rivals

make 90% of your rivals at 20% of your price

by

we carry a pretty crazy week regarding AI model releases. The spotlight may have gone to the long-awaited GPT-5.6 family (with its Sol, Terra, Luna variants), but we have also seen how SpaceXAI Grok 4.5 was released and Meta too was renewed with Muse Spark 1.1. In both cases we are facing a promising qualitative leap, and according to internal and independent benchmarks the performance of both models is already remarkably close to that of the frontier models from OpenAI and Anthropic. Grok and Muse Spark get going That was unthinkable a few months ago. Both Meta and xAI (now fully merged with SpaceX in the new SpaceXAI) seemed to play in the second (or third) division in the AI ​​​​segment. SpaceXAI has barely shared results from Grok 4.5, but the table with Muse Spark 1.1 would look like this. Source: Meta (AI modified to include Grok). In fact, Meta had been working on a mysterious new model called Muse Spark that we finally met in Aprilbut the result was lukewarm: it stood out in practically nothing and it seemed that the firm had not fully met expectations. Something different happened with Grok. Musk’s ambition in this area has always been enormous, but from the beginning Grok’s approach was strange. That first version seemed to want to laugh at itself (and at all of us) with his sarcastic tonebut its evolution showed that we were facing a model with remarkable potential that I couldn’t quite find the key. However, Both companies have given a more than promising push to these models. At SpaceXAI they have probably benefited from the recent acquisition of Cursor, which had its own background with Composer 2.5 and which has undoubtedly influenced the creation of a much more competent Grok 4.5 when it comes to programming. With Meta and its Muse Spark 1.1 the same has been seen: it is much closer to the great rivals, Fable 5 and GPT 5.6. But the striking thing is not that they have come closer, but that they have done so by cutting the price significantly. And that is its great asset. It’s not more for less, it’s almost the same for much less. What is really relevant about these models is not only that the leap in quality has been significant: it is that Its cost is really competitive. In SpaceXAI’s announcement it is in fact something they especially highlight by saying that “Overall, Grok 4.5 offers the greatest intelligence per unit of time and cost.” At the launch of Meta there are no direct statements about this, but there is no need: it is the cheapest frontier model among its rivals, as can be seen in this table: The question is, of course, whether those prices really pay off. If we get “enough intelligence” per dollar, to understand each other. And that is precisely what the Artificial Analysis ranking which evaluates the cost of completing predefined tasks in its benchmarks. At the moment Muse Spark 1.1 is not included in that ranking, but for now there are two clear winners: on the one hand, GPT-5.6 Luna. On the other hand, even better, Grok 4.5. Look at the green quadrant at the top left: the only models that fit are GPT-5.6 Luna and Grok 4.5. Muse Spark 1.1 is still untested. Source: Artificial Analysis. That conclusion seems to make it clear that those specific SpaceXAI and OpenAI models are fantastic for intensive tasks where token consumption skyrockets. They may not be ideal If you are looking to solve very complex mathematical conjecturesbut for agentic tasks we have at least those two clear candidates. Muse Spark 1.1 probably meets those premises as well, judging by the benchmarks and cost. We are therefore facing an interesting turn of events. One that also poses a real alternative to what was until now the great asset of Chinese open models: its cost was much lower than the frontier models of OpenAI or Anthropic, but performance was also lagging behind. What Grok 4.5 and Muse Spark 1.1 (and GPT-5.6 Luna) offer is something like the best of both worlds: they manage to “give us” 90% of what the best models in the world do, but they do it at 10 or 20% of their price. Good, pretty and cheap. Not a bad bet, of course. In Xataka | A teacher suspected that his students had cheated with AI. He took the exam in person and his grades dropped by half.

