Xiaomi has made profits selling cars in its first year. The problem is that it has optimized for an unrepeatable moment

Xiaomi Auto, Xiaomi’s car division, reported a few weeks ago something that is considered impossible in the automobile industry: achieving profits in its first year. It has had a healthy gross margin of 25.5% and a net profit of 680 million yuan, about 82 million euros, thanks to 109,000 cars delivered in a single quarter. Barely a year after selling its first car, the division presents numbers that place a newcomer in the same range as BMW or Mercedes. One that took Tesla years to reach and one that other manufacturers like NIO are still not there. Some They died trying to get there. Lei Jun has executed an impeccable launch and his investors have reason to be impressed, but if we take a closer look at the numbers and break down the origin of the margins (something that must be attributed to Poe Zhao’s wonderful analysis in Hello China Tech), a different story appears: that of a company that has perfectly optimized for a moment that will not be repeated. Two figures: The average price per car in the third quarter was 238,000 yuan (about 29,000 euros). The broadest category was close to 260,000 (about 32,000 euros). Those numbers They are not representative of the market that Xiaomi wants to addressbut rather they represent a temporary concentration. In that quarter, many units of the SU7 Ultra and other premium configurations. The first buyers (the biggest fans of the brand, those who wanted to be the first to drive a Xiaomi) ordered the most expensive versions. It’s not that Xiaomi has fooled anyone, it’s the natural dynamic of any technological launch. The early adopters They always buy the higher versions. The testmotto is to confuse that initial demand with sustained market demand. The 25.5% margin does not validate your business model, it only tells you that you have sold the right product to the right people at the right time. The question is what happens when those people run out. Lu Weibing, president of the group, made this clear in the presentation of results. It said auto margins will likely fall in 2026 due to “competitive factors and normalization of the product mix.” It’s careful business language, but lto translation is simple: When you’re done delivering premium configurations and have to sell entry-level versions to maintain volume, you’re going to find out how much it really costs to compete in this market. Apple experienced something similar with the first Apple Watch. The first few quarters showed spectacular margins, but those numbers reflected sales to enthusiasts willing to pay for novelty, not sustained demand from a mature category. They had to learn to sell beyond the circle of fans. The difference is that Apple was not competing in a market with structural overcapacity and price wars. Xiaomi yes. Xiaomi competes in a Chinese electric vehicle industry where overcapacity is systemicgovernment subsidies have an imminent expiration date and the competition is fierce. There is another detail that should worry: Xiaomi is delivering cars faster than it is selling them. They are consuming the backlog of accumulated orders at a rate that exceeds the entry of new orders. An optimized factory running at maximum capacity is impressive, but if demand is not growing at the same rate, you have built production capacity for a level of demand that you have not yet proven exists. What is coming in 2026 is a kind of convergence of pressures: The portfolio of premium configurations will be exhausted. Subsidies will disappear. And security regulations will be tightened. Xiaomi will have to demonstrate that it can be profitable by selling cheaper cars, without public aid and meeting stricter standards. It is the moment when companies that built a real business are separated from those that surfed favorable temporary conditions. The trap of early profitability is not that the numbers are false. It’s that they make you believe that you have solved the problem when you have only optimized for the easier phase. The real test of Xiaomi Auto is not whether it can make quality cars (it has already proven this) but whether it can build a car business that works when the novelty wears off and it has to compete car for car with rivals that cannot afford to lose. That answer is not in the third quarter report. It’s coming. In Xataka | Xiaomi is no longer a brand: there are several brands fighting over the same logo Featured image | Xiaomi

