bubble

One piece of information perfectly summarizes the book bubble in Spain: 95% of those published do not recover costs

by

The Spanish publishing sector closed 2025 with historic figures: 76 million printed books sold and a turnover that was close to 1,250 million euros. A record. The cold water came a few weeks later, at the annual booksellers’ conference, where it was certified that almost half of the titles available on the shelves had sold absolutely nothing. Who says so. The data was presented by CEGAL, the Spanish Confederation of Guilds and Associations of Booksellers, in theXXVII Congress of Bookstores held in Valencia in February 2026and has been extracted from LibriRed, the confederation’s own tool, which monitors in real time the final sales in more than 1,000 independent bookstores and chains throughout the country. The figure includes novels, essays and comics, both new releases and catalog contents, but (importantly, we are talking about physical bookstores) Amazon and school textbooks are excluded. The specific data. They are that revealing: 13.2% of the titles sell a copy throughout the year. 19.4% do not exceed ten. Only 4.5% of the books that reach bookstores reach 100 copies sold, a threshold that often does not even cover the costs of a launch. In other words, 95.5% of the books available in Spanish bookstores do not have the slightest economic impact on the publishing industry, not to mention that they are directly deficient. In Xataka If you hate justified text, we have good news: you’re most likely right You bill more, you sell the same. This is the paradox that the CgK consultancy put on the table with its Book Market Data 2025 report: The sector reached close to €1,250 million in turnover in 2025, 4% more than the previous year, which represents a historical record. However, total units sold rose just 0.2%, and novelty units sold on average 2% less per title than in 2024. Further analysis of the report They spoke of a statistical illusion typical of inflationary markets, because what has actually grown is the average price of the book. And this benefits the large groups, with catalogs in high rotation. Why is this happening? In its analysis of the Cedal report, El País collected statements from editors such as Enrique Redel, from Impedimenta, who affirms that there are titles that are not published to sell, but to take up space on the shelves, especially by large groups. The strategy is to publish many titles assuming that most will fail, hoping that one or two best sellers compensate for the losses of the rest. More than 90,000 books are published each year in Spain, about 240 newspapers, and theReturn rates range between 30% and 40%. It is a feverish cycle of full-speed rotation, paradoxically inconsistent with the calmest of cultural activities. {“videoId”:”x7zmsee”,”autoplay”:false,”title”:”11 WEBSITES to DOWNLOAD FREE EBOOKS for your KINDLE Xataka TV”, “tag”:”Kindle”, “duration”:”321″} Who can afford it. The two large publishing groups, Penguin Random House and Planeta, in whose shadow it has been for decades the Spanish industry, and which account for more than 40% of the copies sold in bookstores. Fleeing this suffocating single direction are independent bookstores, which offer more than twice the variety of titles than the large chains: more than 525,000 titles compared to 229,633. In this way, visibility is concentrated in a few titles that rotate for a longer period of time, while the rest are buried in excessive catalogs. Some reasons. When looking for factors that exacerbate this situation (the two large groups can suffocate the market with their continuous rotation, but there must be more compelling reasons for so few sales of so many titles), CEGAL points to self-publishing: publishing has been democratized, but the reader’s attention has not. A book without a publisher behind it, without distribution, without promotion and without prior prescription is born practically invisible to the market, and it is normal that many of these launches do not sell anything. ¿AI provides tools to multiply these throws effortlessly? The percentages skyrocket exponentially. In Xataka They are not your imagination: the best-selling books are increasingly simpler and contain less elaborate sentences The difference with other cultural media is in the abundance of second chances. A film that does not perform in theaters can recover the investment in streaming, where consumption already rivals that of theaters. The book that does not sell in its first weeks on the shelf returns to the publisher, returns to bookstores in negligible quantities and is often physically destroyed after months languishing in warehouses. Perhaps finding new ways of dissemination and renewed lives for books would be the solution to this veritable overdose of books without readers. Header | Photo ofBree AnneinUnsplash (function() { window._JS_MODULES = window._JS_MODULES || {}; var headElement = document.getElementsByTagName(‘head’)(0); if (_JS_MODULES.instagram) { var instagramScript = document.createElement(‘script’); instagramScript.src=”https://platform.instagram.com/en_US/embeds.js”; instagramScript.async = true; instagramScript.defer = true; headElement.appendChild(instagramScript); – The news One piece of information perfectly summarizes the book bubble in Spain: 95% of those published do not recover costs was originally published in Xataka by John Tones .