They have five days to open themselves to rivals

by

If there is something that the European Commission does not like at all, it is that technology companies use their power to create monopolies. It takes years actively fighting these practices with chases and historical fines. The objective is to promote competition and when Meta introduced your AI chatbot on WhatsApp, came under the scrutiny of the Commission. Now, have taken a historic decision: order Meta to open WhatsApp to any rival chatbot. And that opens the door, for example, for a Mistral or a ChatGPT to sneak into WhatsApp. In short. This Tuesday, June 9, the European Commission ordered the American giant to restore free access to WhatsApp for competing AI assistants. It is about reversing a situation that Meta has been making more and more complicated for the competition since October 2025. Today, we require Meta to restore access to WhatsApp for competing AI assistants while we investigate whether the restrictions may infringe EU competition rules – Teresa Ribera, European Commission Previously, other companies had access to the WhatsApp API, but Meta changed its conditions to block rival AI services on the platform, something that began to be applied on January 15 of this year. Their own AI chatbot had arrived and they didn’t want anyone stepping on their toes. The Commission did not begin to investigate this out of nowhere, but as a result of complaints from several AI assistant companies who reported that Meta was taking advantage of its position of power and dominance of messaging platforms to ‘sneak’ a single chatbot to everyone: its own. Historical. Meta has five business days to restore that access to rivals and, almost as important as the measure for the user (who will be able to choose which AI system they use in WhatsApp), is the way in which it has been taken. Because it represents the first antitrust precautionary measure that the European Union establishes since 2019 because the investigation has not really ended. The competition commissioner of the European Union, Teresa Ribera, pointed out that this precautionary order is necessary so that competition between companies is fair in these rapidly evolving markets. He assures that, if such a measure is not taken, the damage would be “almost impossible to repair” and assures that they continue to review whether the restrictions that Meta applied may violate EU competition laws… or not. That is to say, the European Commission has been investigating this case for six months and they are not finished, but they have already made the decision to order WhatsApp to open its API. Meta’s Response. Not only the European Commission got involved: Meta was already being investigated for the same reason by the Italian competition authority. Since then, the company has taken some actions to prevent an order like this from arriving, offering access to its rivals for a fee and, just a few weeks ago, offering free access to the API up to a certain threshold and, when they exceeded it, starting to charge for its use. Neither the complainants nor the Commission accepted these measures because they considered that, in practice, it was exactly the same as what Meta was already applying: it was not free access, but, in addition, you had to pay to join the platform. “This will allow free access to OpenAI and some of the largest companies in the world. It is a regulatory overreach subsidized by the many European companies that pay” – Meta Spokesperson Obviously, Meta is not amused by this situation one bit, pointing that the EU is using its power to allow some of the world’s largest companies to use its (paid) WhatsApp Business product for free. They claim that Europe is playing along with OpenAI. Whistleblowers, such as The Interaction Company, do seem satisfied. Now…what. Well there are three options. Or Meta relents and opens its API so that anyone can enter WhatsApp as an AI chat (like when Google in its browser It asks us which search engine we want to use by default Instead of assuming we want yours, go) or pay a fine. That fine is not small: up to 10% of your global annual turnover if you do not comply with these provisional measures. The third option is for Meta to appeal the precautionary order before the courts of the European Union. The Apple thing. As we say, the decision of the European Commission is historic because it is a precautionary measure while they ensure that they must continue studying the case. They give Meta only five days to open its tools and let the competition ‘sneak’ into their home and then to wait to see if the decision is ratified or if the Commission backs down. In any case, it is something that comes just when Apple has once again raised its tone against Europe after the presentation of the new Siri AI by announcing that many of its functions will not be available on iPadOS 27 and on iOS 27 due to EU antitrust policies. But, as with Meta, it is not a fight for privacy (as Apple wants to sell), but rather one for control of its platform and its product so that there is no competition and no one else can play. In Xataka | No, WhatsApp Meta AI cannot be deactivated, but this way you can make it bother you as little as possible

Mercadona’s engine is not the white label, but crushing its rivals in profitability by earning less per product