everything that changes (and what doesn’t) in transfers with this system

There has been a lot of talk about the possible changes that were going to be introduced in payments with Bizum, especially due to the important changes in tax regulations of digital payments. Therefore, we are going to take the opportunity to clarify what are the changes for 2026 in payments on this platform. What we are going to do is give you a list of keys to take into account when using Bizum during 2026. We will confirm something important that does not change, but also everything that is going to change. The changes are to combat fraud in economic activity, but not to control. Changes in Bizum in 2026 Let’s go with the list of news to take into account about Bizum, with what changes but also some things that don’t change although there have been rumors about it. We start with what remains the same, and then we continue with the things that do change. Payments to friends and family are not monitored: Let’s start with the biggest concern. The Treasury has confirmed that it will not monitor payments between friends and family. Come on, without changes between movements between individuals (C2C). Come on, if you use Bizum to pay for a dinner, send a small amount of money or pay for birthday gifts, NOTHING CHANGES. Limit of 10,000 euros. Although not all changes will be monitored, there is a limit to take into account. If you pay more than 10,000 euros per year with Bizum, then you will have to declare them. We could say that it is the spending ceiling. If you do not exceed this amount, you do not have to declare anything. If you exceed 10,000 euros: Whether you exceed 10,000 euros per year with Bizum or if you move more than 25,000 euros with your card, then your bank will have the obligation to report the movements to the Treasury to prevent money laundering. Changes for companies and self-employed workers: From now on, if you use Bizum as a tool for professional collections, entities must report your movements monthly. The limit of 3,000 euros has been eliminated, and is now reported from the first euro. Be careful if you are an individual and receive recurring payments for undeclared work, you could be caught. Bizum to other European countries: Thanks to the alliance with EuroPA and EPI Companyit is expected that in the summer of 2026, starting in the summer you could start sending Bizums to other European countries such as Germany, France or Italy. Payments in dataphones: Bizum wants to allow you to pay with ATMs without needing a physical card or NFC, just with your phone number. There is no date for this yet. In Xataka Basics | Free immediate transfers from banks: what has changed and differences with Bizum

Europe has experienced its cleanest electric Christmas. The problem is what comes next

Europe has just said goodbye to the “cleanest” Christmas in its recent history in electrical terms, but the sector’s toast has been bittersweet. While families celebrated the holidays with electricity prices at a minimum, in the offices of regulators and analysis centers a very different scenario was already being drawn for the near future. We have the sun, we have the wind and we have broken production records, but the system shows signs of exhaustion. The success of this Christmas is, in reality, a reminder of the paradox that the continent is experiencing: we have never produced so much clean energy and, yet, the specter of gas, the saturation of the networks and an imminent rise in regulated costs threaten to spoil the party from 2026. The milestones of December. The fourth week of December 2025 will be recorded as an oasis of low prices. According to data from AleaSoft Energy Forecastingthe prices of the main European electricity markets fell significantly, with weekly averages below €85/MWh. In the Iberian Peninsula, the MIBEL market led this trend with a drop of 20%, the largest percentage decrease on the continent. This phenomenon, dubbed by analysts as the “Christmas effect”, is due to the combination of lower demand due to the festive break and a massive increase in wind and solar production, which put downward pressure on prices across almost the entire continent. The deployment of clean energies. As the report detailssolar photovoltaic production increased by 48% in Portugal and 21% in Spain during the week of December 22. This push was not exclusive to the peninsula: Germany, Italy and France set new historical highs for photovoltaic production for a day in December (Germany generated 87 GWh on the 25th). For its part, wind production maintained its upward trend, rising by 80% in Italy and 21% in Spain. According to the monthly report of OMIEthis force of the wind had already been brewing since November, the month in which wind energy reached a market share of 39.7% in the Spanish system. Abundance vs. rigidity. Despite these records, the transition faces critical obstacles: the disconnection between generation and the capacity to absorb it. According to AleaSoft forecastsAlthough solar production continues to grow, the European grid shows signs of saturation as demand falls. The technical problem is that, at times of maximum solar production and low demand, the system has nowhere to store the surplus. This forces prices collapse non-structurallywhich in the long term puts the profitability of new investments in check. Furthermore, added to this is a fiscal anomaly since in much of Europe, electricity is still burdened with tolls and taxes that make it up to three times more expensive than gas for the end user, slowing down the adoption of efficient technologies. like heat pumps. The Spanish case: the danger of bottlenecks. In Spain, this situation is especially delicate. The country has converted in a “case study on the dangers of saturation.” The lack of investment in networks (only 30 cents for every euro invested in renewables) has caused the curtailment —clean energy that is wasted because the grid cannot transport it—has tripled. The example most critical is Asturias. The network in the central Asturian area is at the technical limit; No more storage projects or new industry can be connected because the cables and transformers cannot support any more load. Furthermore, to avoid blackouts, Red Eléctrica operates in “reinforced mode”activating expensive gas plants to stabilize the tension, an extra cost that ends up in the citizens’ bill. A structural January slope. This Christmas’s price relief could be temporary. AleaSoft Energy Forecasting warns that future of CO2 have reached their highest closing prices since October 2024 (above €88/t), and TTF gas remains stressed due to low temperatures and European reserves below 65%. And in Spain we have to add the regulatory horizon of 2026. As we have detailedthe largest simultaneous increase in fixed costs in years is expected: transport tolls will rise by 12.1% and government charges by 10.5%. There is a real risk of returning to the tariff deficit if electricity demand does not grow as much as the Government expects, which would generate new structural debt in the system. The challenge of not dying of success. The European energy transition has shown that it can expel fossil fuels in certain days. However, this triumph has collided with an insurmountable physical reality: obsolete networks and a cost structure that still penalizes electricity. Christmas 2025 has given us a green market, but the shadow of 2026 reminds us that it is not enough to fill the landscape with mirrors and windmills. Without a real commitment to batteries, a modernization of cables and a reform of regulated costs, the abundance of clean energy will remain a mirage that fades just before reaching our pockets. Image | freepik Xataka | 2026 has not yet started but it has already managed to produce the first bad news: the light goes up