The “bubble” of the eclipse parties reaches Spain and Iceland

by

Next August 12 a long-awaited phenomenon will take place: the first of the three eclipses that make up what many have already dubbed the Iberian trio. In three consecutive years, a solar eclipse will be seen from Spain. Those of 2026 and 2027 will be total eclipses, while that of 2028 will be annular. Be that as it may, it is a statistically improbable event, which excites both astronomy lovers and the general population. That’s why many people have chosen to see it at such imaginative events as music and art festivals designed around the eclipse. The eclipses will not be seen equally in all parts of Spain. For example, in 2026 totality will only be reached in a strip that goes from the north of Galicia to almost all of the Balearic Islands, passing through Asturias, Cantabria, La Rioja, the north of Castilla y León and the Valencian Community, La Rioja, and a part of the Basque Country, Navarra, Madrid, Aragon, Catalonia and Castilla la Mancha. In the rest of Spain it will be seen only partially. Therefore, since the famous sunset of solar eclipses will only be experienced in places where totality is reached, Many of the lucky localities are already preparing events to welcome the eclipse. These are events for the local population, but also for tourists. The emptied Spain will be less empty than ever and tourists will forget about the most typical destinations for a few days to travel to places to which they perhaps would not have traveled under other circumstances. Many hotels have been there for months due to the influx of people who will travel to observe the eclipse outdoors, without many more pretensions. However, there are also those who plan to attend what is possibly the most special festival of their lives. The most unexpected festivals around the eclipse The 2026 eclipse won’t be too long. In Spain, the places where totality lasts the longest They will barely enjoy more than a minute of darkness. Still, multi-day festivals have been planned, with musical performances, scientific talks, workshops and, of course, viewing the eclipse at the appropriate time. These are some of the most striking. Eclipse Festival, in Prades. In this town of Tarragona you will only enjoy 51 seconds of totality. Even so, between August 10 and 13 its Astronomical Park will celebrate a festival with music, workshops, conferences, observations, shows, telescopes and a planetarium. It will also be an ideal time to observe the Perseids. EclipsaFest, in Aldea Santillana. This small village in the also small town of Manjirón, in Madrid, will have 1 minute and 15 seconds of totality. In your case it will be a simpler observation, without the rest of the incentives of a festival, but it will very big. It will only be held on August 12 and admission will cost 147 euros for adults and 117 for children, with a welcome pack that includes glasses and the possibility of guided observation. Playabout Radio Festivalin Ibiza. In Ibiza they will have 1 minute and 6 seconds of eclipse and They will celebrate it as they know best. Accompanied by house and techno music, which will last from August 10 to 14. Umbra Festival, in Agolada. This town in Pontevedra will take advantage of the Brocos Reservoir, which is actually a reservoir, to celebrate a 3 day festival in which visitors will enjoy techno and minimalist music. Of course, also the eclipse, although in this case totality will be fleeting, lasting only 34 seconds. Admission costs 62 euros. Iberia Eclipse Festival, in Vinuesa. In Soria, next to the Duero River, this festival will be celebrated which will consist of four scenarios spread across the hillside and the forest, as well as a camping area and pre-installed tents. For 5 days, attendees, who will have paid an entrance fee of 240 euros, will enjoy music, workshops and a wellness area, which will include yoga, meditation, massages, swimming experiences in nature and art exhibitions. Astral Plane, at the La Pinilla Mountain Station. In this Segovian station you will enjoy the minute and a half of totality in the middle of a set by Detroit techno artist Kevin Saunderson. Admission costs 175 euros. Sizigia Eclipse Meeting, in Alcalá de Gurrea. This town in Huesca has also chosen a reservoir to observe the eclipse in its vicinity, whose totality will last only 40 seconds. Even so, attendees will enjoy 5 days of underground music, among other activities. Admission costs 262 euros, and with an extra fee you can add accommodation in a tipi camp. Also in Iceland Iceland will have its own eclipse viewing events. There, in fact, there will be points where totality can be seen for more than 2 minutes. But perhaps because the weather is less favorable or because Icelanders have a less festive spirit, there will not be as many options to choose from. Some of the most interesting will be the hellissander festivalwhich will include live music and TED talks, and the Grindavíkurbær Blue Lagoonwhich will be held in a spa. Attendees will be able to see the eclipse in an idyllic setting, but they will have to pay 750 euros. The price includes a two-course meal, two drinks, a robe, towel and glasses to view the eclipse. Seeing this, Spanish festivals even seem cheap. Image | NASA and Alfonso Scarpa In Xataka | The trio of eclipses that await Spain on the horizon: an unprecedented and historic chain between 2026 and 2028

