, Intel headquarters in Santa Clara, California The expected increase responds, according to Intel, to greater demand for server processors and infrastructure linked to artificial intelligence. Although GPUs and accelerators attract much of the attention, data centers still need CPU to run general loadsmanage resources and maintain the platforms on which these specialized systems work. The Xeons occupy precisely that space within their catalog. Expanding the volume of Intel 3 would allow us to better supply that market without waiting for a new plant to be ready. The disbursement also comes after a major financial pivot around Fab 34. In 2024, Apollo contributed $11.2 billion and acquired 49% of a joint venture tied to the facility’s production, although Intel maintained ownership and operational control of the factory. The company repurchased that stake in April 2026 for $14.2 billion. Three months later, it again commits capital to the Irish infrastructure after recovering 100% of that company. Intel’s European bet had been much more ambitious. The company introduced Fab 34 as one piece of a future chain that would combine wafer production in Ireland with two new advanced factories in Magdeburg, Germany, and an assembly and testing facility in Wrocław, Poland. That deployment was to cover within the Union several of the main stages necessary to convert a wafer into a finished processor. The projects were postponed in 2024 and definitively abandoned a year later, when Intel decided to adjust its investments to expected demand. The 5,000 million asterisk appears there. Europe will be able to manufacture a greater volume of advanced wafers, but it will still not have the complete framework that Intel had promised to build within the EU. The company maintains its main assembly and testing operations in the United States and Asia, after canceling the Polish facility that was to cover those stages. Leixlip reduces some of the external dependence, although it does not by itself convert Xeon production into a fully European chain. Images | Intel In Xataka | SK Hynix is ​​betting that the memory cycle is dead. He proves it by doing just what always killed him.

A robot rental industry has been created in China that has plunged prices in a year, but it has an asterisk

by

From spring 2025 to winter 2026, renting a humanoid robot for a business event in China has gone from costing between 10,000 and 20,000 yuan a day to being listed at 1,796. Robot dogs already cost 78 yuan a day in JD.comless than 10 euros. A drop of 80% in twelve months. Why is it important. Beyond the price war, this is the first real scale laboratory in the humanoid robot business, and what happens says a lot about the real state of an industry that moves a lot of money in financing but still needs a human behind each machine. In figures: Between the lines. The most interesting number in this matter is not any of the above, but this: every robot deployed today arrives with a human engineer behind it. This technician assumes transportation, calibration, live operation and unforeseen events. The actual model is not ‘Robot as a Servicebut rather ‘Robot + Person as a Service’. The logic of SaaS (marginal costs that approach zero when scaling) does not apply here. Each new unit in the catalog implies a new payroll. The bottleneck is therefore not in the supply of machines, but in the supply of people capable of operating them. The context. Qingtianzu, the platform controlled by Zhiyuan Robotics and backed by Hillhouse Capital, connects more than 200 suppliers with companies that need robots for presentations, inaugurations or weddings. like a marketplace. During the Chinese New Year, their orders grew by 70% and exceeded 5,000 orders in one week. JD.com saw searches for “robot” increase 25-fold. The demand exists, the problem is the cost structure. Yes, but. Rent has fallen by 80%, but operating costs have barely budged: transportation, engineers, insurance, logistics… Everything remains basically the same.. The payback period cited by operators (about six or eight months) assumes about ten monthly orders at 2,500 yuan on average. But that works during peak demand. Outside of the holiday weeks, that rhythm is broken. The big question. 65% of orders are for entertainment and marketing: robots that dance or parade at fairs and those types of cute but short-lived acts. Intermittent uses by definition. To have a stable base, the sector needs to enter factories, hospitals and logistics. But experts have already warned: the majority of current humanoids are in the “cerebellum” phase, executing instructions without autonomous decision. That jump, according to the most optimistic estimatesit will take about five years. The panoramic. In a matter of months, China has built an industry with funded platforms, distributed logistics and real demand. It is the first country that has brought humanoid robots to the mass market, even if it is to perform in shopping centers and shake hands in dealerships. TrendForce foresees more than 50,000 units shipped in 2026, 700% more. The sector has its own precedent: drones for shows, which did not take off for their industrial uses but for the shows nightlife in cities across China. Robot rental can follow the same script. The difference is that an autonomous drone no longer needs a pilot. The humanoid robot still does. In Xataka | There is a Chinese startup creating the most amazing robots of the moment. It’s called X Square Featured image | Andy Kelly