The key is not to have a goal but a path

We face the end of the year and arrive at January full of energy and new purposes for the new year. I’m sorry to be a little “Grinch” in this matter, but the problem is that a large part of those purposes deflate a few weeks later, often before the end of February. Gyms and language academies are witnesses of this. How do those people who manage to maintain their goals for months and even years do it? The answer is that they do not depend on a heroic willpowerbut rather a system that turns purpose into a routine that you want to repeat. The data from a study carried out by researchers from the University of Stockholm and Linköping (Sweden) with 200 people leaves no room for doubt: 77% of the participants fulfilled their resolutions in the first week, 55% kept it a month later and only 40% of the participants remained faithful to their commitment after six months. Other analyzes show that up to 43% of resolutions have been broken by the second week of February. Why resolutions wither in February Like a deciduous tree, the motivational effect of New Year’s resolutions loses its initial momentum in a maximum of five weeks. Science speaks of “fresh start effect“, in which dates like January 1 act as a “clean slate”, a new stage that motivates us to initiate a change. That initial emotion serves as an initial impulse, but it is not enough when the novelty wears off and the daily routine returns. Many times, resolutions are seen as a test of willpower: if you stumble once, you feel like you have failed completely, and that brings guilt and abandonment. Studies at the University of Scranton indicate that 46% of people with a clear purpose feel successful after six months, but only 4% achieve it without setting that well-defined objective, which shows that having a clear goal helps, but it is not everything. A recent study from Cornell University conducted with 2,000 adults in the United States followed their New Year’s resolutions for a year and looked at whether the motivation to achieve them came from external reasons (extrinsic motivation) or because they really liked doing it every day (intrinsic motivation). On average, external motivation obtained higher scores (6.27 out of 7) than internal motivation (5.41 out of 7). That is, external factors had more direct impact about motivation than your own willpower. However, the Cornell researchers discovered something that did make a difference: internal motivation consistently predicted continuity success at all measurement points of the research year, while the external one did not have much influence. Those who completed their goal had 5.73 in internal motivation compared to 5.18 for those who did not. Each extra point increased the chances of success in the goal by 1.60 times. The important thing is not the destination, it is the path As and as I pointed out writer and leadership coach Tiffany Toombs on FastCompanythe most productive people do not see purpose as a fixed and distant goal, but as something flexible to create habits that fit into their daily lives and that work for them. pleasant to carry out. Instead of just obsessing about the bottom line, like “saving more money,” they look for small, daily actions that lead to an identity goal such as “becoming more responsible with money.” To help you on that path, James Clear, author of the bestselling ‘Atomic habits‘, gives some keys to convert those purposes into habits integrated into your routine daily that no longer require effort to make, but rather become almost a reward. For example, choose exercises in which, far from suffering, you have fun. You hate monotonous weights, so sign up for Zumba or a guided class, which will make you return to the gym with enthusiasm. If pedaling for a long time seems boring, put on a cool audiobook or a podcast while you train. The key, according to Clear, is finding the system that allows you maintain consistency through activators that lead you to fulfill that habit. The same applies to eating better or saving: integrating small changes into your daily life that provide you immediate satisfaction. If you have to use willpower, it means that you have not integrated enough incentives to turn that purpose into a routine and you are among that 43% who will abandon their purpose in mid-February. In Xataka | You don’t need more hours in the day. All you need is to understand how the brain works to work better with less. Image | Unsplash (Tim Mossholder)

