stock

The CNMV has tested AI to invest in the stock market for ten months. The conclusions are very revealing

by

In recent months there has been a recurring discourse that we see on social networks and that sell us again that “get rich quick” message. That message is “use AI to invest in the stock market.” The interesting thing comes when we see how the CNMV has published a study in which it has precisely attempted to analyze that premise. Although this organization warns of the risks of investing with AI, there is another important message in the conclusions: LLMs are not bad investors per se. They are bad at following vague instructions, which is just how most people use them. The CNMV study. Two researchers from the CNMV, Ricardo Crisóstomo and Diana Mykhalyuk, have published a study methodologically serious (but imperfect) and very interesting: they used four AI models for ten months live, from April 2025 to January 2026. They chose ChatGPT, Gemini, DeepSeek and Perplexity as models. The process was simple but demanding: each month they asked each model to identify the five stocks in the Ibex35 index with the best expected performance (to buy) and the five with the worst expected performance (to sell short). Then the real result was measured at the end of the month, and here there was no historical data selected just because: the real market was the only arbiter of all the functioning of the models. The models evolved. One of the most significant aspects of the study is that its creators recognized a methodological problem that was difficult to avoid: during those ten months, the versions of the four models were updated several times. The Gemini of April 2025 was not the same than that of January 2026for example, and that could influence the results. The researchers commented that it was impossible to know with certainty whether an improvement or deterioration in performance was due to the prompt strategy, market conditions in that period, or simply because the model changed. The prompt is everything. Three were also tested prompt types very different, and that gave rise to conclusions that were neither alarmist nor did they create false expectations: they were “it depends.” Thus, their results showed that everything depended on the type of supervision that these models had: If the LLMs were asked generic questions such as “What stocks should I buy?”, they failed repeatedly. There were computational errors, incorrect interpretations and also the famous hallucinations of chatbots. Curiously, the only one that made a profit was ChatGPT. The problem is that people who use AI to invest probably use this mode of action. But if prompts prepared with iterative reviews and human supervision at each step were used, Perplexity achieved a monthly return of 3.5% on the IBEX35. Gemini and ChatGPT also improved their behavior if given more precise instructions, and DeepSeek was the worst ranked overall. There is another finding: when models receive official regulatory documentation or business results reports, their predictive accuracy improves significantly. The LLMs they reason better on concrete and verified facts than generating analysis from scratch on information that they themselves search for on the web. financial hallucinations. The CNMV study points out that financial markets are especially demanding for AI models because they require complex processes. They have to retrieve and collect information dynamically, they have to reason in multiple steps, they have to be numerically precise, and they have to know this market, and all in real time. Chatbots are trained to generate “convincing” textsso the incentive here is that the investment recommendation “sounds good” even though it is completely wrong. The confidence with which AI models present incorrect financial analysis is proportional to the risk they pose to those who use them without checking whether what they say makes sense. In short: do not trust AI to invest right off the bat. The Reddit user’s experiment was equally striking, but hardly conclusive. Source: Reddit. The Reddit experiment. A Reddit user named Blotter-fyi rode in November from 2024 a platform called Rallies.ai which gave several AI agents access to real-time financial data and money to make stock market operations. Four months later, with the S&P index down 7% since the start, five of the models are outperforming that index, although only two have positive returns in absolute terms. The author himself was the first to warn that four months are insufficient to reach a conclusion: it could be luck, the market or simply the prompt. Nof1’s experiment was fascinating, but it made it clear that AI models don’t typically make money investing in crypto. Source: Nof1. Nof1 and crypto fascination. Another particularly striking experiment was the one that the company nof1.ai made with its Alpha Arena. He put six AI models to compete, gave them 10,000 real dollars each and gave them two weeks to trade cryptocurrency derivatives without human intervention. The most striking result was not who won, but who lost: GPT-5 ended with more than 25% losses and Gemini with close to a negative 40%. Meanwhile, the Chinese models Qwen and DeepSeek dominated in terms of good performance. They iterated with other models, 32 in total, and of all of them only six achieved a positive return: the rest lost money. Grok-4.20 was the big winner ahead of GPT-5.1 and DeepSeek v3.1. Maybe you shouldn’t just let AI invest for you. The conclusions after these experiments are clear. Four months of a model outperforming the S&P index in a bear market does not prove that AI is a good investor. Only in that specific period, with that specific marketthat model made decisions that turned out to be less bad than those in the index. To see if this makes sense takes years, multiple market conditions, and many instances of the same experiment running in parallel. The same happens with Nof1 – especially short – and with a more serious and methodical process like that of the CNMV, which was also surrounded by events whose impact on the final result was uncertain. Faced with so many unknowns, the conclusion seems clear: … Read more