by

You may like it more or less your cataloghis business strategy or even the forecasts culinary-apocalyptic of its president, but there is something undeniable: Mercadona long ago stopped being a chain of stores to become a social phenomenon. One who depends about 30% of the food distribution market, 51% of the business of prepared dishes and that sets the pace for trends as relevant as that of the merchants. Hence everything that revolves around your finances be interesting. Especially because when studying them in detail and comparing them with other competing chains there is one fact that draws attention: Although its gross margin per sale is lower than that of other rivals, its profitability ratio is much higher. The data: 41,858 million. When Mercadona presented its 2025 economic balance, two months ago, there was a figure that made headlines: 41,858 million of euros. That was the company’s consolidated turnover, an interesting fact because it shows an annual growth of 8%, but it actually hides other even more revealing values. One of them is the turnover, net sales, which amounted to 38,178 million. In that case the increase was 7.2%. If we subtract from that figure what it cost the company to supply its merchandise (28,639 million), we obtain the first relevant piece of information: its gross marginthe money that the company earned after deducting the costs directly attributable to production and sale, such as raw materials or expenses generated during manufacturing. In this case it stood at 9,539 million. Is it important information? Yes. Above all to understand how the Valencian chain makes money and where it has its strengths (and weaknesses) compared to the competition. At first, those 9,539 million may not tell us much, but a few days ago Five Days subjected him to an analysis which does leave a couple of interesting ideas. The first is that this figure shows that Mercadona’s gross margin represents 25% of its sales. That means that of every 100 euros you earn, the supply costs take 75. From the remaining 25 euros you must get enough money to cover other bills and, above all, generate profits. It is not a bad percentage (25%) if we compare it with what the Valencian firm registered in recent years, but it is significantly lower than that managed by other competing companies. The calculations of Five Dayswhich are based on the accounts published by the companies, conclude that this margin rises to 26.2% in Dia, 27% in Eroski and 30.1% in Consum. In theory, this comparison leaves a clear reading: any of these three chains has a larger cushion, once the supply costs have been deducted, to pay the rest of the company’s bills and generate profits. And the surprise comes. The curious thing is that this ‘photo’ changes when we delve a little deeper into the accounts of Mercadona and its competitors. If we look at the operating resultwhich deducts all operating expenses, including for example salaries, rents, advertising, depreciation, transportation or energy, Mercadona is left with 2,061 million of euros. Given that Juan Orig’s chain invoices significantly more than Dia, Eroki or Consum, that operating result is also much higher in net terms. That’s logical. The curious thing is that it is also true in relative terms, based on the total income of each firm. In Mercadona this margin is 5.4% while in the case of Día it drops to 2.6%, in Eroski to 4.6% and in Consum to 2.7%. That’s the first surprise. The second comes when we go one step further and look at the net profitalready discounted the financial and fiscal expenses. It is relevant data because it basically shows what the company ‘earns’, the remainder from which the company takes the money with which it then pays dividends to its shareholders and makes reinvestments. In 2025 that benefit was almost 1,729 million, which is equivalent to 4.5% of its turnover. In Dia this percentage of global sales is 2.3%, in Consum it is 2.6% and in Eroski it remains at just 0.9%. Beyond the numbers. This mixture of percentages can be somewhat confusing, but it is very simple to read: beyond the business volume of each chain, whether it closes the year with more or less millions invoiced, Mercadona has achieved an important milestone. Despite ‘earn’ less per product Than Dia, Eroski or Consum (gross margin), their profitability ratios are much better. How have you done it? In your annual report The firm assures that it has improved its profitability thanks to the “optimization of processes”, which includes energy savings, “elimination of expenses without added value” and “advances in operational efficiency.” Only the use of ovens in ECO mode saved him two million. Outrunning the giants. Mercadona’s formula has not only allowed it to stand out from its most immediate competitors. It has also done so in comparison with other heavyweights in the sector internationally. At least in relative terms. a few weeks ago Expansion public an analysis which shows that the Valencian chain has skyrocketed its net profit margin to such a level that it surpasses giants such as Walmart, Costco or Tesco in profitability. While Mercadona’s net margin is 4.52% (4.52 euros profit per 100 euros in sales), at Walmart it is 3.1%, at Costco 3%, at Tesco 2.52%, at Ahold Delhaie 2.45%, at Dia 2.26%, at Sainsbury’s 0.73% and at Kroger it remains at 0.69%. And that’s just to name a few cases. The Valencian firm not only stands out in the photo finish With respect to its competitors, the figure for 2024 also significantly improves, when the net margin was 3.88%. Images | M. Peinado (Flickr) and Mercadona In Xataka | The gap between what pork costs on farms and in supermarkets does not stop growing. The ranchers have said enough