How much are the discounts extended, what are the prices and how to renew

Let’s tell you How are Madrid transport passes in 2026?which since 2023 has been allowing citizens of the Community of Madrid to have discounts. Since then, these discounts have been extended for periods of 6 and 12 months, and will also reach 2026. We are going to start the article by telling you how long the transport pass discounts are extended, and then we will end by telling you what the quantities are. How much are transport passes extended? Discounts on the transport pass for the Community of Madrid they are expanded throughout 2026. Therefore, everything continues as beforeso that the price of public transport does not rise, and the same economic conditions of the last six months are maintained. What are the discounts The three types of free subscription continue that already existed, being for children up to 7 years old, people over 65, or the card for children under 14 years old. Furthermore, you also have 50% discounts on the Youth Subscriptionaimed at people between 15 and 25, who will be able to continue traveling throughout the region for a price of 10 euros per month. In addition to this, users between 26 and 64 years old They maintain their 40% discount. Besides, 10 trip vouchers also maintain the discount for the Metro, EMO and ML1. In short, the prices are like this: discounted price in 2026 undiscounted price in 2022 zone fertilizer 32.70 euros 54.60 euros zone b1 fertilizer 38.20 euros 63.70 euros zone b2 fertilizer 43.20 euros 72 euros zone b3 fertilizer 49.20 euros 82 euros zone c1 fertilizer 49.20 euros 82 euros zone c2 fertilizer 49.20 euros 82 euros senior citizen pass Free Free youth pass (15-26 years) 10 euros 20 euros subscription 7-14 years Free – children’s card (

Agentic AI was the new race for Big Tech and Meta was far behind. It has bought the company most capable of recovering