The biggest oil crisis is not making them blink for a second in the stock market

by

We have been immersed in what can now be cataloged like the Third Gulf War. Since the United States and Israel offensive against Iran began at the end of February, the world has faced the greatest disruption of energy supply of its history. We are talking about a crisis that has paralyzed 20% of the world’s crude oil, sequestering about 20 million barrels a day They cannot cross the Strait of Hormuz. Missile falls, drones setting fire to infrastructure and thousands of deaths in the region. The impasse. Any basic economics textbook would dictate that financial markets should be in complete panic. However, the opposite occurs. It is enough for the White House to hint at a rapprochement or a vague ceasefire for the stock market to skyrocket, ignoring the physical fundamentals of a war in full swing. Wall Street lives in a parallel reality: the biggest oil crisis does not make them blink for a second. A virtual collapse in the face of a real war. This same week, the markets experienced 48 hours of unprecedented volatility. As detailed oil priceoil prices fell sharply in the Asian session on Wednesday, falling more than 5%. Brent crude oil, the reference in Europe, pierced downwards the psychological barrier of $100, while the US WTI fell to $87.51. The reason for this relief? According to the agency Reutersthe United States would have sent a 15-point peace proposal to Iran through intermediaries in Pakistan. US President Donald Trump boasted to the media that “productive” negotiations were moving toward a resolution. The screens of the traders were automatically dyed green: the European STOXX 600 index rose 1.2% and London’s FTSE 100 rose 1.1%. As Amelie Derambure explainedfrom the manager Amundi, the market simply launched itself to buy the idea of ​​a relief rally (a surge of relief) at the possibility of a temporary ceasefire. The bombs keep falling. However, there is no ceasefire; This should be clear. How to collect ReutersEbrahim Zolfaqari, spokesman for Iran’s joint military command, publicly addressed Trump on state television with these words: “Has the level of your internal struggle reached the stage of negotiating with yourself? We will never make a deal with you.” At the same time, military reality contradicts stock market optimism. The Pentagon prepares the deployment of elements of the 82nd Airborne Division to the region, a drone attack just hit a fuel tank at Kuwait International Airport, and Israel is deeply skeptical of any concessions Washington might make to Tehran in the shadows. Investors “bewitched” by the algorithm. To understand this disconnection you have to delve into the psychology of the market. An analysis published by FortunePaul Donovan, chief economist of UBSclaims that Wall Street is “spellbound” by the good news. “Markets do not react to information, they generally react to social media posts and headlines, even if they are fake news or contradictory,” says Donovan. Investors suffer from a cocktail of loss aversion and confirmation bias. They desperately want the war to end, so they embrace any story that confirms that desire and ignore negative news. Added to this, the “TACO” phenomenon (Trump Always Chickens Outor “Trump always cows”), a belief rooted in the New York trading floor that the tenant of the White House will end up backing down from the economic pain of a prolonged conflict to protect financial stability. Narrative as a weapon of war. Added to this is what energy expert Javier Blas defines in his column Bloomberg as jawboning (verbal intervention). The White House is winning the narrative battle in the markets without moving a single physical barrel. Trump’s constant messages in Social Truth promising a quick resolution—and even lifting sanctions on countries like Russia to flood the market—have managed to stop the panic. Blas sums it up perfectly: “Instead of being a sign of weakness, TACO is playing in Trump’s favor. No one knows for sure when or if he will try to end the war, which has been enough to prevent the traders skyrocket the price of oil.” The desperation to cling to any positive headline is such that it generates episodes of extreme volatility and information chaos. He Financial Times reported in his coverage how crude oil suffered wild fluctuations (Brent fell 11% to rebound shortly after) after a tweet by the US Secretary of Energy, Chris Wright, stating that the Navy was already escorting oil tankers through Hormuz. The message was deleted minutes later and denied by the White House itself, but the effect on the algorithms had already occurred. The bath of physical reality. While Wall Street plays a game of guessing the next tweets from the Oval Office, the physical reality of oil is stubborn. A report from Bloomberg puts his finger on the sore: The physical market continues to deal with shortages, and the war has demonstrated the absolute control that Iran exercises over the Strait of Hormuz. Although Tehran informed the International Maritime Organization that “non-hostile” ships can transit, the route remains effectively closed and reports circulate about the presence of dozens of naval mines Iranians in the area. The mathematics of disaster, detailed by Reutersthey are chilling. After 25 days of conflict, the world has stopped receiving 500 million barrels (the equivalent of five full days of global supply). The logistical desperation is such that Saudi Arabia has boosted its exports from the port of Yanbu, on the Red Sea, to avoid Hormuz. To compound the crisis, Russia has suspended cargoes at its Baltic ports following a vicious Ukrainian drone attack, adding more uncertainty to the global market. Larry Fink, CEO of the management company BlackRocksummed it up bluntly in statements to the BBC: “If Iran continues to be a threat to Hormuz and oil settles between $100 and $150 per barrel, we will have a global recession.” Collateral damage. The narrative chaos has even reached gold, which has lost their protection status. According to Financial Timesthe price of the precious metal has plummeted 16% since the start of … Read more

