figure

The EU has just fined Google 890 million for the DMA. The figure weighs less than a paragraph almost hidden in the statement

by

The European Commission announced this Thursday two fines for a total of 890 million euros to Google for violating the Digital Markets Regulation (DMA). There are 460 million for self-preference in Google Search and 430 million for blocking Google Play developers who want to offer payment alternatives outside the store. It is the first firm sanction against Google under the DMA. Arrives less than a week after the 550 million to AliExpress under the Digital Services Lawthe sister standard that monitors the sale of products on platforms. In detail. The two breaches point to the core of Google’s business. In Searchplaces its own services (Shopping, Hotels, Flights, sports results…) at the top of the page, with rich visual formats, its own filters and graphic elements that rivals cannot replicate. External comparators, such as Idealo, Trivago or Skyscanner, appear below and in a simple list format. In Playdevelopers could not freely link to their own websites or alternative stores to complete the purchase. And when they did, Google continued to charge a commission on those external transactions for a period that the Commission considers “excessive.” The Commission gives Google 60 days to correct both practices. If you do not do so, you are exposed to periodic penalty payments of up to 5% of your daily worldwide turnover. Between the lines. The figure is impressive but is relative if we consider that Alphabet had a turnover of more than $350 billion in 2025 and the DMA allows sanctions of up to 10% of global turnover. That is to say, the 890 million are a lot of money but they are far from the legal ceiling: the penalty could have reached 35,000 million. The most important paragraph is almost hidden in the Commission’s statement: Google has submitted proposals on “how it plans to apply the decision’s principles to summaries and AI mode.” That is, the AI Overviews that already appear on the classic results also have to be subject to the rules of the DMA. There the technical complexity skyrockets because a LLM that synthesizes information from dozens of sources does not separate its own services from those of others with the same cleanliness. The context. Google has already accumulated almost 10 billion euros in European fines since 2017. The difference with the old sanctions is one of method. Those were antitrust: They came after years of investigation and punished already consolidated practices. The DMA operates in reverse. Establish ex ante rules for gatekeepers or designated “gatekeepers” (Alphabet has been since September 2023) and sanctions continued non-compliance. The Commission no longer disputes whether Google is dominant. Part of it is. Yes, but. Kent Walker, president of Global Affairs of Google, has reacted harshly. It has said that complying with the DMA will force the company to “remove real-time search features that Europeans appreciate” and “dismantle security protections on Google Play.” “It is not fair competition, it is a worsening of the product driven by a small group of complainants with particular interests,” he added. The Commission itself acknowledges, however, that Google has made considerable progress during the investigation. It has started testing changes to the presentation of Shopping, Hotels and Flights, and the regulator is examining them. The fine punishes already completed non-compliance. The real negotiation is about what’s next. And now what. Three open fronts: Google can appeal to the General Court of the EU. Given the track record, it likely will. These resources take years. This month, without going any further, the 2018 fine has been settled. Eight years. The technical adaptation in Search and Play will have to be verified. Rivals have been complaining for a decade that Google’s solutions are more cosmetic than anything else and that its dominant position remains intact after each round. The dialogue on AI Mode remains open. This is where how the search for the future is regulated is decided, and where the Commission has fewer precedents to rely on. The next round is not about ten blue links under a search bar. In Xataka | The worst news for Google is not that the EU forces it to open Android to ChatGPT: it is sharing its training data Featured image | Xataka