Meta has closed the purchase of manusa Singapore-based artificial intelligence startup, for more than $2 billion. Throughout this year, Meta has reinforced its AI operations by acquiring several companies focused on different specialties. In July bought Play AIfocused on voice with AI. In August acquired WaveFormsan audio-focused startup. And in September was done with Rivosa company specialized in the design of semiconductors and RISC-V chips. Manus’s is already the fourth major purchase this year, and it is his hope not to be diluted in the race to dominate AI when all this time he has focused his efforts on Llama and his open weights approach. Why it is important. The Agentic AI (agents capable of performing complex tasks with minimal human supervision) has long become the new battlefield for big technology companies. Although companies like Microsoft or OpenAI had sufficient resources to develop in this field, Meta needed to strengthen its position in this segment if it did not want to be left behind. Manus came to reach 100 million dollars in annual recurring revenue just eight months after its launch, which offers Meta a product that generates money right away, something not very common in this sector. What does Manus do? The startup rose to fame in March with a video demo that went viral, showing how its AI agent was able to produce detailed research reports, build custom web pages, filter job candidates, plan vacations, and analyze investment portfolios. All using AI models developed by companies such as Anthropic and Alibaba. At the time, Manus even claimed to surpass OpenAI’s Deep Research. Currently, the company has around 100 employees, mainly in Singapore, offers subscriptions of $20 to $200 per month and already has a user base of millions. Initial success. Manus emerged a few months after the debut of DeepSeekthe Chinese model that shook the foundations of the industry due to its capabilities supposedly developed with less computing power than its American rivals. Just like account WSJ, the startup secured a $75 million funding round led by Benchmark in April, which valued the company at $500 million. Among its investors are firms such as Tencent, ZhenFund or HSG. Untying ties in China. The parent company behind Manus, Butterfly Effect, was founded in 2022 in Beijing by two Chinese entrepreneurs, including its CEO Xiao Hong, known as ‘Red’. Although most of its researchers and engineers were located in China, Manus launched outside the country because it used American AI models that are not available there. Shortly after securing its investment with Benchmark, the company officially moved its headquarters to Singapore. According to account WSJ, Manus has ruled out developing a version for the Chinese market. Goal declared to Nikkei Asia that, following the acquisition, Manus will have no ties to Chinese investors and will no longer operate in China. All existing investors have been excluded from the operation, according to they count from Bloomberg. What’s coming now? Meta plans to keep Manus running independently while integrating its agents into Facebook, Instagram and WhatsApp, platforms where Meta AI is available. According to WSJManus CEO Xiao Hong will report directly to Javier Olivan, Meta’s chief operating officer. “Joining Meta allows us to build on a stronger, more sustainable foundation without changing how Manus works or how decisions are made,” Xiao stated in the official announcement. No return guarantees. Mark Zuckerberg continues his mission to prove that AI can deliver tangible returns. Goal plans to spend $600 billion in American infrastructure over the next three years, much of it related to AI. Just like assures Bloomberg, it is an amount that causes some skepticism in some investors, since there are no guarantees that this expense will generate significant income soon. Cover image | TechCrunch In Xataka | NVIDIA has paid $20 billion to “license” Groq’s technology. He actually bought it

Imoo turned the children’s smartwatch into its own genre. Now all the parents who bought it are stuck

According to CounterPoint Research estimate for the global smartwatch market in 2025… Apple grew 12%. Samsung fell 6%. Imoo grew by 17%. Action replay: A Chinese brand that exclusively sells children’s watches is growing more than Appleand definitely more than Samsung, which is going down. Imoo, what The year has already started growing in quotaalready has 7% of the global smartwatch market. And it doesn’t really compete against the Apple Watch Ultra or the current Galaxy Watch: compete against the anguish of not knowing where your child is when he leaves school. Or rather: against the fear of not knowing if one day something happens. Counterpoint Research projects that the global smartwatch market will grow 7% in 2025 after first falling in 2024. That rebound is partly explained by Apple launching the cheap SE 3 and recovering after seven consecutive quarters of declines. But there is another factor: China went from 25% global share in 2024 to 31% in 2025. And within that jump, Imoo has a specific role that perhaps we are not looking at closely enough. Huawei is reinforcing its focus on health and sports, Apple maintains its inertia, Xiaomi focuses on the watch as part of a domestic ecosystem… and Imoo has turned parental fear into a product category. Their watches have GPS, calls, SOS button or alerts when the child leaves an area geofenced by his parents. As a watch it is not very smart and perhaps fits better in the category of surveillance and emergency aid device. Imoo hasn’t invented parental fear, but it has built a great machine to monetize it. Besides, It is a device that creates functional dependency: Once a parent puts it on their child’s wrist, they get used to the peace of mind it provides. So it is difficult not to renew it when the child stamps it or when it becomes obsolete. This success of Imoo goes beyond technology: when you grow 17% a year selling this type of watches, you do not measure adoption, but rather the number of parents who have decided that the anxiety that would cause them not knowing where their child is (understandable, of course) is worse than the inconvenience of constantly tracking them. Once you cross that threshold, there is no turning back. Previous generations had opaque spacesmoments of disappearance for a few hours before returning to dinner. These spaces are closed with this type of products, colorful and gamified, with a branding questionable but an unquestionable commercial success. Parents do not feel that they “control”, but rather that they protect. And kids don’t feel tracked, at least until they get acne and the bomb goes off, until then they just feel like they have a cool watch. And there is an advantage for parents: if suddenly almost all of your child’s classmates have one, the fact that your child does not have one becomes an anomaly. Imoo’s 7% share (and counting) measures how many children are growing up knowing that their parents can track them at any time. It measures a generation that normalizes permanent connectivity as a default state from the age of six. Counterpoint speaks of the smart watch market with “China-driven growth” and “different strategies to sustain the engagement of the consumer”, but it does not mention that One of those strategies is to redefine a part of childhood. The next son will also wear the watch. And the next one too. Imoo doesn’t need to grow faster than Apple to win. It just requires that each generation of parents find it more unthinkable than the previous one to leave a child unaccounted for. In Xataka | After almost a decade with the Apple Watch, I have switched to a Garmin. And I understood what I was missing Featured image | Xataka