This is how it compares to its rivals in price from POCO, Xiaomi and Samsung

by

Yesterday, Google presented its new economical mobile phone, the Google Pixel 10a and whose greatest bet is photography. Available from 549 euroshas several competitors in the market who will try to tread on an almost guaranteed ground of success (seeing the popularity that Google mobile phones have been acquiring for a few years). Among them are the Xiaomi 15The Poco F8 Pro and the Samsung Galaxy S25 FE. If you have questions about which of these mid-range mobiles but with top features is better for you, we are going to compare them in some of their functionalities, so that you can make the right decision. The price could vary. We earn commission from these links Pixel 10a vs Xiaomi 15T, Poco F8 Pro and Samsung Galaxy S25 FE, at a glance Google pixel 10a xiaomi 15t Poco f8 pro samsung galaxy s25 fe screen 6.3″ OLED 120 Hz, 2700 nits 6.83″ AMOLED 3200 nits 6.59″ AMOLED 3500 nits 6.7″ Dynamic AMOLED 1900 nits processor Google Tensor G4 MediaTek Dimensity 8400 Ultra Snapdragon 8 Elite Samsung Exynos 2400 photographic system Dual main 48 MP + ultra Triple with 2x TV Triple with strong main sensor Triple with 3x TV BATTERY 5100mAh 5500mAh 6210mAh ~4900mAh PRICE From 549 euros From 499 euros From 519.99 euros From 544 euros 10 GOOGLE APPS THAT COULD HAVE SUCCESSFUL How are all these phones different? Design and screen Although today there are few brands that innovate in terms of the design of their terminals (see the case of Nothing) it is true that among all these models you can see design differences that can make you opt for one or another smartphone. The Google Pixel 10a is a mobile phone with a very flat and minimalist design, with a 6.3 inch compact size which makes it a perfect option to hold with one hand. On the other hand, the Xiaomi 15T has a 6.8-inch screen, for those looking for a greater sense of immersion, especially when watching multimedia content. For its part, the Xiaomi Poco F8 Pro has a 6.59-inch screen and also has a slightly larger design than the Pixel 10a, although its edges are soft; Yes, the plastic finish does not give it the feel of a premium mobile. Finally, from the design of the Samsung Galaxy S25 FE, it can be noted that it also has a large 6.7-inch screen and a more refined and premium design than the other models. Your photography system The Google Pixel 10a was born with the intention of offering a excellent photography experience within the mid-range. Of this model, we can highlight its own photo app, with AI that considerably improves your photos. Its main lens is 48 MP accompanied by a 13 MP ultra wide angle lens. Although yes, it does not have a dedicated telephoto like its older brothers. Compared to the Google model, the Xiaomi 15T offers greater versatility in photography, thanks to its triple rear camera. Plus, they are made by Leica, so you can get more artistic photos. As for the Poco F8 Pro, it incorporates a telephoto and offers 8K recording, although the Pixel surpasses it in color balance and night processing. Finally, the Samsung model presents a more advanced photographic system than Google’s, thanks to its 3x optical telephoto, which gives us greater flexibility when taking photos from a distance. Battery and charging The Google Pixel 10a battery has a average capacity of 5,100 mAh and provides autonomy for a full day using navigation, social networks and spending time on video and photos. Of course, you should know that this group of phones is not the largest battery. The Xiaomi 15T has a capacity of 5,500 mAh, so its autonomy is superior even in intensive use. Regarding the Samsung model, its battery capacity is about 5,000 mAh (4,900 mAh according to its technical sheet), so the results are very similar to those of the Google Pixel 10a, although it may become more efficient thanks to Samsung’s optimization. Finally, the big winner in terms of battery is the Xiaomi Poco F8 Pro, with a capacity of 6,200 mAh, giving us more than enough autonomy even if you use the phone for gaming. All of these smartphones offer fast charging, although it is different in each of them. The Pixel 10a supports wired charging up to 45W, thus improving on its predecessors and charging it in half in about 30 minutes. The Xiaomi model supports it up to 90 W and charges to 100% in 50 minutes. For its part, the Poco F8 Pro supports fast charging of up to 100W, which makes it possible to fully charge it in about 37 minutes. Finally, the Samsung phone is fully charged in about an hour. Their brains are different too The processor is the brain of a mobile phone and in this group of phones, each one has a different one. The Google Pixel 10a has the Google Tensor G4whose main strength is its optimization of Artificial intelligence and good daily performance, although it is not the most powerful or the best for gaming. A MediaTek Dimensity 8400 Ultra is the processor of the Xiaomi 15T, which is perfect for multitasking and demanding appsalthough it is true that it gets hot in very intense sessions and is not as efficient as Google’s for use with AI. The most powerful processor on the list is the Snapdragon 8 Elite that the Poco F8 Pro has, which is ideal for powerful games and very demanding tasks, although it may be more than necessary for users who do not need that extra power. Finally, the Samsung mobile has a Exynos 2400which provides a good balance in general, although it is not so good for demanding gaming. Operating system and updates The new Google Pixel 10a comes with Android 16 pure and receive updates for seven years. Without a doubt, it is the champion of Android updates, becoming the phone par excellence for those who want … Read more

Qwen3-Max-Thinking rivals Google’s Gemini 3 Pro more than ever. The key is in what is not being told