the humanoid robot Figure 03

by

In October 2025, Figure showed us what his new humanoid robot was capable of. The Figure 03 boasted a new design and much more precise dexterity, allowing it to move around the house and do such delicate tasks as picking up dishes or folding clothes. Although it is not for sale at the moment, Figure 03 has already started its first work. The chosen place has been the BMW factory in New Jersey. It is not the first time that BMW integrates humanoid robots in their factories. In fact, this new pilot program is the continuation of a first phase in which the previous model, the Figure 02, was supporting the production of the BMW X3 and previously They also had Figure 01 “in practice”. A very technological warehouse boy Figure 02 spent ten months at the BMW plant in Spartanburg, specifically in the body department. His job was to load the sheet metal panels that dress the BMW X3. Specifically, it helped the production of more than 30,000 vehicles. Now, Figure 03 has a totally different job that takes advantage of its new capabilities. The department where Figure 03 operates is assembly logistics. Here they receive the components in large, messy containers. Your job will be to collect them and sort them into carts that are sent to the assembly line in a specific order. In its first job, Figure’s robot had a very specific task in that it always picked up the same pieces, but now it must pick up pieces of different shapes, weights and sizes, which requires greater adaptability and precision. Figure 03 introduces important improvements compared to the previous generation, such as a design with soft parts that makes it safer, cameras in hands that improve the grip of objects and much more sensitive touch sensors. A key test This pilot is part of the initiative BMW iFactoryits global production strategy that seeks to promote digitalization starting with warehouse number 52 of the Spartanburg plant. The BMW X3 is produced here and the BMW iX5 electric. In this plant, technologies such as 3D simulations have already been implemented to optimize processes and vision and sound systems based on AI that provide real-time feedback to operators. Unlike the first pilot, this time they have chosen a more complex task in the sense that it undergoes changes and does not always follow the same pattern, which makes traditional industrial robots not the best option. If Figure 03 manages to maintain the precision and rhythm of the assembly line, it will be the litmus test to see whether humanoid robots can go from being a laboratory demonstration to a real worker. Images | bmw In Xataka | Humanoid robots will be truly ready when they manage to summit Everest. And they are already at it

Threads already boasts 500 million users. The missing figure remains the most important

by

Threads was born at the time when competing with X seemed more possible than ever. The old Twitter was going through a period of profound changes under Elon Musk and Meta decided to enter with its own app in a field that X had dominated for years: brief, immediate public conversation supported by text. What was not clear was whether that window could become sustained use, community and real scale. Almost three years after its launch, the company already has an answer to teach the market. The data comes from the Goal itself, which has announced that Threads It has reached 500 million monthly active users in June 2026. The application would have added about 100 million monthly users since August of last year, when it was already around 400 million. It is a huge figure for such a young network and enough to place it in a very different conversation than in its first months. The company led by Mark Zuckerberg has presented the milestone along with several new features, with special emphasis on reinforcing the role of communities within Threads. In its official announcement, Meta maintains that these groups, organized around conversations on topics such as books, basketball, parenting or musichave helped shape the application. That’s why Communities It is now out of beta and adds functions such as a center to find communities, own icons, progress indicators for topics that are close to becoming a community and more recognition for outstanding users. The figures we have and the figures we are missing Part of the explanation is that Threads didn’t have to convince the user to start from a blank page. It’s no secret that upon its initial launch it benefited from a highly optimized growth strategy: the app was able to build on the connections that millions of people already had on Instagram, and some viral Threads posts even appeared on Instagram and Facebook. This advantage helps to understand why its adoption was so rapid, although it does not solve the underlying question: how many of those users have turned Threads into a commonly used app. This pace places Threads in a striking position if we compare it with other large networks, although with an important caution: not all of them were born in the same conditions nor did they communicate their metrics in the same way. TikTok/Douyin reached 500 million monthly active users in July 2018a little less than two years later of the launch of Douyin in China. Instagram reached that barrier in June 2016some five years and eight months after its premiere. Facebook announced 500 million active users in July 2010a little more than six years after its birth, although that communication did not formulate the metric with the same detail as MAU. Threads did it just before turning three years old. There is the missing figure. Meta has given the global number, but has not published the breakdown by country: we do not know what Threads’ main market is, where the growth is concentrated or how many monthly active users it has in Spain. That gap matters because a social network is not only measured by its aggregate size, but by the weight it achieves in each local conversation. And in Spain, without a public figure that allows it to be measured, Threads does not yet seem to occupy a place comparable to that of X, Instagram or TikTok. Some clues help to read this incomplete map, although none replace the official breakdown that we do not have. In its announcement, the company led by Mark Zuckerberg mentions that local communities will start with native language labels in Japan, Korea and Taiwana clue as to where you are putting the focus. Meta claims to be seeing more traction in Asia, especially in South Korea and Japan, where usage time has increased by 80% and 130%, respectively, compared to the previous year. This is useful data, but it is not equivalent to knowing how many monthly active users there are in each country. Images | Goal In Xataka | “Deepfake” calls have become a top-level security problem: Google believes it has the answer