The most farmed animal on the planet is not chickens, pigs, cows or fish: it is prawns.

Christmas is a time of carols, millions of led lightsnougats, empachos and a particular culinary ‘lore’ in which prawns and prawns are not usually missing. If tomorrow you have the opportunity to taste them during New Year’s Eve dinner, think about the following: what you have before you, on the plate, They are unique animals for humanity. And they are for a very simple reason. There is no other species that we raise more massively, not even chickens. There are those who estimate that approximately 51% of all animals What we have on ‘farms’ are precisely decapods, especially prawns. Prawns galore. If these days (lucky you) you have the opportunity to enjoy a good tray of prawns you should know a couple of things. The first one there are two typesdepending on their origin: there are wild prawns, caught in the ocean and the coasts; and those from aquaculture, which come from specialized farms and play a crucial role to supply the market. These fish farms are also interesting for another reason: they represent the largest farms in the world, at least if we are based on the number of living animals they contain. There are many (many) more breeding animals in them than in farms specializing in chickens, pigs, cows or even insects and fish. Click on the image to go to the tweet. But are there so many? This is what he suggests a study from 2023 that a few months ago rescued in Asterisk Magazine Andrés Jiménez Zorrilla, former investment expert and co-founder of Shrimp Welfare Project (SWP), an organization dedicated precisely to promoting more ethical decapod breeding practices. The report estimates that the planet’s fish farms usually host around 230 billion of these creatures at any given time. To be more precise, between 150,000 and 370,000 million, which exceeds any other known farm animal estimate. Even, the authors clarify, insects. “440 billion (300-620 billion) farmed shrimp are slaughtered each year, far exceeding the number of the most numerous farmed vertebrates used for food production, such as fish and chickens,” specify the articlesigned by Daniela R. Waldhorn and Elisa Autric and published in August 2023 by Rethink Priorites. The photo is completed with the specimens that arrive our months from fishing at sea. Are there more figures? Yes. And they are striking. Although both authors acknowledge that today there is only “partial data”, there are studies that indicate that every year hundreds of thousands of decapods are grown in fish farms on the planet, especially prawns and shrimp, which represent more than 80% of the total. In their report (in English) Waldhorn and Autric generally speak of “shrimp”but when delving into the problems surrounding the aquaculture of these species, both authors provide some extra detail. For example, when listing the species with the highest number of deaths, they specifically cite the P. vannamei and P. monodon. The most correct In Spanish it is to speak of “prawns”, rather than “prawns”. A percentage: 51%. The figures for the aquaculture industry are overwhelming, but they are better understood when compared to those of other sectors dedicated to raising animals in captivity for consumption. Jiménez Zorrilla points out that, in generalregardless of the moment, prawns represent 51% of the total number of animals raised on farms. They are followed at a considerable distance by fish (23%), insects (19%), chickens (7%) and pigs and other livestock (< 1%). Translated into figures, this means that compared to the 230 billion shrimp and prawns that (on average) live in fish farms, there are ‘only’ 779 million pigs and 1.55 billion cattle, 33 billion chickens and 125 billion farmed fish. In case the data were not clear in itself, the activist points out that every year 440,000 million of these decapods are slaughtered for consumption, “more than four times the number of humans who have walked the Earth.” Why is it important? Because Jiménez Zorrilla, like Wadhorn and Austric in their day, do not limit themselves to probing the size of the industry. Its objective is not so much to answer the question of how many shrimp live in the world’s farmers as to draw attention to the conditions in which they develop. “The problem is larger in scale than that of insect farming, fishing or any vertebrate for human consumption,” researchers warn. “If these animals are sentient, current commercial practices pose serious welfare risks during cultivation, handling, sale and slaughter.” Image| Kawê Rodrigues (Unsplash) Via | DAP In Xataka | Prawns, prawns, shrimp, prawns and carabineros: how they differ and which ones are better

