DeepSeek has just released a model that competes with Opus 4.6. It costs seven times less and runs on Chinese chips

by

They have passed 484 days since that “DeepSeek moment“, but the wait It seems to have been worth it, because we have the new DeepSeek V4 with us. We are facing an absolutely gigantic open weights model that once again promises to crack the foundations of the proprietary foundational models of Anthropic, OpenAI or Google. This is moving, gentlemen. Gigantic and open. DeepSeek v4 is an Open Source model and comes in two versions. The first is the Pro, with 1.6 trillion parameters (1.6T), of which it has 49,000 million active. The second is Flash, with 248,000 million parameters (248B, huge for a “Flash” model) of which 13,000 are active. More efficient than ever. Both versions they make use of a Mixture-of-Experts (MoE) architecture, which means that only a fraction of the parameters are activated in each inference. This allows the computational cost to be reduced significantly. Both versions support a context window of one million tokens—to include novels and novels at once as input—when in v3 it was 128,000 tokens. Furthermore, this model is much more efficient than its predecessor in computing per token: it requires only 27% of the operations per token and 10% of the KV cache compared to DeepSeek v3.2. Benchmarks promise. DeepSeek’s internal testing reveals that v4 Pro-Max (the best model with the highest reasoning ability) outperforms or is on par with Claude Opus 4.6 Max, GPT-5.4 xHigh, Gemini 3.1 Pro High, Kimi K2.6 and GLM 5.1. The results, however, are not independently verified, which means we should take them with caution. The numbers are still striking: in LiveCodeBench, a programming test, DeepSeek v4-Pro-Max achieves a 93.5% score compared to 88.8 for Opus 4.6 and 91.7% for Gemini 3.1 Pro. In other tests there is more variability, but at least on paper DeepSeek v4 Pro seems as good as Opus 4.7, which until now was the absolute benchmark. Much cheaper. But as happened with its previous version, the difference in price with those models from US companies is astonishing. As point the analyst Simon Willinson, the official prices of DeepSeek v4 Pro are 1.74 dollars per million input tokens and 3.48 dollars per million output tokens, up to almost seven times less than those of Opus 4.7 and up to almost 9 times less than those of the new GPT-5.5. With DeepSeek v4 Flash the cost is 0.14/0.28 dollars per million input/output tokens, when GPT-5.4 Mini costs up to 16 times more. The conclusion is obvious: if it really does what it says it does, the price is an absolute bargain. That is precisely the challenge: that real experience confirms what the benchmarks say. The hardware mystery. DeepSeek has not revealed what hardware has been used to train this version of its founding model. In the past they did admit that they had used NVIDIA’s H800s. Which yes it is known The thing is that the model has been developed to run on both NVIDIA and Huawei Ascend chips. This last has confirmed Baidu that its Ascend Supernode clusters based on the Ascend 950 will fully support DeepSeek v4 versions. Huawei support is “horrible” news for the US. In The Information they already commented that one of the reasons for the “delay” in the appearance of this model was to adapt it so that it worked without problems with Huawei chips. That support is according to Jensen Huang “horrible” news for the US, because it means that dependence on NVIDIA chips no longer exists or at least is reduced to a minimum. But. The launch comes at a difficult time for the company. Guo Daya, one of the people responsible for the v1 and v3 models, has signed for ByteDance to work on AI agents. Luo Fuli, who led the development of v2, joined Xiaomi last year. This launch also coincides with DeepSeek seeking external funding for the first time. They are expected to raise about $300 million and obtain a valuation of about $20 billion. according to The Wall Street Journal. From the surprise effect to the continuity effect. The launch of DeepSeek R1 in January 2025 was surprising because it demonstrated that China could train competitive models at a fraction of the cost of Western models. With DeepSeek v4 that surprise effect disappears to give way to the continuity effect. This model seems to maintain precisely what made the previous model famous: extraordinary power at a very low cost. Bad news for Anthropic. Such low prices are terrible news for Anthropic, which in recent weeks has been forced to execute a kind of “reduflation” of their new modelswhich are not more expensive but consume many more tokens. We’ll have to see if DeepSeek v4 Pro is as good as the company promises, but if it is, we’ll have another “DeepSeek moment” before us. Maybe not as notable as last year’s, but equally relevant. In Xataka | DeepSeek promised them happiness as the great Chinese AI. I didn’t count on a small detail: Kimi