China has just ranked second in intelligent computing capacity. The important figure is not the most striking

by

There’s a simple way to hype up artificial intelligence: just talk about models. And there is a more useful way to understand it: look at which countries have the capacity to train them, run them and bring them to millions of users without the system breaking along the way. In that second race, much less showy but much more revealing, China has just presented its numbers. The figure draws attention due to the ranking, yes, but the important thing is what it tells about the foundations of its deployment. The figure. The information comes from Digital China Development Report 2025the document with which the National Data Administration summarizes China’s digital development over the past year. There it is maintained that the country reached 1.59 million PFLOPS in FP16 of intelligent computing capacity and that this volume would place it in second place in the world. There is small print. The aforementioned report places China in second place in the world within a specific category: intelligent computing capacity. That is not the same as saying that the country is second in the entire AI race, where models, chips, talent, investment, adoption, regulation and many other variables come into play. What we are looking at is something more limited: the computing capacity prepared to power large-scale artificial intelligence loads. Unity matters. FP16 stands for 16-bit floating point, a way of representing numbers with less precision than FP32 or FP64. It is widely used in artificial intelligence because it allows you to perform more operations and use less memory, a useful balance when we talk about training or running models. PFLOPS, for their part, serve to express how many floating point operations an infrastructure can perform every second. It’s not just power. The report does not stop at 1.59 million PFLOPS FP16. It adds a very specific physical layer: more than 13.73 million standard racks in operation, 42 large intelligent computing clusters, described in the document as “ten thousand card” clusters, and a national testing and verification platform that already supervises 1,129 facilities. This network, according to the document, allows 110,000 PFLOPS to be coordinated for economic, scientific and government uses. The distance with the US. China is placed second and the reference for first place is the United States. That is the reading that appears in Chinese state sources when they talk about a position “only behind the United States”, and also what external analyzes of high-end AI computing draw. The American advantage is not explained only by having more chips: also by data centers, large technology companies capable of financing enormous-scale infrastructures and a highly developed network. The other half of the data is in the use. According to the report, China had 748 registered generative AI services at the end of 2025, of which 446 had been registered during that year. It also talks about 602 million users of generative AI, with a year-on-year growth of 141.7%. These are official figures and should be treated as such, but they help to understand why computing capacity matters: we are not dealing with an infrastructure designed only for laboratories, but for services that are already deployed on a large scale. Images | Xataka with Nano Banana In Xataka | Unitree is doing with robots what DJI did with drones: becoming inevitable