An economic science fiction text has sunk Visa and Mastercard in the stock market. The reason is more disturbing than the story itself

by

Citrini Research, a hedge fund American published this week a text written as if it were a macroeconomic memorandum from June 2028. It is not a prediction, its authors warn. It is a speculative exercise. A feasible scenario. It has achieved 24 million impressions, and counting. It is not an anecdotal tweet. The markets they have responded by sinking. Visa has fallen 4.4%. Mastercard, 6.3%. American Express, almost 8%. And Capital One, 8%. This deserves an explanation. And it’s not what it seems. Between the lines. The market reaction is not explained by the specific content of the Citrini Research report, which includes arguments as debatable as that AI agents will abandon cards to pay with stablecoins in Solana. Antonio Ortiz, technology analysts, has pointed it out precisely: part of the argument “it is from the first of Twitter AI-hype“. The idea that an agent will compare twenty food delivery apps vibecodeadas to find the cheapest one smells like a caricature of the future. But the panic is not irrational. It is precisely the panic of not knowing where the limit is. Why is it importantand. What has moved the market has not been so much the thesis about payments but the thesis about the destruction of value. And that is solid: many billions of dollars of market capitalization have been built on a single foundation: that humans are slow, impatient, forgetful and loyal out of inertia. That we do not compare prices. That we renew subscriptions that we do not use. And that we pay commissions that we do not negotiate. An AI agent has none of those weaknesses. And that changes everything. The backdrop. Citrini’s report comes at a time when the so-called “saaspocalypse“is no longer a metaphor. WSJ states that investors are terrified by the possibility that AI ends up doing the work that large software companies bill for today. ServiceNow, Salesforce, business management platforms… all built on the premise that companies need software for their employees to do their jobs. But… what happens when employees disappear? What if the software itself can be replicated in weeks with agentic coding tools? Citrini’s fiction begins exactly there, in early 2026, when a competent developer can reproduce the core functionality of a mid-market SaaS in a few weeks, and constructs a scenario of systemic collapse. The big question. The report’s most disturbing argument is that in every previous technological cycle, job destruction created new jobs that only humans could do. This time, AI is already occupying those new positions as well. If that’s true—if AI improves faster than workers can reorient themselves—the self-correcting mechanism that has always kept creative destruction from turning into outright destruction wouldn’t work. That is the scenario that the markets have discounted this week, even if only partially and speculatively thanks to a creepypasta financial. Yes, but. The scenario requires assuming a speed of adoption that is not guaranteed, a completely absent political response and a total absence of new economic sectors. None of the three conditions are set in stone. Furthermore, as Antonio points out, there is some collective hysteria in the reaction: each announcement or “scary story catches attention and moves investors.” Markets are trading in panic over the unknown. But there’s an important difference between saying “this scenario won’t happen” and saying “this scenario is impossible.” And that difference is exactly what has the market nervous. The alarm signal. The most striking thing this week is that a speculative text, written in economic science fiction format, has been enough to move billions in market capitalization. That says a lot about the state of certainty in the markets regarding AI: it is practically non-existent. Nobody really knows how much a company whose moat It is human friction in a world where that friction is disappearing. The canary is still alive. But investors have stopped trusting the canary. In Xataka | AI promised to revolutionize all sectors. It has only revolutionized programming while the rest is still waiting Featured image | Avery Evans