They no longer trust their own debt

Deutsche Bank and Morgan Stanley are looking for ways to protect themselves from the debt they have extended to build AI data centers, according to Ed Zitron’s latest report in which he makes a notable criticism of the boom of AI and the stock market in which debt and complacent analysis are inflating an unsustainable bubble, according to their analysis. Both banks are contemplating “synthetic risk transfers.” It is a mechanism that allows the credit exposure of loans to be sold to other investors while keeping the loans on their books. Deutsche Bank even is considering betting short against actions related to AI. Why is it important. These movements clearly show a certain distrust in the economic viability of the infrastructure they are financing. Morgan Stanley, Deutsche Bank, Goldman Sachs, JP Morgan and MUFG have participated in the world’s largest data center financing transactions, including various loans to CoreWeave and the stargate projectsbut now they are looking to reduce their exposure to those same assets. The figures. At least $178.5 billion in data center financing was closed in the United States alone in 2025, almost triple the amount in 2024. CoreWeave, one of the largest operators, carries $25 billion in debt on estimated revenues of $5.35 billion, losing hundreds of millions each quarter. The context. AI data centers are powered by a circular financing model: They sign contracts with their clients before having the physical infrastructure. They use these contracts as collateral to obtain bank debt. They buy NVIDIA GPUs and build facilities that take between one and three years to be operational. Only then do they start generating monthly income. If construction is delayed or the client cannot pay, the loan is up in the air. Between the lines. The banks that have fueled the bubble are now covering their backs. Yes, but. Banks argue that these hedges are normal risk management practices. The problem is that they are hedging themselves against loans that they themselves structured and approved, many of them to clients whose ability to pay is, at the very least, uncertain. CoreWeave has offered OpenAI net 360 payment terms (one year from invoice to settle), depending on your loan agreement. If OpenAI, which needs to raise $100 billion to continue operating, decides not to pay, CoreWeave automatically defaults on its credit obligations. And CoreWeave is probably the best-funded operator in the IT industry. neoclouds. The money trail. NVIDIA announced in October that would guarantee $860 million in lease obligations from a partner in exchange for warrantswith 470 million deposited in a guarantee account. CoreWeave’s third-quarter balance sheet includes a “non-current restricted cash” item of $477.5 million. NVIDIA also signed a 6.3 billion contract with CoreWeave to buy the capacity that CoreWeave fails to sell until 2032. Go deeper. The banks that are hedging their bets are the same ones that have funded most of the global AI infrastructure. They are not selling the risk of any loan, but the risk of data centers that may never turn on, or that if they do, will serve customers who burn billions without generating profits. When the financiers of boom show signs of having stopped believing in boomit is worth paying attention. In Xataka | We have reached a point where not even the CEOs of Google or Microsoft deny that we have an AI bubble Featured image | İsmail Enes Ayhan