by

There are days when it feels like we open the phone and the dashboard changes again. Since ChatGPT broke out in November 2022the artificial intelligence race has continued to accelerate, and every few weeks a new model appears which promises to push the bar a little further. Sometimes it is an update, other times it is a “flagship” with a different surname, but the pattern repeats itself: more power, more ambition and an increasingly global story. In this context, China is gaining visibility in an increasingly evident way, and the name that is now entering the conversation is Qwen3-Max-ThinkingAlibaba’s proposal with which it wants to play in the same league as the great references of the moment. At first glance, Qwen3-Max-Thinking might seem like just another name in the endless list of models. But there is a relevant nuance here: he presents it as his star model for reasoning tasks, and explicitly places it in the same conversation as Gemini 3 Pro. The company says it has scaled parameters and invested computing resources in reinforcement to improve several dimensions at once, from factual knowledge and complex reasoning to instruction following, alignment with human preferences and agent capabilities. In other words: you are not just selling raw power, but a way to “think” better. What benchmarks teach To land that promise, the most useful thing is to look at the comparative table that we have in hand, with 19 benchmarks and a direct count: Gemini 3 Pro leads in 11, Qwen3-Max-Thinking does it in 8. This data, by itself, does not decide “who is better”but it does help to understand the type of fight that Alibaba poses when faced with Google. Here it is worth being very literal with what we are measuring: each benchmark focuses on a specific skill, from general knowledge to programming, use of tools, following instructions or long context analysis. If we look for the point where Qwen3-Max-Thinking really hits home, there is one that stands out above the rest: following instructions and aligning with what humans prefer in a conversation. In Arena-Hard v2Qwen wins with 90.2 compared to Gemini’s 81.7, which is the largest difference in its favor in the entire table (8.5 points above). It is not a minor nuance, because this type of benchmark does not reward only the technical “success”, but rather the final result that a person considers most useful when blindly comparing answers. Added to that IFBenchwhere Qwen wins by the minimum (70.9 versus 70.4). Translated into real life: when the user does not formulate a perfect instruction, when the assignment has ambiguity or requires interpreting intent, Qwen seems more oriented to nailing what is asked of him and doing it in a way that feels natural. The other area where Qwen supports his “thinking model” narrative is mathematical reasoning and logical problem solving. On HMMT, in both the November 2025 and February 2025 issues, Qwen is ahead (94.7 vs. 93.3 and 98.0 vs. 97.5, respectively). And in IMOAnswerBench it also wins, although by a minimal margin: 83.9 versus 83.3. These numbers do not suggest a beating, but they do suggest a consistent pattern: when the problem demands several steps of logic and it is not solved only with memory or a nice answer, Qwen tends to take advantage. To these improvements Alibaba adds a component that is already becoming the new standard: that the model does not remain in the text, but can act. In its presentation, the company talks about an adaptive use of tools that allows information to be retrieved on demand and a code interpreter to be invoked. And this orientation also appears in the benchmarks: in HLE (w/ tools), Qwen wins with 49.8 compared to 45.8 for Gemini, which suggests a better ability to perform when the model can rely on external tools. Here the fundamental change is important: it is no longer just “what he responds”, but how he investigates, how he decides what tool to use and how he synthesizes what he finds. There is a part of this comparison where the Gemini 3 Pro feels more “engineer” than “conversational,” and it is precisely where many professional users put the focus. The Google model wins in MMLU-Pro and MMLU-Redux, two tests closely associated with general knowledge, and also in GPQA and HLE, which in this table appear as demanding evaluation benchmarks and complex questions. In code, Gemini prevails in LiveCodeBench v6 and also in SWE Verifiedwhich reinforces the idea that, for programming tasksis still a very solid bet. Added to this is AA-LCR, where it leads in analysis of long documents. The fine print hides beyond the price At this point, there is a question that weighs as much as any benchmark: how much does it cost to use these models seriously. In standard prices per 1M tokens, the contrast is clear. On Gemini 3 Pro, the entry moves between 2 and 4 dollars depending on the tranche of input tokens, while in Qwen3-Max The input is listed at $1.2. But the most important difference appears at the output, which is where the “thought” of the model is paid: Gemini marks 12 to 18 dollars compared to the 6 dollars of Qwen. Translated into proportions, in standard use Gemini is approximately 1.67 times more expensive in entry and 2 times more expensive in exit in the usual section. If the tranche exceeds 200,000 entry tokens, the distance increases to 3.33 times in entry and 3 times in exit. Gemini is approximately 1.67 times more expensive on entry and 2 times more expensive on exit in the usual section. And here we come to the part that is usually left out of the conversation when everything focuses on power and price: what happens to your data when you use the model, and under what rules. In the case of Qwen, two worlds must be clearly separated. On the one hand there is the consumer web chat, whose terms They contemplate the use and storage … Read more

Microsoft had the deal of the century on its hands. A break of a year and a half was given to one of his rivals on a platter

by

With its early deal with OpenAI, Microsoft was leading the AI ​​race in 2023. A year later it froze its expansion. Now Oracle serves OpenAI models and competitors share what Nadella’s company rejected. Why is it important. This isn’t just about lost data centers. Microsoft has assigned contracts with OpenAI valued at $420 billion to Oracle, equivalent to $150 billion in gross profit over five years. That would have increased its annual profitability by 18%. This means that in addition to losing growth, Microsoft also financed the entry of a rival into the most profitable business of the decade, according to analysis by Semianalysis. The facts. In 2023, Microsoft multiplied its investment in OpenAI tenfold to $10 billion and broke ground on the largest data centers ever built. Represented more than 60% of all infrastructure leases cloud among the greats. In 2024 it stopped everything in its tracks. It canceled 3.5 gigawatts of planned capacity — enough to power 2.5 million homes — and projects in a dozen countries. Its share of contracts fell below 25%. Between the lines. The company has used the argument of financial prudence: it did not want OpenAI to represent 50% of Azure’s revenue with lower margins than the traditional business. But the reality is simpler: he couldn’t keep up: OpenAI demanded a speed that Microsoft couldn’t match. Yes, but. The company has returned to the market with some urgency. The problem is that the options have been running out. Now rents capacity to neoclouds —specialized companies that build infrastructure—to resell it to third parties. It is a business with worse margins. The company that refused to build now pays commissions for having miscalculated. The money trail. Oracle is not the only winner. CoreWeave, Google, Amazon, Nscale and SB Energy have signed large contracts with OpenAI. In 2025, the story of OpenAI has been the story of its diversification away from Microsoft, although it is true that What seemed like a bad divorce ended in a separation of assets with forced smiles. The world’s most valuable AI lab had to fragment its infrastructure across multiple vendors because its original partner couldn’t—or wouldn’t—scale. In applications, Microsoft’s historical dominance with GitHub Copilot is also eroding. There are startups that have built more integrated code editors and scaled beyond Copilot. Microsoft has been forced to add the models of its rival Anthropic on GitHub Copilotwith a brutal cost for their margins. The company that had exclusive access to OpenAI now depends on its competitor to keep its code editor relevant. And now what. Microsoft has until 2032 before its agreement with OpenAI expires. It has Copilot with 100 million users. You have Office 365, Azure, and a business ecosystem that no one else can match. But the “great pause” of 2024 will take years to heal. The company has bet that the future of AI will be enterprise – with security and localization requirements – and not centralized in remote megacenters. You may be right. But 18 months of technology advantage is worth billions. And Microsoft just gave them away to its rivals. In Xataka | OpenAI has to pay debts of $400 billion in 2026. Nobody has the slightest idea how it is going to pay them Featured image | Simon Ray in Unsplash