Today torreznos are a delicacy of Spanish gastronomy. For years they were despised for being a shepherd’s dish

by

The pattern is so consistent it’s almost laughable: the lobster “it was a punishment” for servants and prisoners In colonial New England, oysters were Dockers’ food in Victorian Londonthe oxtail was second meat, what you took home when there was no sirloin. All current haute cuisine is built, to a large extent, on the recipe book of survival. But the case of the torreznos is even more interesting. A day for history: November 19, 2024. On that day, the European Union enrolled the Protected Geographical Indication (PGI) “Torrezno de Soria”. A PGI is, in essence, a seal that recognizes the reputation of a product. In this case, it recognizes the way in which the people of Soriano salt, marinate and cure white pork belly. A preparation that, moreover, moves almost five million kilos of bacon every year (growing almost 30% a year) and more than 20 million a year. Why is it more interesting? Today, Soria torrezno is consolidating commercially at a national and international level and it is curious because this process coincides with the end of domestic slaughter. And it is not an impression: the data from the Health area of ​​the Government of Castilla y León point to a 64% reduction in the number of home killings in the last decade in areas like Tierra de Campos. The interesting thing is that this decrease in slaughter is, in a way, the necessary condition that allows the industry to develop. This is how culinary nostalgia works: the same people who eat it today in trendy bars are usually the grandchildren of those who ate it not because they liked it, but because there was nothing else. Am I implying that the torrezno are not good? It wouldn’t occur to me. Only, as always, the story is more complex than it seems. The first written reference we have (or that, at least, I have been able to find) is a reference to the usefulness of torreznos to identify insincere converts. Then much more appears: bacon in Spain was not something frowned upon, it was something central to the diet… for ideological reasons. It was when the obsession with old Christians disappeared that the torreznos began to become a stronghold for the poor and shepherds. From there, the story (as I said) is a classic: offal, barnacles and sea urchin have gone from being ‘offal’ to being gourmet delicatessen. The torrezno too. Eating is something full of ideas. That is perhaps the most revealing thing about the Torreznos case, the confirmation that we eat with our mouths, yes; but above all we eat with ideas. The torrezno has only been able to be renamed as something gourmet once it has ceased to be anyone’s food out of obligation (out of that economic or political-social obligation). Maybe it’s the right time to think about how we think about food. Image | DAP In Xataka | Hearts, bowls of torreznos and raw milk: what the ancestral diet fad consists of