We thought we had an AI bubble. There are powerful arguments that indicate that we were wrong

by

You either love AI or hate it. Either you are a (deluded?) optimist, or you are on the bandwagon of skeptics and bet due to an imminent puncture of that AI bubble that everyone talks about. The well-known analyst Ben Thompson has been in the second group for some time and stated that in fact we were in a “good” bubble and beneficial even if it bursts. The annual NVIDIA conference a few days ago has made him change his position, and for him there is no bubble. It doesn’t have just one argument, but three. Or rather, three jumps. The first jump: ChatGPT. The launch of ChatGPT in November 2022 was an eye-opener and demonstrated what generative AI could do. That first model, yes, had two serious problems. The first, that he was frequently wrong. The second, that when I didn’t know something, he invented it and hallucinated with astonishing security. That made ChatGPT something awesome but unreliable, like a cool toy that needs constant user supervision to be truly useful. The second leap: reasoning. Almost two years later, another unique revolution occurred in the field of generative AI. In September 2024 OpenAI launched its o1 model and with it there was a spectacular novelty. For the first time, the model did not simply blurt out the first thing that came to mind: he reasoned about his answer before giving it, evaluated whether it was correct, and considered alternatives. The result was an AI significantly more reliable and, therefore, more useful. The price? More computing. AI models with the ability to “reason” consume many more tokens than those that respond directly, and that triggered demand for infrastructure. Or what is the same: data centers. The third jump: the agents. These two revolutions have been joined by the third, that of AI agents. Claude Code and Codex at the end of 2025 showed that AI agents were no longer a promise and became something that really worked. From then on it is possible to give them instructions so that they can then start executing nested tasks that can keep them working for hours. These agents verify their own results and correct errors without the human having to intervene. The difference with what we had before is notable, but it also dismantles the bubble theory. Bubble? In a bubble, Thompson explained, investment exceeds real demand. However, in his opinion, the opposite is true here, because each hyperscaler—Microsoft, Google, Amazon, Meta—has made it clear that the computing demand is surpassing them, and to solve it they are all announcing astronomical investments in AI data centers. These investments exceed market expectations, but not those of these companies, which like Thompson are clear that in reality the demand is going to end up being so enormous that the current infrastructure will fall very short. Millions of users are not needed. Even more striking in this analysis is another nuance that this analyst points to. Chatbots were supposed to need mass adoption to generate economic impact, but on the other side we have agents, who don’t have that requirement. A single person can control thousands (millions?) of agents simultaneously, creating complex tasks. That means it doesn’t take everyone to use AI for computing demand to skyrocket: enough people just need to use it as they are likely to use it: to create those “one-person businesses” where one human being has thousands of AI employees. Companies will pay. The reality is that the vast majority of consumers are not going to want to pay for AI. Companies do, because they pay for productivity and AI seems start fulfilling that promise. But the argument goes beyond cost savings: agents not only make the work that humans do more efficient, but they allow a small group of people with a clear strategic vision to execute it on a scale that previously required hundreds of employees who also had to be coordinated. Large companies have been adding layers of management necessary to scale for decades, but all that hierarchy disappears with agents. But. This analyst is also clear that the wave of layoffs is going to be increasingly evident and it is evident that AI is going to have a clear impact. However, he explains that many of these current layoffs correspond more to the overemployment experienced with the COVID-19 pandemic. What will happen now is that companies will no longer wonder if they hired too much for the “pre-AI” world, but rather if they hired too much for the “post-AI” world. In fact, those that don’t ask will probably end up competing with smaller rivals, built from the ground up with AI and with radically lower cost structures. For him two things are clear. The first is that the demand for computing will not stop growing. The second, that the bubble, if it exists —and according to him, the answer is that he doesn’t—, it’s not going to explode. In Xataka | His dog had cancer, his vets had no solutions and he found an mRNA vaccine elsewhere: ChatGPT

If anyone was waiting for the AI ​​bubble to burst, NVIDIA’s results have a message: sit tight