Steam Deck is out of stock, and PS6 and Switch 2 are already shaking

by

Considerable tension in the video game industry at a point in which it is perhaps not going through its healthiest moment. The last slap that shakes a sector that already balances in a moment of perfect storm is the shortage of DRAM chips, driven primarily by the demands of the artificial intelligence industry and its demands regarding data centers. Valve has already announced that hard times are coming for its warehouses and stocks, and rumors about changes in stock, prices and launches of Switch 2 and Playstation 6 are constant. The first, the Steam Deck. If you tried to buy a Steam Deck from the United States last week, you would see the ‘Out of stock’ message. Was a disappearance without warning and that affected the 512 GB and 1 TB OLED models (the 256 LCD has not been available for several months). It is something that, at the moment, only affects to Valve’s home country and Canada: In most European countries and some Asian markets it is still available. With a brief note, Valve has confirmed it from the website: The component crisis due to the demands of the AI ​​market has left the console models that were still available out of stock. And remember that when the LCD 256GB stock runs out, that will be all for that model. At the moment there are no clues about when the missing consoles will be back on sale or if prices will increase. Delays at Valve. There is another consequence of this component crisis: the delay of the company’s next releases, the Steam Machine, the Steam Frame and the new Steam Controller. In it steam hardware blog The company acknowledged that it was not able to offer prices or launch dates, despite the initial intention to launch them in the first quarter of the year. The launch window was moved from “beginning of 2026″ to “first half of 2026”. As for the price, which remains unknown, the analysts’ first intuitions set it for the Steam Machine between 700 and 800 dollars. Since then, subsequent calculations have skyrocketed beyond a thousand dollars. Could be, As some analysts commentthat Valve is prioritizing the available memory stock towards the Steam Machine, leaving its laptop on a secondary level, and hence the lack of stock? In any case, the second-hand market rubs his ditto. The devastating DRAM crisis. The problem behind this shortage is the supply crisis of DRAM, a type of memory present in computers, consoles or smartphones. The three manufacturers that control 90% of production (Samsung, SK Hynix and Micron) have been overwhelmed by demand for artificial intelligence data centers. Some reports speak of something that goes beyond a conventional cyclical shortagewhich points to a necessary reformulation of needs and prices. The problem is that manufacturing HBM (High Bandwidth Memory), the high-density and extremely high-performance variant that powers the AI ​​centers, makes Samsung and company have to limit DRAM production, and LPDDR5 production, the standard used by devices like the Steam Deck, plummets. Consequence: DRAM prices increased by around 50% throughout 2025, with another 30% increase in the fourth quarter of the year, followed by a further increase of around 20% so far this year. The situation does not seem to be going well finish in the next few months. Collateral damage 1: PS6. Bloomberg was the one who gave the alarm voice: Citing sources close to Sony, he revealed that the company was thinking of delaying the launch of PS6 until 2028 or 2029. If this last possibility happens, we would be in the longest interval between Playstation generations in history: nine years. Beyond the wait it entails for the players, it also disrupts developers’ plans and projects which, according to most sources, would have planned their launches around 2027. Technically, and although with PS6 we are in rumor territorythe console would arrive equipped with 30 GB of GDDR7 RAM, fourteen gigabytes more than PS5. At a time when Samsung has sold all its production capacity by 2026 and Micron has confirmed which does not have the capacity for new contracts until 2027, mass manufacturing PS6 is unfeasible. At least with PS5 there is no need to worry: Sony has confirmed which has enough memory to cover its needs for 2026 and 2027, which also guarantees that there will be no price increases. Collateral Damage 2: Switch 2. The ballot that Nintendo has in its hands is even more problematic, because Switch 2 It is already on the market, but it may have no choice but to raise the price in 2026. Bloomberg reported it in the same PS6 article, also citing people familiar with the crisis. Significantly, Nintendo was one of those responsible for the memory supply during 2025 being strained: its splendid launch and sales are, in part, responsible for the current situation. According to different sourcesNintendo would currently be paying 41% more for the 12 GB of RAM that is in each Switch 2 unit, compared to the cost anticipated in its original projections. Maintain current price would suppose a loss of $50 per unit sold during 2026: a drain (the console has already sold almost 18 million units) that not even the Japanese giant can afford. At the moment there are no official decisions, but the president of Nintendo recognized that “the current memory market is very volatile”: in the presentation of results admitted that the increase in component prices “exceeds our expectations”, although also that the company has accumulated inventory and closed long-term supply agreements. Nobody escapes. We have before us the three great sectors of the world gaming: PC, home console and hybrid consoles. All three have the same problem, and it cuts across the most powerful companies, which tells us not about a specific crisis, but about one that has structural overtones, and that goes beyond the video game sector. For example, PC users are already facing price increases or manufacturers like Dell and Lenovo considering reducing the RAM of your laptops mid-range. … Read more

and the stock market gets restless

by

SpaceX is exploring two merger scenarios: joining with Tesla or xAI, as reported Bloomberg and Reuters citing sources familiar with the matter. This above all is a warning for those who currently have investments in one of Elon Musk’s companies, although it is also a situation that can generate a business earthquake. Perhaps the biggest stock market movement in history. We are going in parts to explain everything we know so far. Two paths on the table. According to Bloomberg, SpaceX has internally discussed the viability of a merger with Tesla, pushed by some investors in the aerospace company. For its part, Reuters point that a combination with xAI is also being evaluated, which could materialize before the IPO scheduled for this year. Of course, it should be noted that none of the companies involved have publicly confirmed these conversations, and sources warn that no final decisions have been made. Movements. Last year, SpaceX agreed invest $2 billion in xAI, according to The Wall Street Journal, and this week Tesla also revealed that it had allocated the same amount to the AI ​​startup, according to account the middle. On the other hand, it is worth mentioning that xAI acquired X (formerly Twitter) in 2024, in an operation that valued xAI at $80 billion and the social network at $33 billion, as Musk himself mentioned. Here we once again witness a circular economy movement among its companies, with consequences that remain to be seen. The logic. There are several legs here. On the one hand, it is worth mentioning that Musk has a thorn in his side build data centers in space to feed AI, something he has expressed on several occasions. In this way, a merger with xAI would allow SpaceX to accelerate this project, taking advantage of xAI’s computing capacity for its in-orbit facilities. In the case of a merger with Tesla, the move could serve well to leverage its ability to manufacture energy storage systems, something that could help SpaceX expedite the development of systems that use solar energy in space to power these data centers. Musk has also talked about using Starship rockets to transport optimus robots from Tesla to the Moon and Mars. The interest of investors. Regardless of the path chosen, any deal could attract considerable interest from many investors, especially infrastructure funds or Middle Eastern sovereign investors, according to they count from Bloomberg. The outlet also relies on public documents in Nevada that show that on January 21, two legal entities were formed with the expression “merger sub” in their names. The name of Bret Johnsen, SpaceX’s chief financial officer, appears in these documents. Between the lines. The operation raises some questions. If SpaceX acquires xAI, SpaceX shareholders could see their stake diluted. If Tesla buys xAI, its shareholders would effectively be financing a bailout of Musk’s other businesses. In any case, it is clear that the possible merger would raise several blisters. The numbers at stake. SpaceX reached a valuation of approximately $800 billion by the end of 2025, according to counted TechCrunch, becoming the most valuable private company in the United States. Tesla has a market capitalization of around $1.56 trillion. According to BloombergSpaceX is contemplating an IPO that could value the company at around $1.5 trillion, possibly in June, close to Musk’s birthday and with a fundraising of up to $50 billion, which would make it the largest IPO in history. Initial reaction. Tesla shares rose as much as 4.5% in after-hours trading after news the news from Bloomberg. Johnsen told employees in December that a possible IPO would help drive “a huge rate of flight” for the Starship rocket and a possible base on the Moon. Meanwhile, the media also says that Bank of America, Goldman Sachs, JPMorgan Chase, Morgan Stanley and Robinhood Markets are already positioning themselves to host the operation. Cover image | Flickr (Ministry of Communications) and SpaceX In Xataka | There is a perfect time of year to ask for a raise: January to March