240 km without curves, in the middle of the desert and with truck traffic

by

Imagine driving for more than two hours without turning the steering wheel even a single degree. No curves, no noticeable slopes, no changes on the horizon. That is the reality of Highway 10 (Highway 10) of Saudi Arabia, which holds the Guinness record as the longest straight road on the planet with a completely linear section of 240 kilometers. A highway born for a king. Highway 10 stretches 1,480 kilometers from Ad Darb to the border with the United Arab Emirates, but it is its segment between Haradh and Al Batha that has received all the attention. The road was originally built as a private road for King Fahd of Saudi Arabia, although today it has become a fundamental artery for the transport of goods between the center and west of the country with the Emirates. The Empty Quarter desert as a setting. The road crosses the Rub’ al Khaliknown as the Empty Quarter, the largest sand desert in the world. The area itself explains why it is possible to build such a straight line: there are no mountains to surround, valleys to cross or geographical features to avoid. Just sand and more sand as far as the eye can see. The infrastructure is completely paved and has mainly two lanes in each direction, supporting intense truck traffic that crosses the desert. Speed ​​limits adjusted for heavy traffic. The maximum speed allowed on this highway varies depending on the type of vehicle: passenger cars can travel up to 120 km/h on fast sections, buses 100 km/h and trucks 80 km/h. Although in 2018 were announced Upper limits of up to 140 km/h for light vehicles in certain sections, the constant presence of heavy transport makes maintaining these speeds complicated in practice. A mental challenge more than a physical one. Believe it or not, driving on the straightest road in the world is not as easy as it seems, especially due to fatigue. The monotony of the desert landscape and the total absence of visual stimuli can cause drowsiness and even a dangerous disconnection while driving. Added to this is the occasional threat of camels wandering across the road. So, although the route is ‘easy’ to handle, mentally it can become a nightmare. Not for nothing is it found in Dangerous Roads website. Reinforced security measures. Aware of the risks involved in driving on such a monotonous road, the Saudi Ministry of Transport and Logistics has implemented various improvements safety features, including paved shoulders, reflective pavement markings (known as “cat’s eyes”), protective barriers, kilometer signs, and directional and warning signs. Here the driver’s attention must be vital, especially on a road with so few changes. Other legendary straights. Before Highway 10 snatched the title, the Australia’s Eyre Highway boasted the record with a 146 kilometer straight stretch through the Nullarbor Desert. Although almost 100 kilometers shorter, this Australian road remains one of the most unique driving experiences on the continent. Also noteworthy are roads such as ND-46 in North Dakotathe United States, or some sections of the Argentine Route 40which although they do not compete in length of absolute straightness and offer a great variety of landscape that softens the eye, also encompasses endless kilometers of monotonous movement. Cover image | City Vibes In Xataka | Yes, the V16 beacons transmit your position in the event of an accident. No, the DGT cannot “spy” on you with them A version of this article was published in 2025

with the ‘Hero of the Seas’, Royal Caribbean believes that

by

At Royal Caribbean they continue to explore a very specific idea: that the cruise stops being just the place from which you travel and becomes, in itself, the great argument of the trip. That’s where it comes in Hero of the Seas, presented by the company as the fourth ship of the Icon class. Rather than stopping at the beginning of the list of novelties, what is relevant is the movement it reflects: when it seemed that this type of boat had already taken its formula very far, the move points to a new attempt to take it one step further. The Icon class already has the Icon of the Seas and Star of the Seas at the top, while below are giants of the Oasis class such as Wonder of the Seas and Utopia of the Seas. This helps us read the company’s move: Hero is not presented as a discrete evolution, but as a direct continuity of the range with which Royal Caribbean has reserved its largest ships and its densest offering for the family audience. If there is something that we already know well about the great Royal Caribbean cruises, it is their way of divide the ship into neighborhoodsalmost as if each area had its own character. Hero of the Seas will follow that same scheme with eight neighborhoods on boardbut the company wants to accompany it with more layers of offer: nine swimming pools, 28 restaurant spaces and new accommodation options for multi-generation families. Seen this way, what is drawn is not just a large ship, but a structure that, on paper, seeks to fit different ages and rhythms within the same trip. As we can see, what the firm proposes for Hero of the Seas is a more segmented aquatic offer: a new Caribbean-style pool, a larger exclusive area for adults, improvements in bathing spaces already known from the Icon class and new water games for children. Royal Caribbean uses several proper names to dress up that proposal, but the substance is easier to explain. Divide these aquatic spaces into several areas aimed at different audiences and times. A ship that wants to look less and less like a ship The other great leg of Hero of the Seas has not to do with rest, but with reinforcing the offer of active leisure throughout the day. Royal Caribbean includes here two new family slides with floats and an update to Storm Chasers, to which it adds several elements that are already part of its usual recipe on large format ships: a combination of aerial walkway and zip line over the seasurf simulator, climbing wall, minigolf and sports courts. We are not facing a break with what the shipping company was already doing, but rather an expansion of that mix between water park, sports deck and resort leisure. The accommodation also falls into that segmentation logic that we have been seeing in the rest of the ship. Royal Caribbean talks about very attractive options for families, such as a three-deck tree house or rooms directly connected to areas designed for traveling with children, but at the same time maintains premium suites and more standard cabins. This allows the proposal to be read better: the shipping company combines more attractive accommodations with a broad base of options aimed at different audiences, from large groups to passengers looking for something more conventional. Beyond what it can offer on board, Royal Caribbean has already made it quite clear how it wants to put Hero of the Seas into circulation. The company has placed its premiere in August 2027 from Miamiwith seven-night itineraries through the eastern and western Caribbean, and with a fixed stopover at Perfect Day at CocoCay, its exclusive enclave in the Bahamas. From there, the ship will move between destinations such as Roatán, Cozumel and Costa Maya or Philipsburg and Charlotte Amalie, depending on the route chosen. It is also the part that completes the commercial fit of the project. Images | Royal Caribbean In Xataka | A millionaire has been living from cruise to cruise for more than 25 years: his biggest problem is not money, but balance