by

NVIDIA just published your results of the fourth quarter of its last fiscal year and has left Wall Street speechless. Revenues of $68.1 billion, a net profit that almost doubles that of the same period of the previous year, and a forecast for the following quarter that has far exceeded analysts’ expectations. And all this in a turbulent context where more efficient models and other alternatives are beginning to appear. The crash of DeepSeek is far away, and the demand for chips does not slow down. We tell you the numbers in detail. In case your position was not clear. Only a handful of companies in history have exceeded $100 billion in annual profit. Alphabet, Microsoft and Apple are in that club. NVIDIA has just joined them, with $120 billion in profits in the last twelve months, according to the report. The difference is speed: just three years ago, its annual profit was 4.4 billion. We can say with certainty that no technology company has ever grown so quickly on that scale. AI, and more AI. The engine that has driven these profits is its data center business, which generated $62.3 billion in the quarter, 71% more than a year ago. Within that segment, if we focus on their Blackwell chips, they have gone from entering 32.6 billion to 51.3 billion, while the networks (NVLink, Spectrum-X and InfiniBand) grow from 3,000 to 11,000 million. Gross margin is 75%, and earnings per share nearly double to $1.76 in GAAP terms (which is the official rulebook that companies follow to demonstrate transparent accounting). What Jensen Huang says. “Without computing, there is no way to generate tokens. Without tokens, there is no way to grow revenue.”, counted directly the CEO of NVIDIA in the meeting with investors. Their thesis is that in the new AI economy, computing power directly equates to revenue for their customers. That is why the large cloud service providers (Google, Amazon, Microsoft, Meta) continue increasing your capex budgetswhich together will exceed 500,000 million dollars in 2026 to build AI data centers. And NVIDIA is the main beneficiary of that expense. What DeepSeek has not broken, but accelerated. At the beginning of 2025, the emergence of the Chinese DeepSeek model generated an unprecedented tremor in the markets, leaving a simple question in our minds: if AI becomes more efficient, why do we need so many chips? The answer from NVIDIA’s results is that efficiency does not reduce infrastructure demand, it multiplies it. Every improvement in inference efficiency lowers the cost per token, encouraging more companies to deploy more AI applications, which in turn requires more compute. It’s like Jevons’ paradox, but applied to AI: efficiency expands the market instead of contracting it. Agentic AI as the next catalyst. On the same call with investors and analysts, Huang stood out that “enterprise adoption of agents is skyrocketing.” AI agentsthese systems that make decisions and execute tasks autonomously, require many more inference cycles than chatbots. They are the next step in the AI ​​value chain, and NVIDIA is once again in a privileged position. Colette Kress, CFO of the company, confirmed In addition, the first samples of Vera Rubin, the next generation of chips that will arrive later this year, have already been sent. China and the competition. Not everything is green. NVIDIA acknowledged that its forecast for the next quarter ($78 billion) does not include computing revenue in China. The company has generated just about $60 million from H20 chips since the Trump administration reapproved some sales in August 2025, according to SEC filings, and has yet to earn revenue from the most recently approved H200. Regulatory uncertainty with Beijing remains a small China in Huang’s shoe. In parallel, competitors such as AMD, Broadcom or Google’s own custom chips (TPUs) are gaining ground. But the NVIDIA CEO remains focused on his vision. And according to pointed at the meeting: “Every company depends on software, and all software will depend on AI.” As long as this is fulfilled, everything indicates that NVIDIA will continue selling the blades and picks. Cover image | NVIDIA In Xataka | NVIDIA was founded by three engineers, but only Jensen Huang remains CEO: “I wish I had kept some shares”

OpenAI’s biggest fear is not that the bubble will burst. It’s just that I do it ahead of time

by

Sam Altman has admitted in an internal memo published by The Information that Google is catching up technologically with Gemini 3. That’s a real problem for OpenAI, but OpenAI’s real concern isn’t that. It’s just that he needs the party to last long enough to give him time to build his own infrastructure. Why is it important. OpenAI plans to burn more than $100 billion in the coming years pursuing AGI. But it is completely dependent on Microsoft for servers, NVIDIA for chips, and external investors for financing. Google, on the other hand, already has its own TPUs and generates 70 billion in free cash flow per year thanks to Search, YouTube and Google Cloud. If the music stops early, one survives and the other doesn’t. The paradox of timing. OpenAI faces a very peculiar race against time: If investment in AI slows in 2026 or 2027, it will have spent tens of billions but will not have completed its own infrastructure. You will remain tied to expensive suppliers. You will not be able to compete on costs with Google. Staying halfway is the worst possible scenario. Instead, if the bubble lasts until 2030 or beyond, OpenAI will probably have reached the threshold of self-sufficiency. It will have its own chips, its own data centers, economies of scale. It will be able to survive even when the investment tap is turned off. It’s like building a bridge: it doesn’t matter how much you’ve spent a lot. If you only get halfway, it’s of no use. The absence of moat. OpenAI cannot protect itself with sustainable technological advantage. In AI there are no defensive moats (moats) real. Every time OpenAI or any other lab makes a breakthrough, the rest replicate it within months. The only sustainable advantage OpenAI has left is cost. If you control your infrastructure, you can offer prices that no one else can match. If you do not control it, you become a dispensable intermediary between the end customer and whoever does have the chips and servers. The context of the memo. The document published by The Information reveals that Altman anticipated turbulence after the launch of Gemini 3. Google’s new model stands out precisely in the areas that generate the most revenue for OpenAI: automation of web design and programming. Altman acknowledged to his team that “Google has been doing an excellent job lately” and warned that he expects “the environment to be tough for a while.” But he urged them to stay focused on “achieving superintelligence”, admitting this would mean being left “temporarily behind in the current regime”. The figures. OpenAI went from almost non-existent revenue in 2022 to projecting 13 billion this year. It is one of the fastest business growth in history. But it plans to earn 200 billion in 2030. To achieve this, it will need to multiply its current income by 13 in less than five years. Meanwhile, it plans to spend $90 billion on R&D alone through 2030. That represents 45% of its projected revenue. Large technology companies allocate between 15% and 30% of their gross profit to research, not their total income. If OpenAI falls short of its billing goal, that percentage will be even higher. Yes, but. Google has structural advantages that are difficult to overcome: Generates a huge cash flow thanks to consolidated and very profitable products. You can afford to burn money on AI for years without too much trouble. And it already has its own infrastructure after a decade developing TPUs. OpenAI, on the other hand, lives off external funding. His recent agreement with Oracle to design data center components in the United States is an attempt to build that self-sufficiency. Altman presented it as “a step to ensure that the core technologies of the AI ​​era are built here.” At stake. OpenAI’s technological advantage over rivals such as Google and Anthropic has narrowed. Investors have sunk more than $60 billion into OpenAI, recently valuing it at $500 billion, betting that it will continue to dominate the market for AI that creates content and reasons like humans. That bet falters. Anthropic, founded four years ago by former OpenAI employees, is skyrocketing its valuation and aiming to generate more revenue than its former home selling AI to developers and companies. Their models specialize in generating computer code. And ChatGPT is still far ahead of Gemini in usage and revenue, but the gap is narrowing. Between the lines. Altman concluded his memo by acknowledging the pressure: “It sucks that we have to do so many hard things at the same time: the best research lab, the best AI infrastructure company, and the best AI platform/product company. But it’s our destiny in life. And I wouldn’t trade positions with any other company.” The question is not whether OpenAI can technically compete with Google. It’s whether you can hold on financially long enough to stop depending on others. Featured image | Xataka In Xataka | There is a generation working for free as a documentarian of their own life: they are not influencers but they act as if they were.