Spain already sells more electric cars and plug-in hybrids than gasoline. With a (big) asterisk

by

The plug-in vehicle is expanding in Spain. For the first time, our country has recorded more sales of plug-in vehicles (plug-in hybrids and electric) than gasoline and, of course, diesel cars. Or, in other words, they add up to more than pure combustion vehicles per fuel type and come close to exceeding the sum of both. The data, however, has important nuances. you will have read it. And it makes sense, because the data is striking. For the first time, Spain has added more sales of plug-in vehicles than pure combustion vehicles. The figures for last November are, according to ANFACthe following: Gasoline cars: 21,147 units Diesel cars: 4,979 units Plug-in hybrid cars: 11,999 units Electric cars: 9,316 units Therefore, the duel is as follows: Sum of combustion vehicles: 26,133 units Sum of plug-in vehicles: 21,315 units. The first. The news is that for the first time the sum of cars with plug They have surpassed pure combustion gasoline. Cars that do not have any type of electrification continue to represent 22.47% (28.15% if we extend the photograph to the entire year 2025) but this energy is clearly declining. Cars with a plug have already reached 22.65% market share. But the big change is in the year’s accumulated results. This has shot up to 19.29% when a year ago it stood at 11.06%. Growth between January and November 2025 has skyrocketed by 100.12%. That is, twice as many cars of this type have been purchased. The hybrids. Once again, the non-plug-in hybrid is the best-selling type of car. According to ANFAC data, it was the best-selling type of car last November, with 41,034 units and a market share of 43.60%. This data does not stop growing. In the accumulated of the year, the market share is 41.85% and is almost four percentage points more than in the same period of 2024 (38.09%). In total, they have grown 26.04% in sales so far this year. These hybrids are mostly gasoline. But of the more than 40,000 units last November classified as hybrids, 3,852 of them are hybrids with diesel engines, which begins to give some clues about what we are talking about. Right now, non-plug-in hybrids that run on diesel are 1,000 units away from surpassing pure combustion diesels. Why do we talk about an asterisk? Because in their accounts, the microhybrid cars They count the same as a hybrid. There is no way to know how many of the more than 40,000 hybrids sold in Spain in November 2025 and the more than 437,621 units sold so far this year actually correspond to electric hybrids. What is popularly known as a “Toyota hybrid.” In fact, among the best-selling hybrids so far this year we find cars like the Citroën C4, the Dacia Dusterhe Renault Austral or the Nissan Qashqai. All of them have electric hybrid versions but also light hybrids (also called mild hybrid or microhybrids). In fact, the last two only have versions with the ECO label and although of the four engines, two are mild hybridthey all add up as hybrids in the final count. The controversy of mild hybrid. The controversy with the light hybrid or mild hybrid It comes because it is an effective formula for manufacturers to minimally electrify a car to receive approval from the authorities but with a purely cosmetic impact on the car’s consumption or emissions. With the same engine, a car that uses this type of hybridization barely improves the data approved by its pure combustion brother. In Spain, these cars have some advantages over pure combustion cars despite the fact that their real impact is minimal. In MadridFor example, a car mild hybrid It is exempt from paying 75% of the Tax on Mechanical Traction Vehicles (IVTM) during the first six years. These cars also receive the ECO label from the DGT, which is key when receiving more benefits in Low Emission Zonesspaces where circulation is restricted taking into account the car’s environmental labeling. In some cities they also have advantages such as discounts when parking on the street. A redefinition? It is not expected. Neither when it comes to defining them as hybrids nor when it comes to giving them the ECO label. Recently, in the Congress of Deputies The new Sustainable Mobility Law was approved. It was intended to include the study of a review of environmental labeling, but an amendment by the Popular Party prevented it from being included. this will take place. The creation of a new category or the non-provision of the ECO sticker to these cars is, however, a problem. The main obstacle is what to do with the thousands and thousands of cars mild hybrid that have already been sold and that have received their ECO sticker. Provide different labeling to the new cars, despite the fact that they are in the same situation as the current ones, can create a discriminatory situation, but a retroactive withdrawal of the stickers already delivered is not contemplated either. Photo | juice In Xataka | Catalonia wants to restrict circulation to cars with DGT label B in the ZBE: these are the deadlines and the cities