It has been the biggest marketing campaign for Apple. and free

by

I don’t know what someone who saw the moon landing felt. Apollo 11 in 1969but I do know that a couple of weeks ago half the Internet was glued to social networks to see everything the crew of Artemis II I was sharing. The moment in which the capsule returned to Earth It was exciting, but this time humanity had something unthinkable 50 years ago to follow the event. A iPhone 17 Pro Max. NASA didn’t want… When you travel 400,000 kilometers from home, you have to think very hard about what equipment you carry in your suitcase. Like any tourist, American NASA astronauts go with a camera, a Nikon Z9. It has its reason: NASA and the Japanese firm have an agreement to use these cameras. However, the most iconic photo of this trip wasn’t taken with a Nikon: it was taken with an iPhone. The Photon by Christina Koch With an iPhone 17 Pro Max, specifically. For years, NASA has refused to allow its astronauts to carry their cell phones because it is not essential for communications and, in addition, it can interfere with the systems of the ships and capsules. But For this occasion they relaxed the rules a little and things went well. The iPhone was going in airplane modeconnected solely to Orion’s internal network to transmit data to stations on Earth. And, seeing the success, they will surely repeat. The video. Because they are no longer just the photos: it is the tremendous video that Reid Wiseman, commander of the mission, has shared with the world. Before going into detail, here it is: It’s taken with that iPhone 17 Pro Max, using the 8x 406,771 kilometers away from Earth to capture how the planet was hiding behind the Moon. According to Wiseman, it is a focal length very similar to that of the human eye (you have to trust it, obviously) and the experience was like watching the sunset on the beach, but from the cosmos. For her part, astronaut Christina Koch also captured a video in which the iPhone looms mischievously in the reflection: The best camera. The iPhone is the mobile phone that captures the best video, although its telephotos are not even close to what they offer chinese mobile phonesnot on video either. However, as they say, the best camera is the one you always have with you. This is as true as a temple, and this case is a perfect example because the Nikon Z9 with its huge lenses could not fit through the gap in the docking hatch window to make that video. The iPhone does. Apple, so happy. After all the commotion caused by the photos taken on this mission (due to the quantity of them, the quality and how quickly they have been shared), we might think that Apple and NASA signed an agreement, but it seems that is not the case. Simply, it is one more tool than astronauts they came on board to document everything, since the agreement that does exist is the one they have with Nikon. If there was one with Apple, it would be public. But the truth is that Apple has just had the best ‘shot on iPhone‘ of his story. The company has been pushing that narrative of “we record our events with the iPhone because it is a beastly camera” for years, and now, four astronauts have put those iPhones to the test more than 400,000 kilometers from Earth. I have no doubt that in the September keynote we will see Apple take advantage of those photos captured with the iPhone on the Moon. And I can only think of how eager Chinese companies will be for their missions to start taking off so they can do the same. Images | Artemis II, NASA In Xataka | The mobile phones with the best camera 2026: next level photography in your pocket