Spend more on R&D than any of its rivals

by

Intel has not just lifted. With An intractable nvidia in the segment of artificial intelligence and with a Unattainable TSMCthe company tries to adapt to The new times. The hope is that the latest capital injections help: in recent days we have seen how Softbank invest 2,000 million euros euros in it and then the United States government has bought 10% of Intel to save it from burning. However, the company Directed by Lip-Bu Tan He is putting all the meat on the grill to try to do it. It is demonstrated by the Spectacular amount of money invested in your R&D department. The surprising thing is not to invest a lot of money; That is expected in a technological company. The really surprising thing is that he invests much more than his rivals. Intel invest more than anyone in R&D, but be careful: their rivals squeeze the step An analysis by Techinsights cited In Korea Joongang Daily It collects the financial results of Intel and some of its rivals in 2024 to study its investments in R&D. According to these data, these were those amounts for various companies: Intel: 16,546 million dollars (PDF) Nvidia: 12,914 million dollars (Nvidia) Samsung: 9,500 million dollars (Joongang) TSMC: 6.5 billion dollars (Joongang) AMD: 6,456 million dollars (PDF) That means that Intel invests 28% more than Nvidia – which right now has a much greater dimension – 47% more than Samsung and nothing less than 156% more than AMD or TSMC, which despite the most important company in the world in the manufacture of semiconductors does not dedicate more money than its rivals to innovate. The thing becomes even more interesting if we compare these R&D investments with the income of these companies. In 2024 Intel invested 31% of its net incomewhile AMD invested 26%. Nvidia only invested 10% – but not to make money with their AI products – but here the truly striking is that Samsung does not seem so interested in its R&D department, because it only invested 4% of its net income. It is also important to note that in AMD they do not have their own production plants such as the rest of their competitors, so although their investment in R&D was the lowest of these four companies, all that investment went to the chip design in full. The problem for Intel is that its growth in investment in R&D grew only 3.1% that in 2023. Compared Samsung grew by 71.3% more, NVIDIA 47% more and TSMC 8.8% more. They all seem to press the step more than Intelwhich raises just the opposite. In fact it is known if in That cuts policy which is carrying out Lip-Bu so will also end up cuts in the division and cost of R&D, but It seems possible that it is so. If so, this 2025 NVIDIA may end up being the company that spends the most in R&D, especially to try not to lose its privileged position in the AI ​​market. Image | Intel In Xataka | Intel’s fall symbolizes the end of an era: the model that dominated technology for 50 years has died