The AI ​​bubble is so obvious that not even Sundar Pichai or Satya Nadella make an effort to deny it

by

The thing about bubbles is that we are certain that there is one only when they burst. And with all this artificial intelligence, is talking a lot about whether or not there is one around this technology. Of course there are indicators that set off alarm bells, but the curious thing is that we would not have believed that two of the greatest exponents in contributing to the development of this technology would maintain reservations. And Sundar Pichai, for Google, and Satya Nadella, for Microsoft, have not made much effort to deny the doubts. Irrationality. Pichai declared to the BBC in an interview he noted “elements of irrationality” in the current AI market and warned that no company, including Google, will be immune if the bubble bursts. His words are especially striking because they come at a time when Alphabet shares have doubled in seven months, reaching a market capitalization of $3.5 trillion. The CEO compared the situation with the Internet bubble of the late 90s, recognizing that although there was excessive investment that ended in bankruptcies and layoffs, today no one questions the profound impact of the Internet. “I hope AI is the same. I think it’s both rational and there are elements of irrationality in a time like this,” he explained. When the numbers don’t add up. Skepticism is based on concrete data. OpenAI, Google’s most visible competitor in this field, has committed to spending $1.4 trillion in infrastructure for eight years while it expects to generate just $13 billion in revenue this year. Just like share In the Ars Technica media, Sam Altman himself, CEO of OpenAI, acknowledged to journalists in August that investors are “overly enthusiastic” about AI models and that “someone” will lose an “incredible amount of money.” Microsoft also shows the cards. For his part, Satya Nadella has been equally forthright about the current limitations of the sector. At the beginning of the year already pointed out to claim that a milestone has been achieved in AGI (general artificial intelligence) is “just hacking the tests without meaning”, downplaying the benchmarks that so much marketing generates. According to Nadella, the true metric of AI success should be reflected in countries’ gross domestic product: “When we say ‘this is like the industrial revolution,’ we should have that kind of growth that caused the industrial revolution,” he explained, referring to increases of 5-10% in GDP. That growth has not yet come. Jensen Huang says exactly the opposite. While Pichai and Nadella talk about irrationality, NVIDIA founder and CEO Jensen Huang has presented spectacular results in the third quarter and settled the debate in his own way. “There has been a lot of talk about an AI bubble. From our perspective, we see something very different,” he commented. NVIDIA reported revenue of $57 billion in its latest quarter, up 62% from a year earlier, with net profits of $32 billion. Its data center business has generated $51.2 billion, a record boosted by the sale of its Blackwell chips. According to Huang, sales of these GPUs are “skyrocketing” and cloud chips are out of stock. NVIDIA also projects a fourth quarter with revenues of $65 billion. AI still doesn’t make money. NVIDIA does make money, a lot of money, but He does it by selling the shovels during the gold rush. The vast majority of companies that develop large language models are losing money spectacularly. OpenAI is the most obvious examplebut not the only one. Microsoft, Amazon, Meta and Google they are allocating tens of billions of dollars to build data centers dedicated to AI in a colossal bet whose profitability is not guaranteed. For Nadella, what AI needs is something equivalent what Excel and email meant for the PC, that is, an app that makes the majority of users understand how to use AI. At that time we saw that the PC took a long time to find its place, especially until it reached mass adoption that transformed real processes. There are chips but there is no energy to power them. In addition to the profitability problem, there is an immediate physical limitation. Nadella revealed recently that the biggest obstacle is not the lack of chips, but the energy needed to power them. “If you can’t do something like that (supply enough power), you’re going to have a bunch of chips sitting around in inventory that you can’t plug in. In fact, that’s my problem right now: It’s not that I don’t have a sufficient supply of chips: it’s actually the fact that I don’t have places to plug them in,” he admitted. Microsoft, Google and other big technology companies are resorting to drastic solutions such as building their own small nuclear power plants (SMR reactors) to supply their future data centers. ARM CEO Rene Haas noted that energy needs could triplea challenge that calls into question the sustainability of the current expansion. Of course we don’t know how things are going to end, but no one doubts that we’re going to have a good time with it. Cover image | Microsoft and Bloomberg In Xataka | Gemini 3 promises more quality and precision than ever in its responses. The question is whether we will really notice the difference