Japan has just landed with a weapon to take down its shaheds

by

Sometimes wars change because of an unexpected solution to a seemingly minor problem. There we have the case of the Second World War, when the Allied pilots began to use simple aluminum strips launched from aircraft to confuse enemy radars, saturating them with false echoes and completely disrupting German air defense. The idea, as simple as it was cheap, showed that in certain conflicts the key is not to have the most powerful thing, but to find the most effective way to neutralize what already exists. Japan enters the drone war. Yes, Tokyo has taken an unprecedented step in the Ukrainian war by directly introducing proprietary technology on the battlefield, something unusual in its recent defense policy (although not for the future). Through the Terra Drone companyTokyo has not only invested in the Ukrainian Amazing Dronesbut has taken one of its systems from the laboratory to the real front. The result is a new type of cooperation where Ukrainian combat experience is combined with Japanese industrial capacity, creating a hybrid actor that did not exist until now in this conflict. An interceptor to knock down swarms. A proper name appears here. The key system is Terra A1 interceptor dronedesigned specifically to address threats like the Shahedthe same ones that Russia has used massively since the beginning of the invasion. We are talking about devices with speeds close to 300 km/h and a range of about 32 kilometersdrones that can detect and attack targets in the same mission cycle. Their advantage is not only in their features, but in their approach: they are designed to combat cheap drones with equally cheap solutions, avoiding the use of much more expensive missiles for lower value threats. The Terra A1 interceptor The cost war. Here is the key change of the conflict. While a Shahed drone can cost tens of thousands of dollars, such an interceptor can cost just a few thousand. Faced with this and how we have been countingtraditional systems such as anti-aircraft missiles can easily exceed one million per unit. This difference allows Ukraine to raise a volume based defensecapable of responding to massive attacks without exhausting strategic resources in each interception, something critical in a war of attrition. Ukrainian technology, Japanese industrial muscle. In reality, the alliance is anything but casual. Ukraine provides direct knowledge of combat, with systems adapted to electronic warfare, jamming and real front conditions. For its part, Japan provides production capacityfinancing and industrial scaling. The objective here is clear: to move from improvised or limited solutions to mass production capable of sustaining the pace of the conflict, all with a view to even exporting this model to other scenarios where cheap drones have become a dominant threat. Towards more autonomous drones. In fact, the next step is already practically defined and is none other than reduce human intervention. Current developments seek to ensure that these interceptors can take off, locate targets and attack automatically, without the need constant control. In theory, this not only increases efficiency, but allows you to respond faster to crowded attacks, where reaction time is key. In this field, the combination of Ukrainian software and Japanese technological development aims to accelerate a trend that is already transforming the air war into other conflicts like the Middle East. A new front for Russia. It is the last of the legs to analyze, because the arrival of the Terra A1 It means that Russia now faces a different problem than usual. These are no longer just traditional Western systems, but a new layer of defense based on cheap, scalable drones specifically adapted to your tactics. Japan’s entry into this field introduces an unexpected factorthat of a country with great technological capacity that is beginning to directly influence the balance of the battlefield, and does so by providing tools designed precisely to neutralize the type of weapon that Moscow more has exploded on Ukrainian territory. Image | Amazing Drones In Xataka | Ukraine is shooting down Russian drones with “pilots” 500 kilometers from the front: it is a radical revolution in war In Xataka | In 1914, submachine guns forever changed the way war was waged. In 2026, it’s algorithms’ turn