what SpaceX needs to figure out before thinking about the Moon

by

Finally, after several postponements and even a scrub During the countdown, flight 12 of the SpaceX Starship has been able to take place. Elon Musk’s company has considered it a success, taking into account its complexity and everything that could have failed. However, it should be noted that it has been a partial success. The performance of the Starship S39 has been very good, but the Super Heavy B19 rocket has had some incidents. So many that they end up disintegrating upon re-entry into the atmosphere. Logically, this implies that there is a lot of work to do before the next flight. Which went well. The launch occurred successfully at 22:30 UTC (00:30 Spanish peninsular time) on Friday, May 22. It was achieved reach a thrust of 8,240 tons, double that achieved by the SLS rocket that NASA is using in the Artemis program. Even the acceleration was greater than expected. The separation of stages occurred properly and the ship fulfilled what was expected, landing in the Indian Ocean as planned. The release of Starlink mockups traveling as payload on Starship was also properly carried out. What went wrong. One of the biggest innovations of Starship Flight 12 was the introduction of version 3 of its Raptor engines. There were a lot of hopes for them, but some have not worked as well as could be expected. The first failure occurred 1 minute and 42 seconds after takeoff, when one of the outer ring engines of the Super Heavy rocket shut down. This consists of 3 central motors, an outer ring with 19 motors and an intermediate one with 11. The failure in the outer ring was already a relevant incident, but it was not the worst. The separation of the two stages occurred at 2 minutes and 30 seconds and precisely there it was seen how the ship’s 6 engines partially burned the surface of the rocket. At that point, the Super Heavy’s engines began to turn on, but some did not activate. 8 seconds later, one of the engines in the intermediate ring exploded, affecting several of the engines surrounding it. With the entire engine system damaged, only 5 of the intermediate ring engines were ignited during the return burn, so the rocket was unable to brake properly during reentry, which occurred at 1,450 kilometers per hour. The rocket disintegrated and what was left of it impacted the ocean 300 kilometers from the place planned by SpaceX. There were also failures on the ship. Although most of the failures occurred in the Super Heavy rocket, there was also an incident with the engines of the Starship ship. This consists of 3 vacuum engines in the center and 3 sea level engines around it. The difference between them is that those in the center are prepared to operate in conditions of spatial vacuum. Since there is no atmospheric pressure, They can have larger nozzleswhich allow greater thrust with the same fuel. On Starship Flight 12, one of those 3 engines shut down early, so to compensate, it was necessary to keep the sea level engines on for longer than planned. At least it was a mistake that SpaceX engineers were able to correct. Most of the failures were concentrated in the engines And now what? As SpaceX has pointed out, this has been a partial success. There have been many points of the mission that have gone perfectly, but it is clear that there is a lot of room for improvement. To begin with, some questions should be asked, such as whether the shielding system that the engines previously had had prevented the explosion that caused the rocket to disintegrate. In version 2, the external piping system It left the engine so exposed that each of them had individual shielding. In version 3 this shielding has been considered unnecessary when improving that space plumbing system. However, it is clear that it will be a point to review. On the other hand, it will be necessary to study step by step the on and off systems that have not worked properly. Next challenges. On upcoming flights, SpaceX will have to meet several challenges. The first will be to demonstrate the possibility of doing an orbital ignition. It was planned to fire one of the engines individually in orbit, as it is a key step for orbital insertion and controlled returns to Earth. Unfortunately, given the problems that were occurring with the engines, the plan was finally cancelled. On the other hand, it remains a challenge rocket recoverywithout disintegrating. And, finally, we will have to try to make these vehicles quickly reusable, as the Falcon 9 is now. In short, Starship flight 12 has been a success, but there is a lot of work ahead. SpaceX should not rest on its laurels if it wants to stay alive your lunar dream. Image | SpaceX In Xataka | SpaceX is preparing the largest IPO in history: the fact that it is doing so right now is no coincidence

Tesla’s enormous problem in Germany has an alarming figure and a clear person responsible: Elon Musk

by

Three out of four potential buyers of an electric car reject the idea of ​​buying a Tesla. The study points to the German market, which is the first electric car market in Europe by sales volume, and explains an important part of Tesla’s failure in Europe during 2025. Three out of four. 75% of potential buyers of an electric car in Germany do not value the idea of ​​buying a Tesla car, according to a study by the German Institute of Economics in collaboration with the Technical University of Dresden. The figure, which in itself is bad, has even more meaning. And that 75% is made up of potential customers who believe it is unlikely to buy a Tesla (15%) and those who completely reject buying a vehicle from this brand (60%). The reason, as we could imagine, is not a question of competition or price. The disaster. Last year, 545,142 electric cars were sold in Germany. It was, by far, the strongest electric car market in Europe. The growth was 43.2% compared to 2024, the year in which just over 380,000 electric cars were sold. Its market share reached 19.1%, above the European average, according to ACEA. For Tesla, however, it was not a great year. In Europe, 150,504 electric vehicles from Elon Musk’s company were sold, 37.9% less than the previous year when 242,436 registrations were registered. The most problematic thing is that the company had achieved a market share of 2.3% (a good bite to eat on the electric car pie, which in 2024 was only 13.6% in the European Union. That is, almost two out of every 10 electric cars sold in Europe were from Tesla. The drop was even more pronounced in Germany. There, the drop was 48.4%, as recorded Reuters at the beginning of the year. And, with everything, It has not been its strongest percentage drop in European countries but the damage in volume is more than evident. The politics. The decision by which the Germans seem to completely reject Tesla is evident to the creators of the study: Elon Musk’s political positioning. According to the authors, political positioning influences the purchase of a car more than sociodemographic characteristics. They point out that young people, those with a higher level of education and those who live in urban areas are more inclined to purchase an electric car. In political terms, Green supporters are the most open to acquiring this technology and AfD (German far-right) voters are the least enthusiastic. On average, they say, the potential customer for an electric car has grown by over 40% and those who reject it outright have also fallen. But the problem for Tesla is that it is not attractive to either group. Among the Greens, only 10.8% value the purchase of a Tesla as their first option and the percentage grows among AfD followers to 15.2% but it must be taken into account that these voters are also less in favor of buying a car of this type. Just lose. The study concludes with a statement: Elon Musk has lost support for buying cars among progressive groups (those who buy the most electric cars or are willing to buy) and has not attracted enough conservative groups to alleviate this disadvantage. The result is a direct consequence of a year 2025 that began with Elon Musk doing a Nazi salute during Donald Trump’s takeover of the United States and which continued with a explicit support of the company’s head for AfD and other far-right parties in Europe. It must be taken into account that this type of political positioning in Germany is much more delicate than in other countries. In Germany the Nazi salute is a crime punished with a fine in minor cases but which can be grounds for imprisonment in more serious cases. Study on preferences when buying an electric car in Germany segmented by political parties. Source: German Institute of Economics The worst option almost always. The image above shows the predisposition of Germans to the type of electric car they want to buy, segmented by their origin and the political parties that these potential customers vote for. According to this data, Tesla is the last option in four of the six political parties studied, even behind Chinese cars as the first option. The latter always surpass him except among CDU and SPD voters (although in both cases a greater percentage considers it possible to buy a Chinese car over a Tesla if we add the second level of predisposition). Tesla reaps the worst results among the Greens and Linke (The Left) and the absolute rejection is greater among the supporters of the latter political party. Chinese cars are, in all cases, the second option chosen when considering those who are willing to buy an electric car and those who value it as a possible purchase. The Germans are the ones who obtain the most support and the first option in all cases, with the greatest support among Green voters and with the AfD as the party with the greatest reluctance to buy it. Photo | Elon Musk in X and German Institute of Economics In Xataka | Tesla is discovering in real time that the most difficult thing was not to build a car brand from scratch: it was to maintain it