Sandisk has risen 1,000% in the stock market since the summer. Its advantage is called Kioxia

by

In just five months, Sandisk shares have soared 1,000% in one of the most astonishing recoveries in Wall Street history. The company has been the latest big beneficiary of the AI ​​boom and the rush to build data centers full of advanced AI chips… and also the memories that accompany those chips. That’s where Sandisk’s great asset comes in, called Kioxia. Value of Sandisk shares in the last six months. Source: Google Finance. Without knowing it, SanDisk was ready for the revolution. HBM memories were traditionally the favorites to accompany GPUs that were the great “brain” of AI, but the scarcity of these components with high bandwidth has meant that the spotlight has been focused for a few months on DRAM and NAND memories, two types of storage in which sanDisk is a dominant player. Like other manufacturers in its segment —Micron is one of the outstanding—, SanDisk has suddenly found itself in a situation that benefited it enormously. free money. The memory chip market works like a commodity market in which leverage can be significant. That means that when prices rise, companies like SanDisk don’t need to invest in new factories or employees to earn more — although they can build them if they deem necessary. It is as if for Micron or SanDisk this phenomenon is equivalent to “free money” because they are receiving much more income for the same products they sold a year or two ago. Not even they themselves expected it: SanDisk CEO David Goeckeler talked about the rise of AI in June, and commented “We try to estimate demand. We think demand is good. What we need is to get supply to match that.” He couldn’t anticipate what would happen with memories starting in September. DRAM and NAND memory prices are skyrocketing from the end of 2025. Source: Sherwood. The key alliance: Kioxia. In recent times SanDisk has grown significantly in your solid state drive business (SSD) for enterprise data centers. But it also maintains a historical strategic alliance with the Japanese company Kioxia, which allows it to obtain NAND chips at a much lower cost than its rivals. The profit margin skyrockets, and so do the shares on the stock market. A relationship with ups and downs. The relationship between Sandisk and Kioxia (formerly Toshiba Memory) is based on a Joint Venture from more than 20 years ago focused on the development of NAND memories. This alliance has achieved advances such as the memories BICS Flash (with 3D storage technology), the wafers that leave their factories are shared between both companies. Kioxia went through a difficult time after Toshiba’s financial crisis and failed merger attempts with Western Digital. They survived all this, and together with Sandisk now the Japanese company controls 30% of the global NAND market. Some win, others lose. The investment fund Elliot Management pushed in early 2025 for SanDisk will separate from Western Digital. They believed that at that time it was worth about $20 billion—as when he bought it a decade ago—, and that fund sold its stake just before the total market explosion. Today that stake would be worth more than $340 million. Bad business for users. But in addition to that background, the ones who have it most complicated are the users, who will continue to suffer the consequences of this phenomenon for months, and perhaps years. Neither Micron nor Sandisk/Kioxia appear to have any intention of significantly expanding production capacity. They already did this during the pandemic and that caused excess inventory when demand fell after confinement. Now they do not want to expose themselves to the same thing, and there is talk that the price increase will continue throughout 2026 and may let’s take a long time in seeing memories at prices “like those before”… if we end up seeing them. Image | Igor Shalyminov In Xataka | Japan has taken out the checkbook to once again dominate the chip industry. Prepare a plan of 325,000 million dollars