NASA’s alliance to finally understand dark matter

by

This week, NASA launched the Nancy Grace Roman Space Telescope, better known as the Roman Space Telescope. With its launch scheduled for September of this year at the earliest, it will become the space agency’s newest space telescope. It will coexist with others like Hubble or James Webbbut it has something that these don’t have. The ability to track vast expanses of the Universe at once. That’s what makes it special. Much more space. The Roman Space Telescope has 18 detectors that give it a panoramic view of space. It has been baptized with this name in honor of what is known as the mother of Hubble, for her important role in the development of this other space telescope. However, both have major differences. It is capable of looking at a field 100 times larger than that of Hubble. As a result, is expected that discovers tens of thousands of planets, billions of galaxies and stars and thousands of supernovae. An ideal companion for James Webb. The Roman Space Telescope also has advantages over the James Webb. If it is capable of analyzing a field 100 times larger than that of Hubble, in the case of James Webb exceeds it by 50 times. This allows you to observe without a clear objective on the part of the researchers. When exploring such large expanses, you may find something unexpected at any time. That’s where James Webb comes into play. And, although it can analyze less space at once, it is much more precise. Its mirrors are larger, so it captures more light and can discern more details. If the Roman detects something interesting, the James Webb analyzes it with a magnifying glass. Context matters. We have already seen that the James Webb can study the Roman detections with more precision. However, they can also help each other in the opposite direction, since the Roman is capable of providing context around James Webb’s objectives. Together to unravel dark matter. The biggest difference between the Roman Space Telescope and the James Webb compared to Hubble is that they can analyze space by focusing on emissions in the infrared spectrum, rather than visible light. As a result, it can see through cosmic dust, detect cold objects, and look further back in time. The latter is extremely useful for understanding the expansion process of the universe and, incidentally, unravel some mysteries about dark matter. The Universe expands. We have known for a long time that the universe is expanding. That is, the galaxies are moving away from each other, but not because they are moving, but because the space between them is stretched, like a balloon that is inflating. It is also known that this is happening more and more quickly. But why does it happen? It is not clear, but it is suspected that it may be due to dark matter. Supernovas that act as lighthouses. To better understand what is happening, it is important to measure very well how galaxies are separating. One of the ideal ways to do this is by using Ia supernova explosions as beacons. They are phenomena with a known maximum brightness, so they are used to measure distances, taking into account the analysis of their relative brightness from Earth or the place where a space telescope is located. The problem is that they only occur once every 500 years in the Milky Way. A telescope that measures in the infrared can travel very far back in time, but the James Webb only does so in small pieces. The Roman, on the other hand, can analyze such large areas that several of these explosions could be detected at the same time. That would allow several beacons to operate simultaneously to better map the Universe and understand why it is expanding as it does. Once the beacons were located, the James Webb would enter the game to do its detailed analysis. Together they can unravel very ancient mysteries of astrophysics. There is no one better than the other. Image | POT In Xataka | We have been studying the planets of TRAPPIST-1 for years with great hope. James Webb just knocked it down

We have stuffed the Gibraltar monkeys with Doritos. His solution has been to eat dirt as if it were omeprazole