We believed that ‘Air’ and ‘Edge’ mobile phones were synonymous with cuts. Huawei wants to explode that idea with a figure: 6,500 mAh

by

The surname “Air” (or “Edge” in another case) is usually synonymous with an ultralight design and, therefore, of sacrifices. We have seen it in smartphones like the Galaxy S25 Edge from Samsung or your own iPhone Air of the signature of the bitten apple. The battery is the first victim in the quest for extreme thinness. However, Huawei seems willing to break this rule with its next Huawei Mate 70 Air after return to the top of the market in his native country. According to a wave of leaks and accompanying photos, the Chinese giant is preparing a device that not only claims to be the thinnest ‘Mate’ in history, but does so by integrating a huge battery. We knew that China had the solution for the battery of ultra-thin mobile phoneshere comes the first demo. A “normal” smartphone battery. This is the figure that is focusing all the attention of Huawei’s next launch. Leaks echoed by media such as Android Authority They point to a massive 6,500 mAh battery. If confirmed, in addition to being the highest capacity of a Huawei mobile to date, it would also dwarf the direct competition in the segment. slim: The iPhone Air has a 3,149 mAh battery, across the street, Samsung puts a 3,900 mAh battery in the Galaxy S25 Edge. An engineering challenge. How has Huawei managed to integrate this battery? Leaks indicate that the phone is built on an aluminum and glass chassis with a thickness of around six millimeters which would help. The images seen online confirm an extremely thin terminal that maintains the aesthetics of the Mate family, including its characteristic circular camera module. This is what the Huawei Mate 70 Air looks like in leaked images. Image: Weibo But without a doubt, the silicon-carbon batteries They are what have allowed the Chinese firm to take the leap. We have seen how these have allowed us to stretch the energy capacity up to 15,000 mAh in the case of Realme (still with certain unknowns about its durability) u 8,000 mAh in the Honor one. Without reaching these figures, the 6,500 mAh of the next Huawei Mate 70 Air seems feasible. It won’t skimp on photography either.. The Mate 70 Air looks at a triple system with a 50 megapixel main sensor (possibly 1/1.3 inch), a 13 MP ultra wide angle and an 8 MP periscope telephoto lens. It seems that it will not have to concede in the field of cameras, an ambition that aligns with Huawei’s strategy in recent times. one that has taken him to the throne of mobile photography recently with his Pure 80 Ultra. In addition, they may use again image sensors manufactured on national soil. Huawei is supported by SmartSensa Shanghai-based manufacturer of CMOS sensors: has more than 350 customers and 420 patents of which 190 are of its own invention. Reservations in physical stores of the Huawei Mate 70 Air. Image: Weibo Kirin Heart. And in two flavors? As expected in post-veto Huawei, the terminal will use an in-house Kirin 5G chip. Curiously, at Huawei Central They talk about two variants: the 12 GB RAM model would use a Kirin 9020B (a version with reduced clock frequencies), while the 16 GB model would use the Kirin 9020A, a SoC that we already knew in the Mate 70 family. It is, again, a reflection of the steps that Huawei has been taking in recent times in order to diversify some chips that no longer hidesas well as to ensure your HarmonyOS ecosystemkey in times when you need resilience. Imminent launch. This is not a long-term rumor: according to multiple leakers, the device is already in the reservation phase in physical stores in China and its official launch could be as soon as November 6. That is, in two days. All this happens while Huawei is already preparing new flagships: the Mate 80, which will try to demonstrate power by compensating for the hardware limitations (more evident in chip manufacturing) with custom software. Cover image | Composition with images of Huawei and Jose García for Xataka In Xataka | With HarmonyOS NEXT Huawei has achieved something incredible. Neither Samsung, Microsoft nor Mozilla achieved it