Their companies lack the scale of their rivals

by

The Japan government needs its semiconductor industry to be great again. The biggest. In fact, it was in the past. In 1988 NEC, Toshiba, Hitachi, Fujitsu, Mitsubishi, Matsushita and other Japanese companies hoarded nothing less than 50% of the chips industry. However, Today none of these companies It is positioned among the leaders of A sector dominated with iron fist by Taiwanese, American, Dutch, South Korean and German companies. Japan is currently investing more money in its sector of integrated circuits than the US, Germany, France or the United Kingdom. Not in terms of net value, but its effort is greater if we weigh the investment of these countries on their gross domestic product (GDP). The US dedicates 0.21% of its GDP to its semiconductor industry, and Germany 0.41%. France, according to Nikkei Asia0.2%, and, finally, the United Kingdom 0.04%. The difference is very significant and puts on the table the effort that Japan is making with 0.71% of its GDP. However, this country will not be easy to compete from you to you with Taiwan or South Korea in the integrated circuit industry. Toshikazu Maeda, the general director of the company specialized in the manufacture of equipment to produce Marumae chips, holds that many Japanese companies lack the necessary scale to compete effectively and increase their income. In fact, he regrets that most of the Japanese companies are not growing in full rise of the artificial intelligence (AI). To remedy it, it proposes a solution: smaller companies should merge to grow and be ready to react to the next great opportunity. Rapidus is Japan’s best option to compete with South Korea and Taiwan Japan currently has dozens of very specialized small businesses that manufacture components for ASML either Tokyo Electronwhich are two of the largest manufacturers of photolithography and wafering processing equipment. As Maeda defendsits production capacity is too modest to compete with giants from other countries, such as South Korean companies Samsung or SK Hynix, which produce some of their integrated circuit manufacturing equipment, or the American applied materials, among many others. However, if we stick to the manufacture of Japan Chips already has a company that aspires to compete with TSMC, Intel or Samsung. Rapidus corporation It has been expressly created to replace Japan at the forefront of integrated circuits. Interestingly, it is a very young company. It was founded on August 10, 2022 By the Japanese government With an initial capital of 7,346 million yen (just under 46 million euros) contributed by, and here comes the interesting, Sony, Toyota, Nec, Softbank, Kioxia, Denso, Nippon Telegraph and Mufg Bank. The initial capital invested in the constitution of this company is not very bulky, but there is no doubt that the companies that participate in it have an indisputable relevance in the sectors of technology, automotive and telecommunications. Japan currently has dozens of very specialized small businesses that manufacture components for ASML or Tokyo Electron Rapidus is currently putting a circuit manufacturing plant integrated in northern Japan, in the city of Chitose (Hokkaido), in which it plans to produce semiconductors of 2 Nm. The first prototypes of these chips They are already readybut large -scale manufacturing will not arrive at best until 2027. So far there is nothing really surprising because presumably at that time TSMC, Samsung and Intel will already be manufacturing integrated circuits with comparable lithographs. What is causing the new Rapidus factory to monopolize the looks of the semiconductor sector is that, according to Atsuyoshi Koike, which is the president of the company, it will be completely automated. Its purpose is resort to robots and AI To set up an automated production line that will be specialized in the manufacture of 2 Nm chips for AI applications. Its plan consists, in short, to produce integrated circuits faster, with a lower and more quality cost. To manufacture these semiconductors, equipment of extreme ultraviolet lithography (UVE) produced by the Dutch company ASML, and practically all manufacturing processes are automatic. However, the tests of test and validation, interconnection and packaging of the chips are still largely carried out manually in most manufacturing plants. According to Rapidus, its automation technology of all these processes will allow you to reduce the delivery time of your chips by 66% compared to the times they usually offer TSMC and Samsung. If this Japanese company finally achieves its purpose and its competitors do not improve its efficiency will be able to deliver its semiconductors In a third of the time spent by their rivals. A priori is a stinging enough asset for Rapidus to grow in a perceptible way, although for the moment it is just a conjecture. Whatever this company seems to have everything well tied. More information | SCMP In Xataka | Japan takes the initiative with nuclear fusion and sets an extremely ambitious date: the 2030s In Xataka | Japan has taken the carrier to dominate the chips industry. Prepare a 325,000 million dollar plan

The most obsessively competitive gamers have found a trick to win their rivals: Give Electric Downloads

by

Today video game developers are really strict in terms of cheating in their creations, with different systems Anti -che that They even affect kernel From our computer as is the case in ‘Valorant‘. But a new generation of Modders It is demonstrating that traps can be done without touching a single line of game code: with electrical discharges. If I can’t hack the game, I’m looking for alternatives. For some people, it is good for it is the same as Apply different tricks to games that they use daily to be able to cross walls, be immortal or have an automatic point when talking about Shooters. But this is something that leaves the rest of the players of a game. Having the player himself. Two Youtubers They have put the solution on the table to skip the barriers that are applied at the hardware level. One of them is ‘Basically Homeless’ that has led the concept of ‘player improvement’ to a new level. And instead of installing software that points for it in Counter-Strike 2, has created a system that electrocutes your arm to react at a superhuman speed. The mechanism is fascinating. An external software analyzes the screen in search of enemies. As soon as one detects one, it sends a signal to a Raspberry Piwhich in turn activates muscle stimulation diodes placed strategically in its forearm and hand. These little cramps force their muscles to get contracting, moving the mouse towards the target and clicking. Surprisingly, it is something that works. The results are very good. The own Youtuber He managed to reduce his reaction time of about 200 milliseconds approximately at only 100 milliseconds. According to its calculations, using an Ethernet cable connection instead of Wi -Fi to communicate the PC with the Raspberry Pi, it could lower this figure to 40 ms, a practically unbeatable speed for a human being. It does not consider it a trap. For this creator this is not something that threatens ethics when playing against other people. It is only a help, or as he calls it, ‘neuromuscular aim assistance’, since it is technically his own body who performs the action, although induced by a machine. Although it remains to be seen what large companies would say if this is popularized. Robotic carpets that point for you. In a similar line, the Modder Kamal Carter has presented another solution of hardware to dominate in Valorant. In his case, the protagonist is not his arm, but a robotic platform located under the mouse mat. And the system is similar. A screen reader identifies the enemy bots in the game’s shooting field in the first place. Next, a program that emulates the techniques of aimed at professional players sends instructions to the platform. This moves with a millimeter accuracy, displacing the mouse to achieve perfect shots. After adding a system that automates the click, Carter achieved almost perfect scores in practical mode. The most advanced anti-cheat systems are weak. “Made the law, made the trap,” says the saying. And in this case it is fulfilled. Large companies no longer know what to do to avoid tricks in their games, coming to ‘invade’ the kernel of our computer with maximum Windows privileges. Something that is designed to detect the use of unauthorized software. And there are many consequences that are being presented by cheating. Valve managed to block 40,000 cheats In ‘Dota 2’ or in ‘Deadlock’ They transformed into frogs to the most cheats to become aware that they should not do that. But also in Warzone they bet on Block the opening of the parachutes with the aim of crashing directly with the ground. But now these new tricks can be more difficult to detect and apply a punishment. In Xataka | Nintendo Switch 2: 17 tricks and tips to squeeze the portable console to the maximum