a stock of billions of dollars

by

These are not easy times for the alcohol industry. And not only for the crossfire of the trade war, ups and downs of prices or the apparent loss of interest of Generation Z for drinking. As customers demand less whiskey, cognac and tequila, the giants of the sector have found themselves with a growing stock that some estimates already put at 22 billion dollars. Thousands and thousands of bottles that threaten to strain the finances of large manufacturers and (even worse) unleash a price war that clouds their future. “The accumulation of inventories is unprecedented,” warn. A huge “lake of liquor”. So recently described Financial Times the panorama that the large distillate manufacturers have encountered, giants of the sector that have their warehouses full of bottles that cannot be disposed of. To be exact, the newspaper claims that five of the titans of the industry that are listed (Diageo, Pernod Ricard, Campari, Brown Forman and Rémy Cointreau) have a stock of aged spirits valued at 22 billion. Click on the image to go to the tweet. Breaking schemes. If the figure seems high, it is because it is. Those $22 billion mark the highest level of stock in more than a decade and there are already those who warn that they paint a delicate picture. “The accumulation is unprecedented,” recognize FT Trevor Stirling, analyst at Bernstein. The most extreme case would be that of the French cognac manufacturer Rémy Cointreau, which according to the newspaper accumulates aged stocks worth 1.8 billion euros. Almost double its annual income and close to its global market capitalization. Is there more data? Yes. In recent years the percentage of mature stock over total net sales has increased clearly at Rémy, but also at Brown-Forman, Campari, Diageo and Pernord Ricard. For example, in the case of the British multinational Diageo, this ratio has clearly increased in just a few years: if in the fiscal year of 2022 it represented 34%, it is now 34%. 43%. The problem is not just the big manufacturers of Scotch whiskey, cognac and tequila. Data from the Tequila Regulatory Council show that at the end of 2023 the Mexican industry had a stock of 525 million liters of that popular distilled beverage. The figure (sum of the product in barrels or pending bottling) is almost equivalent to the country’s annual production. “Much more new liquor is distilled than is sold, and the stock begins to accumulate,” duck Bernstein. A ticking bomb. The accumulation of stock is not worrying only because of what it suggests to us about the past and current sales pace. It is especially so because of its implications for the future. With more barrels and bottles gathering dust in the warehouses of the big manufacturers there are those who already fear that a price war would break out, a pulse in the market that would aggravate the situation. For now, FT recalls that there are manufacturers who have chosen to hit the brakes in their factories. This is the case of Diageo, which has production suspended of whiskey in various factories, or from the bourbon producer Jim Bean (Suntory), which has done something similar with its main Kentucky distillery. The problem: often the aged drinks sector works for several years, so pausing its production today can compromise the supply in five years or a decade. What is the reason? To understand the current stock of the industry, it is necessary to understand several keys. For example, the fluctuations in demand in recent years and the forecasts with which manufacturers have worked. At the beginning of the pandemic, the sector registered an increase of distillate consumption in the US, which led to an increase in production. After the health crisis and with inflation as a backdrop, however, the market returned to normal. What’s more, the industry had to deal with new challenges that a priori have little to do with its business. The first is the trade war unleashed last year, a scenario in which the alcohol industry was not foreign. In fact, if Jim Bean considered suspending production at his main distillery it was precisely due to the increase in stock and the uncertainty generated by tariffs. Another key factor is that alcohol consumption (at least of certain types of alcohol) appears to be moderating as more people focus on their physical health and take weight loss medications, such as Wegovy or Ozempic. It is nothing exceptional if you take into account that in 2023 a study Walmart already warned that consumption of Ozempic was reducing food sales. The big question. Beyond these current factors, a key question for the industry hovers: Are we consuming less alcohol in general? Do we drink less than our parents and grandparents? And will the new generations on whom the sector will have to rely on in a couple of decades drink less? There is data that suggests this. The Our World in Data platform has developed a graph on per capita consumption that reflects that almost all the countries analyzed consume less alcohol than a few decades ago. It is not the only study that points in that direction. Another recent one from Gallup confirms that in the US consumption has fallen to its lowest level since at least 1930 and OECD tables They also show that many of their countries (not all) have seen how the intake of liters per person per year decreased between 2013 and 2023. There are those who already warns that the trend does not look like it will stop, fully affecting to the accounts of the distillate industry. Images | Paolo Bendandi (Unsplash), OECD and Our World in Data In Xataka | There is an age at which we should stop drinking alcohol forever. Neuroscience is clear why