Openai signs with Samsung and SK Hynix for a potential chips demand of 900,000 wafers per month. It is an absurd figure

by

In Seoul A package of agreements was closed which reflects how far the career for artificial intelligence is coming. Openai sat down with Samsung and SK to advance his project Stargate And the companies pointed to a goal that surprises on its own: 900,000 DRAM wafers per month. The plan, according to the parties, goes through reinforcing memory production and studying new data centers in South Korea. All this was announced after a series of meetings of Sam Altman, business leaders and President Lee Jae-Myung himself. The appointment at the Seoul presidential office brought together Sam Altman With the leaders of the aforementioned Asian technological conglomerates, in the presence of the president Lee Jae-Myung. The tone was shared: Korea seeks to consolidate as one of the three global powers in artificial intelligence and OpenAi needs to anchor its Stargate project in regions with technological muscle. This lace explains the interest of both parties in formalizing agreements that cover from the memory supply to the construction of new data centers, with a long -term view. An objective that can tension the entire memory sector The volume that has been put on the table is disproportionate if compared to the market. According to Techinsightsthe global capacity of production of 300 millimeter DRAM was about 2.07 million per month in 2024 and would grow to 2.25 million in 2025. reaching 900,000 would mean about 39% of all that capacity. No individual manufacturer reaches such a figure alone, so that the magnitude of the agreement reflects both Openai’s ambition and the growing pressure to ensure the supply of advanced memory. Signed documents include preliminary commitments to expand memory production and evaluate additional infrastructure in South Korea. Among them is the participation of Samsung SDS in the development of data centers, as well as Samsung C&T and Samsung Heavy Industries in its design and construction. The Ministry of Science and ICT contemplates evaluating site outside the Metropolitan Area of ​​Seoul, and SK Telecom has signed an agreement to study the viability of a center in the southwest of the country. It is also proposed to explore the deployment of Chatgpt Enterprise and API capabilities in corporate operations. A key point in all this is in the difference between using and training a model. When someone consults a chatbot, infrastructure of inference is activated, much less demanding. But to train a new generation system, thousands of chips are needed working in parallel, each accompanied by High performance memory modules. This scale multiplies the need for servers, cooling systems and electrical power. In that context, guaranteeing hundreds of thousands of wafers per month does not seem an excess, but a way of ensuring that the next wave of models has the necessary material support. Stargate Data Center in the United States Openai’s computing muscle relies on huge draft alliances. With Oracle and SoftBankthe company prepares five data centers that would provide several capacity gigawatts. Nvidia, meanwhile, has announced that it would invest up to 100,000 million dollars and that would give access to more than 10 gigawatts through their training systems. Openai’s trajectory is not understood without Microsoft, his first great partner. The Initial bet of 1 billion in 2019 and the subsequent investment of 10,000 million gave access to the Azure cloud, Key to train models They promoted Chatgpt. Over time, however, Sam Altman’s company has begun to reduce that dependence. The last movements mark a change of course towards infrastructure in which OpenAI has more direct control, a way of making sure they are not conditioned to a single supplier. It should be remembered that many of the ads remain preliminary. Letters of intention and memoranda mark the will to advance, But concrete details have not yet closed. At the scale that Stargate raises, the risks are evident: from bottlenecks in the production of high performance memory to energy availability to feed facilities of several gigawatts. To this are added the necessary permits and the complexity of coordinating projects with so many actors. At the moment, the signed opens a path, but it remains to be seen what materializes and in what deadlines. Images | Sam Altman | Samsung | SK Hynix | Xataka with Grok In Xataka | I’ve been hooked to Sora 2 for two days: I’m generating absurd memes where I am the protagonist and I can’t stop