Bill Gates and Linus Torvalds had been rivals for 30 years. The funny thing is that they have just known and a selfie has been made

by

Being fond of technology between the late 90s and the first years 2000 implied Be a witness of a fierce War of sides Among the supporters of Windows, Macos or being an alternative “outsider” that renegated both (and even Graphic environments) and hugged free software from the hand of GNU/Linux. As standards of each of those sides: Bill Gates, Steve Jobs and Linus Torvalds. Between Bill Gates and Steve Jobs there has always been a certain Love -od relationship What He led to collaborate and throw yourself The hit to the head recurringly. However, after more than three decades dedicating expletives, we have had to wait until 2025 so that Bill Gates and Linus Torvalds are face to face and appear together (and well avenues) in a photo. A historical encounter between two legends Bill Gates has defended throughout his career the right of developers to license your programs I already charge for them who uses them. For its part, on the opposite side, Linus Torvalds bet on a Open and collaborative modelin which each user can adapt their tools to their measure freely. Bill Gates and Microsoft. Linus Torvalds and Linux. Both represent the different ways of understanding the evolution of computer science in the last three decades and, despite being the main protagonists and flag bearers of the most fierce dialectical battles In the technology blogs of the 90 and 2000, they did not know each other in person. The person responsible for facilitating the historical meeting has been Mark Russinovich, creator of the popular SysInternals softwarewhich currently acts as Director of Technology (CTO) of Microsoft Azure. Russinovich, always has been very supportive of the Open Source solutions and it was one The drivers of Azure support for Linux. That position has assured him a Good relationship with Torvaldswho accepted the invitation of the director for dinner with Bill Gates and Dave Cutler, one of the main Windows NT developers and Azure promoter. It has transcended very little about what was discussed during the unique dinner, but what has transcended has been the historic selfie that Russinovich did to immortalize the meeting and that he has published in Your LinkedIn profile. It is undoubtedly an unusual image of two staunch enemies for more than three decades, which They just met And they pose with relaxed and affable attitude. Touch the photo to go to the original message Next to the photo, the host could not hide his emotion for having achieved such a milestone: “I had the emotion of my life, organizing dinners for Bill Gates, Linus Torvalds and David Cutler. Linus had never met Bill, and Dave had never met Linus. No important decisions were made about the kernel, but perhaps the next dinner.” Linux very touched and more integrative Microsoft The moment in which it occurs does not go unnoticed either. Linux has gained market share thanks to the proliferation of cloud infrastructure and is already installed in about 100 million equipment (mainly servers), achieving a 4.13% market share. In 2021 and 2022, the system had a 2% share, which has been growing since then. According to data Statcounter, the Windows fee remains in a loaf 70.21%. For its part, Microsoft left behind the times of Steve Ballmer describing “cancer” To Linux. Since the arrival of Satya Nadella to the Microsoft direction, they have been carried out different approach samples between Windows and Linux. One of the most important was the possibility of deploying a Windows subsystem for Linux (WSL) that allows you to execute an environment Linux in Windows computers. Perhaps, as Russinovich said in his post “important decisions were made about the kernel” but, taking into account the attendees of that dinner, they may talk about greater Linux integration On the Azure Platform of Microsoft. Maybe they decide at the next dinner. In Xataka | Bill Gates’ fortune records an unprecedented fact in the last 33 years: Millionaires’ Top 10 ‘has been left out Image | Mark Russinovich

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.