why OpenAI is installing Boeing 747 engines in its data farms

Just three years ago, Blake Scholl, CEO of aviation company Boom Supersonic, had a linear business plan: He would first build the supersonic plane of the future and, much later, retrofit its engines to generate power. However, a phone call changed the order of factors and revealed the desperation of the technology industry. On the other end of the line was Sam Altman. The OpenAI CEO’s message was a direct plea: “Please, please, please get us something.” Altman wasn’t looking for plane tickets; I was looking for electrical power. This anecdote, reported to the Financial Timessummarizes the state of emergency in the sector: artificial intelligence is advancing at breakneck speed, but it has hit the wall of physical infrastructure. While the AI evolves in monthspermits to connect to the electrical grid can take up to ten years in some regions. Faced with this paralysis, the industry has opted for “Plan B” which consists of bypassing the grid and manufacturing its own energy on site. The tall price of urgency. This strategic shift has profound consequences. The first is economic, the “delay” is expensive. According to BNP Paribas analystspower from a gas plant built for Meta in Ohio costs about $175 per megawatt hour, nearly double the average cost for an industrial customer. The second is environmental. Mark Dyson, from Rocky Mountain Institutewarns that the emissions of these plants are much worse than those of the general network, which combines efficient gas with renewables. Despite this, the urgency is such that the authorities are giving in. In Virginia, the world’s data center heartland, it is considering relaxing emissions rules to allow generators to run more frequently. Even polluting plants that were in retirement, like the Fisk plant in Chicagohave canceled their closure to feed the demand for AI. From the sky to the data center. The most surprising solution comes from aeronautical engineering through aeroderivative turbines. The ProEnergy Company are buying motor cores CF6-80C2 of the iconic Boeing 747 to rebuild them as ground power units. A single one of these turbines generates 48 megawatts, enough for a city of 40,000 homes. It is not an isolated case. GE Vernova already supplies this technology for the gigantic Stargate (OpenAI/Microsoft) data center in Texas. Blake Scholl himself confirmed that it will sell Crusoe turbines “practically identical” to those of his supersonic planes to finance his aeronautical project. The return of diesel. Beyond aviation turbines, the sector is rescuing the most reviled fuel: diesel. The manufacturer Cummins has already sold 39 gigawatts of energy to data centers, doubling their capacity this year. What was once emergency equipment for power outages is now in demand as a primary energy source. The situation has escalated to the US Government. Secretary of Energy, Chris Wright, suggested on Fox News an almost war economy measure: requisition the backup generators from data centers or large stores like Walmart to turn them over to the network when the general system falters. The ignored alternative: Is smoke necessary? Not everyone agrees that the return to the fossil is inevitable. A study by researchers at Stripe, Paces and Scale Microgrids maintains that the future is in “off-grid” solar microgrids. According to their calculations, a system with 44% solar energy is already as cheap as gas, and one with 90% renewables would surpass nuclear projects in profitability. The advantage is speed since these solar farms can be built in less than two years in desert areas from Texas or Arizona. Giants like Google have taken note, buying the electric company Intersect Power for 4.75 billion dollars to protect its clean supply and not depend on the network. However, the majority industry prefers diesel and known gas due to a matter of technical inertia, due to the prosaic fear that the cloud will go out if the sun does not shine. AI goes physical. The industry finds itself in a technical paradox. To power the most advanced software on the planet, big technology companies are resurrecting combustion engines and burning fossil fuels on a massive scale. Although these “bridge turbines” allow AI to continue growing today, experts cited by the Financial Times They warn that this fever could cool as the tech giants reduce their capital spending. For now, the cloud has had to come down to earth. The future of artificial intelligence, ironically, depends not only on brilliant code, but on who controls the underground and who manages to turn on enough “plugs” so that the greatest technological revolution of our era is not left in the dark. Image | freepik and Harpagornis Xataka | The exorbitant deployment of data centers for AI has a new problem: salt caverns

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.