NVIDIA and OpenAI know that the AI ​​bubble can burst in their faces. His solution: let dad pay for the state

by

Too big to fail or, in English, “too big to fail.” It is a theory of economics and finance which argues that certain corporations, especially banks, are so large and so interconnected that their failure would have catastrophic consequences for the global economy and therefore must be rescued by governments. The speech gained traction in the 2008 financial crisis and is beginning to sound again from the mouths of NVIDIA and OpenAI, no less. Government support. At an event of WSJSarah Friar, CFO of OpenAI, stated that the company will not go public in the short term (she says until at least 2027) and that its priority is growth and investment in R&D, above profitability. The most striking part of his speech was when he said that they hope that the government will support the financing of future agreements related to data centers. That OpenAI is burning astronomical amounts of money to lead the AI ​​race is something we have been discussing for a long timebut it is the first time that they directly appeal to the state to guarantee it. Shortly after, Friar collected cable in a post on LinkedIn: “OpenAI does not seek government support for our infrastructure commitments. I used the word ‘support’ and that confused the message,” but the seed was already planted. Depreciation. OpenAI is closing deals to secure computing capacity. We have seen it with his alliance with NVIDIAwith amdwith Broadcom and more recently with amazon. The complexity of the situation is that the depreciation rates of AI chips remain uncertain. As it says Washington Post’s Gerrit de Vynck in XOpenAI is going to need the best chips to be at the forefront of the AI ​​market, but financing this demand is not the same if the life cycle of the chips is seven years, as if it is only two years. The money is flowing, the question is for how long. In this uncertain scenario, government support would act as a safety net so that banks and private equity firms would feel more comfortable and continue releasing billions for OpenAI. China will win. NVIDIA is also appealing for government involvement in subtle ways. In a Financial Times event in London, Its CEO Jenshen Huang has warned that “China is going to win the AI ​​race.” Their arguments are that China has more flexible regulation and government subsidies for the energy your data centers needthat It is not little. This energy advantage allows China to compete even if they cannot buy NVIDIA’s most powerful chips. Huang doesn’t say it directly, but it is a clear wake-up call: either you subsidize the energy our data centers need or China will win. The fear. The question has been hanging over the air for a long time: Are we witnessing a new bubble? The investor Michael Burry thinks soand he is not just any investor, he was the one who made gold when the real estate bubble burst in 2008 (the movie ‘The Big Short’ is based on his story). The thing is, Burry just bet short against NVIDIA, which recently It was valued at 5 billion dollars. Fear of the bubble continues to grow, according to a Coatue report and the number of fund managers who believe we are in a bubble increased to 54% in October, up from 37% in July this year. 48% of the S&P 500 index corresponds to AI-related stocks. Fountain: Bianco Research Numbers. The fear is not at all unfounded and all you have to do is take a look at the numbers. Account Tomás Pueyo in Uncharted Territories that the economy should be in recession, but the numbers show the opposite and AI is behind this growth. The S&P 500 index is through the roof and 48% of this growth corresponds to AI-related stocks. The share price is far above what it was in the dotcom bustall with ridiculous benefits. And that’s not all, the economic growth of the United States in 2025 is due almost entirely to the construction of data centers for AI. According to the Economist Jason Furmanwithout taking data centers into account, the GDP of the United States would have grown only 0.1% in 2025. The creator of the newsletter Today in Tabs He gave a very graphic example: “Our economy could be reduced to three AI data centers in trench coats.” Tightrope. Returning to OpenAI, its financial director assured the Financial Times that it could be profitable simply by stopping investing too aggressively since it has a “very healthy” margin structure. The thing is, they can’t do it. OpenAI needs to achieve AGI, its great promise and the only thing that could justify this insane investment. If it fails, will cause a shock wave that can impact NVIDIA, AMD, Oracle… and end up dragging down the global economy. The competition tightens, Anthropic is eating the business market’s toast and Google is not only winning every time more users with Geminireached record revenue in the last quarterwhile OpenAI lost $11.5 billion in the same period. It doesn’t look good. Images | Wikipedia In Xataka | NVIDIA will invest 100 billion in OpenAI so that OpenAI buys chips from NVIDIA. And it’s a disturbing sign