There is still stock of all these bargains

by

November is coming to an end and, although we already have Christmas on the horizon, we still have a few more days to take advantage of the offers of Black Friday. It is true that many bargains have disappeared from the main stores, but that does not mean that we do not still have some powerful discounts on technology. In fact, a store that still has several very attractive ones is, without a doubt, PcComponents. This store, immersed right now in its final Black Friday fireworkshas several deals that can suit us very well if we are thinking of renewing our old computer for a new one or if we want to change some of its parts in case we already have a PC assembled. There is a lot to choose from, but below we leave you some very powerful offers that are still available. MacBook Air M4 by 849 eurosa very interesting laptop for those looking for outstanding performance and Apple ecosystem. AOC Q27G4ZR Monitor by 159 eurosgaming monitor with Fast IPS panel and a refresh rate that can reach 260 Hz. AMD Ryzen 5 5500 processor by 70.90 eurosan economical CPU option with good performance. AOC CU34G4 Monitor by 229 eurosideal if you are looking for an ultrawide monitor to play. Forgeon Solum Liquid Cooling by 54.99 euroswith two powerful fans and RGB lighting. Radeon RX 9070 XT graphics card by 649.90 eurosa minimalist assembly ideal for playing at 1440p or even 4K. MacBook Air M4 If we want an Apple laptop and we are looking for something current, we will find few better options right now than this one. MacBook Air M4. This is the 13.6-inch version with 256 GB of storage, available right now for 849 euros when placing the equipment in the cart. A compact and very manageable device, but that does not give up offering a large dose of power and autonomy. Ideal for working, studying or watching multimedia content. Apple Macbook Air Apple M4 Laptop 10 Cores/16 GB/256GB SSD/GPU 8 Cores/13.6″ Silver The price could vary. We earn commission from these links AOC Q27G4ZR Monitor If we are looking for a new gaming monitor, this AOC Q27G4ZE could be a great fit for us. It is a model with 27 inches, QHD resolution and 240 Hz refresh rate, although it supports overclock to raise the rate to 260 Hz. It uses a Fast IPS panel and barely has 1 ms GtG response. It has support for HDR 400 and has speakers, which is a plus for single player games where we don’t want to use headphones. It is available for 159 euros. AOC Q27G4ZR Gaming Monitor 27″ QHD Fast IPS 240Hz Overclockable 260Hz 1ms HDR400 Speakers The price could vary. We earn commission from these links AMD Ryzen 5 5500 processor Not all users want to have the most powerful or the latest of the latest on their PC. If you are just looking for an affordable gaming device to play at 1080p without too many complications and without spending too much, this Ryzen 5 5500 may be ideal for you. It has 6 cores and 12 threads, more than enough also for simple and undemanding tasks. By 70.90 eurosa CPU with very good quality-price ratio. AMD Ryzen 5 5500 3.6GHz Box Processor The price could vary. We earn commission from these links AOC CU34G4 Monitor Another monitor option, also from AOC, although in this case significantly different from the previous one. In this case, we have a 34-inch ultrawide screen with 1500R curvature, which is ideal for gaining immersion while we play. Its resolution is WQHD (3,440 x 1,440 pixels) and it has a refresh rate of 180 Hz. It is very adjustable, so we can place it as it best fits our setup. It is at a minimum price: 229 euros. AOC Monitor CU34G4 34″ Curved WQHD 3440×1440 180Hz HDR10 Fast VA 1ms Black The price could vary. We earn commission from these links Forgeon Solum Liquid Cooling Accompanying our CPU with good liquid cooling will allow us to have sustained performance without temperatures rising excessively. This one from the Forgeon brand is ideal if we take into account that it also barely costs anything right now. 54.99 euros. It has two 120 millimeter fans and a square pump. In addition, the system has RGB lighting that will allow us to give a different touch to our tower. Forgeon Solum Liquid Cooling 240 ARGB Liquid Cooler Kit Black The price could vary. We earn commission from these links Radeon RX 9070 XT graphics card We close this selection of offers with a graphics card, this time an RX 9070 XT. This assembly, from XFX, has an arrangement of three fans and a fairly sober and minimalist appearance, ideal if we are not concerned about having lighting or flashy details on the PC. At the performance level, it is a graphics card capable of offering outstanding performance at 1440p, as well as in 4K, relying on technologies such as AMD’s FSR. comes out for 649.90 euros. XFX SWIFT AMD Radeon RX 9070 XT Triple Fan 16GB GDDR6 FSR 4 Graphics Card The price could vary. We earn commission from these links Some of the links in this article are affiliated and may provide a benefit to Xataka. In case of non-availability, offers may vary. Images | PcComponentes, Apple, AOC, AMD, XFX In Xataka | Best wireless headphones. Which one to buy and 21 models from 15 euros to 470 euros In Xataka | Best gaming keyboards. Which one to buy and 11 recommended gaming keyboards for different users and budgets

Oracle signed a 300 billion agreement with OpenAI. Two months later it has lost 315,000 million in the stock market

by

Since Oracle announced its $300 billion deal with OpenAI On September 10, its shares have lost $315 billion in market capitalization, as they have stated since Financial Times. The technology company He has bet everything on a single card: Become the premier infrastructure provider for the world’s most valuable AI lab. Investors are not convinced. The most expensive bet in its history. Oracle has tied its future to OpenAI in an unprecedented way in the technology industry. According to estimates At Jefferies, 58% of its future order book comes from a single customer: OpenAI. To put it in perspective, Microsoft has just 39% concentration with its largest customer, and Amazon 16%. Oracle has gotten into a mess and its business diversification has become a critical dependency on OpenAI. The plan is ambitious but risky. Oracle’s strategy is to reach $166 billion in cloud computing revenue by 2030, according to counted the company last month. To achieve this, its investment budget in the current fiscal year ending in May amounts to $35 billion. The analysts wait that this annual expenditure will stabilize around 80,000 million in 2029. But here’s the problem: Starting in 2027, most of that revenue would come from OpenAI, according to the calculations from RBC Capital Markets. That is, Oracle is not just building massive infrastructure, it is building massive infrastructure for a single tenant that has yet to prove its long-term commercial viability. The numbers don’t add up yet. Oracle’s net debt already stands at 2.5 times its ebitda (earnings before interest, taxes, depreciation and amortization), more than double what it was in 2021, and is expected to almost double again by 2030. Its free cash flow is also expected to remain negative for five consecutive years, according to the forecasts collected by Bloomberg. The company is financing with debt a gigantic server farm with the hope that OpenAI will generate enough revenue to justify the investment. Meanwhile, as has shared Financial Times, investors are so restless that the cost of insuring against a potential Oracle default is at a three-year high. The contagion effect of OpenAI. Oracle is not the only company that has suffered after announcing agreements with OpenAI. Broadcom and Amazon too have seen their shares fallwhile NVIDIA has barely moved since its investment agreement in September. A few months ago, any type of association with OpenAI caused prices to rise, considering himself the King Midas of AI. The most notable case was AMD’s in Octoberwhen its shares rose 24% after announcing a chip deal that included company warrants. That halo effect seems to have completely faded. Between the lines. The initial theory was that OpenAI was in a frantic race to catch up. general artificial intelligence (AGI) and that Oracle was the only company capable of scaling the necessary computing capacity at the required speed. Oracle promised the lowest upfront costs and the fastest path to revenue generation because it acted as a data center tenant, not an owner. Now investors are sending the signal that partnering with OpenAI is no longer a guarantee of success. The alternative reality is less rosy: Oracle doesn’t have as much operating profit as its competitors to burn on R&D, so it’s betting everything to keep its only big customer in exchange for a promissory note. Amazon, Microsoft and Meta can afford to spend between 70,000 and 130,000 million a year in infrastructure. Oracle is juggling financials to keep pace. And now what. Oracle has until mid-2026 to prove that your Abilene data center in Texas, with capacity for more than 400,000 GPUs and 1.4 gigawatts of power, can generate the promised returns. Meanwhile, the market has spoken and is awaiting evidence that this partnership will bear the promised fruits. Cover image | Oracle and OpenAI In Xataka | As if there weren’t enough AI companies, Jeff Bezos has just returned from the shadows to build another one, according to the NYT