by

On the Rock of Gibraltar, one of the most classic scenes that can be found without a doubt is that of the Barbary macaqueswho seem to be the absolute kings of the entire terrain. Here it is quite likely to see tourists end up stealing a piece of sandwich, chips or a sweet that they have in their hands, especially when the rules are not followed that are marked. But this, which may have its funny point, the truth is that it generates quite a few alterations in the animals, to the point of needing to eat dirt. It has been studied. That a macaque takes a bag of Doritos from a tourist and begins to eat it, among many other types of junk food, is something that has focused the attention of science. And it has reached such a point that it has materialized in a study published in Scientific Reports where numerous continuous episodes of geophagy have been documented among these primates. This is nothing more than the usual consumption of red soil, which is known as terra rossa, and even, in some specific groups, a strange mixture of dirt and tar. Its consequences. What at first glance may seem like simple animal curiosity is actually a survival mechanism, as researchers have found a statistical correlation between the consumption of this junk food, loaded with sugars, salt and dairy, and the frequency with which they eat the earth. In fact, the data shows that this practice increases in summer, when there are a greater number of tourists in the area, and decreases drastically when the influx of people and, therefore, the availability of junk food in the area decreases. Why do they do it? Here we must take into account that the stomachs of these macaques are not prepared for the junk and processed food that humans consume. This is why the earth can act as a gastric cushioning system and can be compared in our minds to a stomach protector. After a heavy meal, macaques use clay to purge themselves, absorb toxins and relieve the digestive discomfort caused by our diet. In addition to providing them with certain minerals that junk food lacks, scientists point out that swallowing dirt helps them try to rebalance an intestinal microbiome that has been altered by the excess of salt and sugar. A cultural tradition. The most interesting finding of this research is that eating dirt is not a mere isolated instinct, but has become an anthropogenically induced tradition of primates. That is to say, we are dealing with a behavior of social transmission, since the macaques learn from each other that eating dirt relieves the belly pain after stealing an ice cream from a tourist. This, logically, is a cultural component of self-medication that has arisen purely and exclusively due to the alteration of their environment at the hands of human beings, demonstrating the great capacity for adaptation that they can have to any situation that comes upon them. Images | wirestock 8photo In Xataka | We have been believing for 50 years that the Strait of Gibraltar was “closed” with an apocalyptic cataract. Now we have nuances

A survival thriller coming to Netflix tomorrow that pits Charlize Theron against a psychopath in the depths of Australia

by

A murderer chooses his victim carefully. That’s what Ben (Taron Egerton) does when he determines that Sasha (Charlize Theron), an elite climber who has entered the outback Australian to overcome a duel, is his next objective. What Ben doesn’t know is that Icelandic director Baltasar Kormákur has spent twenty years specializing in placing extraordinary actors in devastating environments. The result of this clash receives the poor translation of ‘Dominant predator‘in Spain (the original is the most elegant’Apex‘), premieres on April 24 on Netflix and is pure popcorn party. Shot entirely in real locations, the film put Egerton and Theron, who are joined by Eric Bana, to the limit of their strength. In fact, Theron suffered a broken toe during filming and continued filming without anyone on the crew knowing. A good example of how the film posed a challenge for its performers that went beyond the strong psychological tension that is drawn between the two rivals. In fact, The approach of ‘Apex’ is one of the oldest in action cinema: The short story ‘The Most Dangerous Game’ by Richard Connell, published in 1924, described an aristocrat who hunted humans on his private island, and inspired a total classic of the genre in 1932, ‘The Evil Zaroff’. From there, multiple variants of the “people hunting” trope, and which is often, as here, accentuated by a danger that is added to the relentless hunters: an environment that functions as a gigantic trap. ‘Apex’ director Baltasar Kormákur has spent two decades building a filmography around a single question: what does extreme adversity do to people? His most relevant survival films, ‘Everest’ (inspired by a real catastrophe in 1996), ‘Adrift’ (about a couple trapped in the ocean after a hurricane), and ‘The Beast’ (where Idris Elba confronts a lion in South Africa), have common characteristics: real locations, actors subjected to harsh physical conditions and landscapes that are both scenery and threat. This new Netflix exclusive ‘Apex’ plays in that same league. In Xataka | 16 premieres on Netflix: this week, the new ‘Stranger Things’, a rare British series and the return of Charlize Theron