Openai estimates that it will enter 200,000 million dollars in 2030. The figure, like everything in OpenAi, is extremely ambitious

by

OpenAI has set a target of 200,000 million dollars for 2030, as reported The Information. Your own internal documents reveal that to achieve this you will need multiply by 13 your current income In less than five years. Why is it important. The company is burning billions per month and plans to spend 90,000 million only in R&D by 2030. This represents 45% of its projected income, well above the percentage allocated by large technological ones, which remain mostly between 15% and 30% of their gross benefit, not even their income. If Openai’s income is below the goal, that percentage will be even greater. The figures. Openai expects to move from 13,000 million income at 2024 to 200,000 million in 2030. Its R&D expenditure would be proportionally double that of the most successful technological technological ones, much more mature and settled. To achieve this, it basically depends on large companies continue to invest in generative. If there is A brake on investmenteven if that does not imply the burst of a bubble, OpenAi will have accounting problems. In addition, this projection rises up to only one semester. OpenAI has increased the expected billing by 2030 by the beginning of the year. The big question. Is a business model sustainable where almost half of the income – even the gross benefit – is destined for research and development? If business income does not rise as Openai projects, the company will have a serious problem. Yesterday it was announced Your agreement with Oracle committing to a huge investment level to which you can hardly face except that you change the screws, or to deliver a good part in kind (business use licenses), as Microsoft did with it paying in Azure credits. In Xataka | Baidu is no longer satisfied with being the Chinese Google. His new AI model also wants to turn it into China Openai Outstanding image | IlgmyzinXataka

The domestic robot with which many dream is already real. F.02 Figure has just learned to load a dishwasher

by

We have been imagining a future with robots that help us at home. From robotine of ‘The supersonic‘even Andrew from’The Bicentennial Man‘, Fantasy has always been there, but it had never seemed so real. The advances in recent years have taken us to a point where automatons begin to touch the field of everyday life. Figure AI, one of the most ambitious companies in humanoid robotics, has taken a step that brings that future. Your model Figure 02known as F.02, has shown that it can Fold clothes, reorganize packages and load a dishwasher: simple tasks for us, but greatly complex for a robot. A jump that redefines the potential of humanoids With 1.68 meters high and 70 kilos, F.02 is a fully electric robot with an autonomy of about five hours per load. You can walk 1.2 meters per second and transport objects of up to 20 kilos, which places it near the physical range of an average adult. What makes this robot special is that F.02 does not depend on human operators to execute their functions. Unlike Optimus, The Tesla robot we saw in the event ‘We, Robot’ with several assisted capabilitiesthis model works totally autonomously thanks to an artificial intelligence system called Vision Language Action (VLA), designed to interpret instructions and act on your own. The second key point is its ability to learn new skills from training data. F.02 could already fold clothes and move packages, but has incorporated the task of loading a dishwasher No need to redesign your hardware nor to create a new AI model. The company indicates that this task, which seems trivial, hides huge challenges: separating disorder of messy batteries, reorienting them and manipulating them with centimeter precision, coordinating both arms and handling fragile or slippery objects. Each dishwasher is different, each load is a puzzle, and the system must adapt in real time to unexpected errors or collisions. The most interesting thing is that everything has been achieved with the same robot and the same digital brain, only adding data. This draws a future in which domestic robots can receive updates for Expand their abilities without replacing them. They could even learn from their own experience at home, closer to a realistic vision of autonomous and useful robots on a day -to -day basis. Images | Figure ai In Xataka | Anthropic is worth 183,000 million even though he invoices 5,000 million a year. Or it is the business of the century, or it is the madness of the century

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.