Michael Burry just shorted NVIDIA. All good except because he was the one who predicted the 2008 real estate bubble

by

Michael Burry, the well-known investor and fund manager who predicted the 2008 financial crisis, has recently shown his bearish positions against NVIDIA and Palantir just after launching on social networks a warning about excess optimism in the market. Warning which the Bloomberg media has qualified ‘cryptic’, for several reasons. The movements, made known in regulatory documents filed on Mondayhave reopened the debate on whether artificial intelligence is generating a speculative bubble. What exactly has Burry done. His investment fund, Scion Asset Management, has bought put options (puts) worth $186.5 million against NVIDIA and $912.1 million against Palantir, according to mandatory filings with the SEC. These options benefit if the stock price falls. Burry also took bullish positions (calls) in Pfizer and Halliburton, two stocks that have underperformed the market this year. Why does it matter? Burry is not just any investor. Its history is marked by having bet short against the US real estate market two years before the 2008 crashenduring criticism from his investors until Lehman Brothers went bankrupt and his fund multiplied its profits. His story inspired the film ‘The Big Bet‘. Having gained that fame, when Burry bets against something, the markets pay attention, although his track record is not infallible, as he has been wrong in the past with other bubble predictions. Click on the image to go to the post The context of their movements. Days before these positions became known, Burry broke two years of silence on social networks with a disturbing message: “Sometimes we see bubbles. Sometimes you can do something about it. Sometimes the only winning move is not to play,” accompanied by an image of his character in the film. On Monday night he posted again, this time sharing a Bloomberg chart about concerns about circular financing between OpenAI, NVIDIA and other AI companies. Market reactions. Palantir shares fell more than 10% following the news, even though the company had just raised its annual revenue guidance. NVIDIA also fell by up to 2.9%. Palantir CEO Alex Karp responded in an interview with CNBC calling the idea of ​​shorting against companies like Palantir and NVIDIA, which he says are doing “noble tasks,” “crazy.” The bubble debate. For months, many investors have expressed concern about whether the AI ​​boom is being artificially sustained. Ray Dalio, founder of Bridgewater Associates, warned recently told CNBC that “there are many things that look like bubbles,” although he clarified that bubbles do not usually burst until the Federal Reserve tightens its monetary policy. According to its “bubble indicator”, approximately 80% of market gains are concentrated in large AI-related technology companies. An important nuance. It’s not entirely clear whether Burry is betting directly on the downside or whether these options are part of a more complex strategy to protect other investments. And just as share Bloomberg, regulatory filings only reflect long positions, so if you were using these puts as a hedge for other investments, we wouldn’t know. The curious thing is that its first quarter presentation did include a note explaining that puts “could be used to cover long positions”, but the third quarter presentation does not say anything about it. Scion’s recent history. This is not the first time Burry has bet against NVIDIA. During the first trimester He has already liquidated almost his entire portfolio of listed shares and bought put options against the chipmaker. However, it has also achieved success: in the third quarter it closed positions in Alibaba (with a 36.5% profit), Estée Lauder (27%), ASML Holdings (45.7%) and Regeneron Pharmaceuticals (10.8%). Canary in the mine or false alarm? The question on Wall Street is whether Burry is once again detecting a bubble before anyone else or if he is wrong this time. NVIDIA is up 54% this year until reaching a capitalization of 5 billion dollarswhile Palantir has soared 173% thanks to its expansion in AI-related businesses. Valuations are high, but both companies continue to grow and expand their business. Be that as it may, if there is a bubble, we will find out in the worst possible way: when it bursts. Cover image | Solen Feyissa and ‘The Big Short’ In Xataka | The geopolitical irony that we are experiencing in the chip war has an unexpected beneficiary: Russia