We thought that AI was going to collapse the electrical grid. The solution is to “unplug” it 18 days a year

by

Daily headlines bombard us with the insatiable hunger for Artificial Intelligence, painting a future where data centers will devour our infrastructure. However, reality hides a fascinating irony: the same technology that clutters cables today could be our greatest ally. According to estimates of DeloitteAI will optimize global systems saving more than 3,700 TWh by 2030, almost four times the energy consumed by all data centers on the planet combined. But to get to that stage, you first have to turn on the machines today. And the solution is surprisingly analog. Paweł Czyżak, from the Ember analysis center and one of the most authoritative voices in the European energy transition, sums it up with a simple idea: A data center does not need to operate at full power every hour of the year. In the face of system collapse, the industry’s new survival dogma is clear: “Connect now and operate flexibly.” The heart attack of the network. We have been victims of what we once defined as “tyranny of 24/7”. Algorithms do not sleep and demand uninterrupted supply. This voracity has caused a heart attack in the traditional data epicenters in Europe (the “FLAP-D” markets: Frankfurt, London, Amsterdam, Paris and Dublin), almost completely paralyzing new deployments. The bottleneck is no longer the latest generation microchips; transformers and free electrons are missing. Added to this physical collapse is the bureaucratic one. The European University Institute (EUI) warns that connection queues are a critical funnel: in countries such as the United Kingdom or Italy, the requested capacity exceeds the peak of national maximum demand by more than 10 times. All of this is aggravated by speculative “zombie” projects that block entry to legitimate developers. The obstacles are, as detailed in the recent study by Camus, encoord and Princeton ZERO Laba double wall: there is a lack of cables for day-to-day operations and a lack of clean capacity built to provide backup. Flexibility as a lifesaver. Is it possible to “turn off” part of the AI ​​brain without the system crashing? Yes. A recent trial led by Nebius, Emerald AI and National Grid showed that an AI cluster was able to cut its consumption by 30% in just 40 seconds to relieve the network, keeping critical tasks intact. Even Google already boasts of having reached 1 GW of “demand response” by combining batteries and the ability to move loads between regions. As Czyżak explainsmoving just 5% of the load (the equivalent of a few critical hours per year) unblocks the grid massively. In fact, this strategy would save more natural gas than a country like Denmark consumes in electricity generation, by preventing electricity companies from having to turn on expensive and polluting combined cycle plants to cover demand peaks. For its part, the Camus and Princeton report proposes to scale this with two mechanisms: Flexible connections: The center operates normally 99% of the time, but in the scarce 40 or 70 hours a year of extreme network saturation, it reduces its computing or draws on its own batteries. BYOC agreements (Bring Your Own Capacity): Big tech finances its own clean energy capacity instead of waiting for the state to modernize infrastructure. The combination is magical: it reduces the wait to connect to the network from 7 to just 2 years. For a technology company, this means starting to bill three years earlier, generating net returns of between 1,000 and 4,000 million dollars per site. The citizen will not pay the bill. On a social level, the transition towards this flexible model brings excellent news for the average citizen. The detailed modeling of Princeton’s ZERO Lab confirms that a flexible data center (under BYOC schemes) assumes practically all of the incremental costs it generates to the electrical system. In other words, the billions needed to host the cloud will not be transferred to household electricity bills. On the contrary, by making the most of the existing network instead of building massive new lines, the fixed costs are distributed among more actors. In Spain, organizations such as the CNMC are already applying “flexible access permissions”forcing by law to accept controlled cuts in emergencies to protect the stability of the country. The plug that will rule the world. In the frenetic geopolitical and business race to dominate the future of Artificial Intelligence, the narrative has changed. It is no longer enough to design the fastest microchip or have the most brilliant engineers. Today absolute victory belongs to whoever has a free plug. But rather than desperately burning gas or waiting a decade for governments to bury thousands of kilometers of copper, the industry has found a pragmatic way out. Demand flexibility from Big Tech Not only does it allow them to turn on their servers years earlier; It protects citizens’ bills, squeezes the infrastructure of the 20th century and banishes the dangerous ghost of a Europe forced to relapse into its old addiction to fossil fuels. Image | Photo by Scott Rodgerson on Unsplash Xataka | There is no energy for so many data centers and the consequence is clear: half of those planned for 2026 in the US are in danger

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.