Anthropic trained his AI with millions of books with copyright. To a judge that has seemed correct (with a great asterisk)

by

Anthropic has just achieved a very important legal victory in that legal battle that the world of AI maintains with copyright and copyright for years. The sentence, favorable to Anthropic, can sit a great precedent for the rest of the cases in which AI companies have been sued for training their models with works with copyright. But be careful, because it has not been a total victory. ANTOPIC WIN. In the demand of three authors against Anthropic, the company was accused of downloading millions of books with copyright, in addition to buying some of them to scan and digitize them. The objective: train their AI models. Judge William Alsup has made clear In his sentence that “the use for training was a fair use.” Companies that develop AI models have always shielded in that concept of just use to argue how their models with all kinds of works, including those protected by copyright. Fair use. This legal criterion maintains that limited use of protected material is allowed without needing permission from the owner of those rights. In the laws of Copyright, one of the ways that judges have to determine if that type of activity is a fair use is to examine whether that use was “transformer.” Or what is the same, if something new has been created from these works. For Alsup “the technology in question is one of the most transformatives that many of us will see in our lives.” A victory with a great asterisk. Although the judge indicated that this training process was a fair use, he also determined that the authors could lead Anthropic to trial for hacking their works. The company argued that this was justified because it was “at least reasonably necessary to train LLMS.” For Alsup the issue is precisely that although they ended up buying some of them, he built a huge library for which he did not pay: “Anthropic downloaded more than seven million pirate copies of books, did not pay anything and retained these pirate copies in his library even after deciding that he would not use them to train their AI (at all or never again). The authors argue that Anthropic should have paid for these pirate copies of the library. This sentence coincides with it.” Thomson-Reuters’ precedent. A few months ago Thomson Reuters won a 2020 demand Against a so -called Ross Intelligence Startup. According to them, the company had reproduced material from its legal research division, called Westlaw. The judge rejected the arguments of the defense and declared that the argument for fair use could not be applied in that case. The sentence against Anthropic is right in the opposite direction and blesses that type of use … while companies buy the works with which they train their models. The company of AI, by the way, had already achieved a small legal victory In a previous case against Universal Music. Anthropic downloaded piecework books. In the trial it was revealed how the co -founder of Anthropic, Ben Mann, downloaded in winter 2021 data sets such as The so -called Books3 or libgen (Library Genesis) that they are nothing more than gigantic book compilations, many of which are protected by copyright. Goal is in the same. All companies that develop AI models have been trained with all types of data, including works protected by copyright, and they all face a similar situation. Goal, for example, downloaded 81.7 TB of books with copyright via Bittorrent to train their AI models. That makes the company of Mark Zuckerberg can end up suffering a destination similar to that of Anthropic, which has before him a new very dangerous judicial process for his finances. A potential fine of billions of dollars. As indicated in Wired, the minimum fine for this type of copyright rape is $ 750 per book. Alsup indicated that the illegally unloaded library of Anthropic consists of at least seven million books, and that means that the company faces a potentially huge fine. At the moment there is no date for that new trial. The endless battle of AI and copyright. This is the last episode of a soap opera that we will undoubtedly see many more chapters. Companies like Google, OpenAI either Perplexity They have been equally voracious when training their models and have devastated public (and not so public) data on the Internet. Copyright’s rape demands are accumulating, and cases such as Anthropic may sit a predictive disturbing for all of them if they did not buy the books they used to train their models. Image | Emil Widlund In Xataka | 5,000 “tokens” of my blog are being used to train an AI. I have not given my permission

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.