Duolingo was the fun, brave company we loved that taught us languages. Today it is sinking in the stock market

by

Most people never manage to turn their ideas into business successes. Luis von Ahn (Guatemala City, 1978) has achieved it twice. The first, when he created reCAPTCHA and sold it to Google in 2009 for a small fortune. The second, years later, started from a much simpler concept. Learning languages ​​was a painso von Ahn wanted to turn that into just the opposite: something fun. This is how it was born Duolingoa company that taught how to speak languages ​​with a strong component of gamification. You already had to go to an academy or spend long periods of time in online courses: you could learn words, phrases and pronunciation through small tests when you were on the bus or waiting in a queue. Duolingo achieved the most difficult thing: making us like each other (and fall in love) Learning with Duolingo was fun and comforting. The small rewards worked and turned it almost into a video game that little by little more people became fond of. The snowball got bigger and bigger and Duolingo became one of those companies that already seemed likeable at first. It seemed that everything it did was done well, and little by little the company took important steps to become the giant it is today. The certifications arrived who wanted to rival the famous TOEFL exams, their platform for schools, and more and more languages. Some, like japanesewere a challenge. Others, like the Klingon or the high valyriumwere above all a diversion that consolidated the fun and cool image of the company. Then things started to get interesting because Duolingo wanted to not only teach us languages ​​to speak, but also programming languages. He was encouraged to want to serve as a tool so that the little ones They learned to read and write. And for the young and not so young, Duolingo wanted to become private mathematics teacherof music or even chess. All of this ensured that over the years Duolingo managed to solidify that company image that Not only did he solve real problems, but he did it in a friendly, friendly and fun way.. In 2021 the company decided go public and after a couple of relatively calm years, the shares began to rise in value significantly. Everything seemed to be going great for the company. And then everything went wrong. AI has mortally wounded Duolingo, but not because of what we think When OpenAI presented GPT-4o in June 2024, many of us saw the future. One in which you no longer typed on your computer or on your mobile screen: it was enough to talk to him. That promised to transform many segments and kill some others, and among those threatened were companies like Duolingo. At the time it wasn’t so obvious, but when we saw that kid solving a math problem With the help of AI, it was not difficult to imagine that education, as we had known it, could have an expiration date. Curiously, that didn’t seem to affect Duolingo too much. The company continued to grow, but then two things happened. First and foremost, a major blunder. Luis von Ahn advertisement in April an “AI First” vision in which I would bet on artificial intelligence as a new great tool for your growth. The message sounded like “let’s do without the human being,” and although von Ahn tried to clarify things, the damage was done. After that, the debacle. Duolingo shares began to plummet. But the thing didn’t end there. The second of those turning point events occurred in August, when GPT-5 demonstrated that one could build a custom Duolingo for, for example, learn french in a fun way. People stopped being in love with Duolingo and they began to criticize her precisely because of what had made her succeed. There was too much gamification and, as i said a user on Reddit, “for me the reward for learning a language is learning the language.” Source: Cinco Días. Stocks continued to fall almost steadily. These days Duolingo presented financial results, and the curious thing is that although they were good, they were not good enough for Wall Street. The firm reached 135 million active monthly users (50 million use it daily), 20% more than in the same period of the previous year. It also rose 34% in paying users. Although one would think those numbers were fantastic, they also warned that the forecasts for the fourth quarter were not so optimistic. Result: new stock market debacle. So much so that the shares have plummeted 64% since reaching their highs on May 1, just after the “AI First” announcement. Since then, Duolingo’s drift has been worrying, and the coming months will undoubtedly mark its future even more. The company is in a difficult moment, and the rise of AI may end up causing those experimenting with their chatbot to realize that starting to learn languages ​​​​is as easy as telling ChatGPT “I want to practice my English with you a little. Correct me when I say something else and suggest small exercises” out loud. That is the great challenge for Duolingo going forward. In Xataka | How to practice languages ​​using artificial intelligence

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.