If you thought the AI ​​bubble was worrying, it’s because we hadn’t entered its next phase: debt

by

Big technology companies have issued $75 billion in bonds and loans between September and October 2025: Meta leads with 30,000 million. Followed by Oracle (18,000 million in bonds plus a loan of 38,000 million). And Broadcom (27 billion). The figure is equivalent to what these three companies used to borrow in an entire year. Why is it important. The shift from liquidity to debt marks a turning point in the AI ​​race. For years, these companies financed their infrastructure with cash flows, but now they are resorting to debt: Debt not linked to bonds has gone from 15% to 30% of its capital. The money trail. Oracle has closed the largest syndicated loan (a joint loan by several banks to a single client) in its history: 38 billion for data centers. Meta, for its part, is allocating its 30,000 million to campuses in Virginia and Oregon. And Broadcom uses them to strengthen its semiconductor division and its network equipment. The threat. Paying the interest on all this debt now consumes 15% of these companies’ operating profits, compared to 10% a year ago. And the cost of borrowing has risen: corporate bonds are near their most expensive levels since 2022. If the energy bill rises by 20% – a more than likely scenario given the stress on electrical networks – or if AI does not generate the expected revenue, these companies could see their credit rating reduced and trigger a chain crisis. Yes, but. Large investors continue to buy these bonds, attracted by returns of 6%. Money flows because official interest rates are at 3.75%so lending to these technology companies seems like a good deal. The problem is that any sudden change in rates can make these bonds lose value. And fast. At stake. Debt finances the AI ​​revolution, but also makes it more fragile and technology companies continue to increase their investment. If inflation returns or profits fail, the same debt that accelerates innovation could become a liability. Investors, meanwhile, continue to win; but they assume the risk of the storm. In Xataka | Apple is resisting the push for AI PCs because AI PCs have caused complete indifference Featured image | Towfiqu barbhuiya

When asked if AI is a bubble about to burst, big technology companies have just responded: hold my cap

by

The AI ​​race is about computing power and data centers the size of entire cities. And that doesn’t exactly come cheap. Big Tech is spending indecent amounts of money so as not to be left behind in AI and the fear that everything is a bubble flies over the environment. That doesn’t seem to stop them. Microsoft, Google and Meta have announced that they are increasing their planned spending on AI. what’s happening. Microsoft, Google and Meta have just presented their results for the last quarter and there are two pieces of news. The good thing is that all three have managed to increase their income. The not-so-good news is that they have sent a message to their worried investors: they are going to spend even more money than they planned on data centers and AI infrastructure. More wood. That AI is a bonfire of money we already knew it. Now we know it’s going to get even bigger. Meta had planned that Capex (capital expenditures) for 2025 would be $66 billion. Now they just said that The total will be between 70 and 72,000 million. And not only that, next year it will be even bigger. For its part, Alphabet (Google) had planned a Capex of 75,000 million, but they confirm that They will spend between 91 and 93 billion dollars. Finally, Microsoft has not given the annual data, but in this quarter They have spent 34.9 billion dollars5,000 million more than planned. In 2026 they expect spending to be even higher. Planned CAPEX REVISED CAPEX goal 66 billion 70-72 billion +24% GOOGLE 75 billion 91-93 billion +23% microsoft 30,000 million (quarterly) 34.9 billion (quarterly) +23% Also more income. Don’t panic, or at least not too much. All three have achieved record profits in this period. Meta earned 51.24 billion, Google 102.3 billion and Microsoft 70.1 billion, an increase of 26%, 16% and 13% more than the same period last year. All three assume that the numbers will continue to grow, and that is precisely what Those who warn of a bubble are not so clear. It’s not AI, it’s the cloud. In the case of Microsoft and Alphabet, the main vector of revenue growth is their cloud business, a trend that It started in the previous quarter and has continued to increase. Google Cloud generated 34% more revenue thanks to growth in “core products, AI infrastructure, and generative AI solutions.” In the case of Microsoft, its cloud services brought in 26.8 billion, 33% more than last year. And I published it. Meta is building data centers like there’s no tomorrow, but it doesn’t have a cloud business. Mete has something else: Facebook and Instagram. Its income comes largely from advertising and Zuckerberg assures that the good numbers come precisely because They are applying AI to improve their advertising systems. Not so fast, Zuck. Although Meta is the one that has increased its income the most compared to last year (26%), its shares have fallen 8% after announcing that it would continue to increase spending on AI. It seems that investors have quite a few doubts about their latest decisions, such as spend a million to create your superintelligence team or the plan to spend $600 billion in data centers. Image | Pixabay In Xataka | OpenAI is burning money like there is no tomorrow. The question is how long can he last like this?

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.