meta

Meta spent 2 billion on a Chinese AI startup. China is clear that it was a conspiracy

by

With China and the United States dancing the dance of artificial intelligenceboth countries and companies want to get the best cards for their decks. Meta is investing millions in the development of AI and, even so, it seems to be lagging behind. To turn the tables, he closed 2025 with a $2 billion purchase: that of a Chinese startup called Manus. The operation was so notorious that the Chinese government itself raised an eyebrow and undertook an investigation to see what was happening there. And they are already clear. It was a conspiracy. The Manus case. Although it has rained a lot and these last few months in China AI companies have come out from under the stones, during the first half of 2025 the proper name was that of deepseek. It was the great competition from the Western OpenAI or Google Gemini, but in March something that looked like an AI agent began to appear: Manus. That’s how they sold italthough it was really a deep investigation mode that helps you perform actions, but does not do them for you. It didn’t matter: the expectation was there and, although there were doubts about his behavior and limits, Manus began to move a lot of money (more than 100 million in estimated income) and attract attention from the big players. One of them was Meta, who took over the company. The purchase. A good question is how China let something like this slip away for a technological and strategic rival to buy. And it’s a good question, but the answer is that, at some point, Manus stopped being a Chinese startup. In the middle of last year, Manus moved to Singapore, allowing the company to bypass export and import controls imposed on China. To the not having your own LLMthey depended on others like Claude which they could more easily access from outside China. This already set off alarm bells in the Government, but with the purchase of Meta the bells echoed. China put to work to various organizations to see what was really happening, the largest of them being the Chinese National Security Commission, which is commanded by President Xi Jinping himself. The reports prepared by this body are directly supervised by the leaders of the Communist Party, so it is a voice that must be taken into account. Conspiracy. And the result of the investigation is clear. As they comment in Financial Timesthe conclusion is that Meta’s acquisition of Manus is a conspiratorial attempt to try to undermine China’s technological capabilities. These are big words that do not remain in a vacuum, since the founders of Manus – Xiao Hong and Ji Yichao – were summoned by the NDRC last March to address issues such as possible violations of foreign investment rules in China. He did not stay for a meeting and, as the FT points out, both have been prohibited from leaving the country during the review process. In fact, there are sources that suggest that Manus would be considering backing out of the agreement, but even so, it is not clear that the Chinese authorities will be satisfied. For his part, Meta points out that they did everything according to the law and it seems that he has already started to integrate Manus systems into their tools, so taking that step back would be very complex. And now… what. That the National Security Commission has classified the case as “conspiracy” is something serious, since it was the trigger for a broader review that involves more agencies in the country that are currently reviewing everything. And the underlying problem is the speed with which everything happened. Manus took off and, just four months later, they moved everything to Singapore to break away from China just before the purchase of an American company. The investigation is shaking the Chinese technology sector because it is not the first time something like this has happened. Although on a smaller scale, it is an operation called ‘Singapore washing’ in which startups founded by Chinese move to the city-state to bypass China’s control and have a more direct line with the United States. The problem is that, at a time when the commercial and strategic war has intensified, calling the Manus case a “conspiracy” sets a precedent. One in which it is stated that China does not want to let artificial intelligence talent and technology escape because this advance has become one of the country’s strategic legs for the next five years. We will see what happens when the case is resolved, but it is clear that Beijing’s objective, like Washington’s, is to prevent its assets from escaping, and Manus can be the example for national technology companies do not follow a similar model in the future. In Xataka | We don’t know if “crisis” means “opportunity” in China, but there is one business where it does: RAM memory

Meta will surpass Google in digital advertising for the first time in history

by

That Google is the queen of online advertising is one of the great constants of the Internet, but everything indicates that the reign is approaching its end. If the predictions come true, for the first time in history, Meta will be the company that generates the most advertising revenue. Projections. At the moment the surprise has not occurred, but the projections of the advertising analysis firm Emarketer are clear: Meta is going to snatch the throne of online advertising from Google in 2026. Specifically, they project that Meta will earn 243.46 billion dollars from advertising, while Google will earn 239.54 billion. Why is it important. Google’s dominance in the online advertising market was absolute. In fact, that domain has been in the regulators’ crosshairs for years and It has become very expensive for Google. Meta’s surprise, although not by a huge difference, is confirmation that the internet has been reconfigured with social networks and that the cake is much more distributed. Considering that Google has built its empire on the foundation of online advertising, it is even more relevant. The Meta Boost: AI. Meta has a portfolio of products with millions of users such as Instagram, Facebook, Threads and WhatsApp. According to Emarketer, the company has been “incredibly patient” in building solid usage habits in its user base before introducing ads. But what has caused this acceleration has been the integration of AI in content recommendation systems. This has allowed them to increase the viewing time of Reels by 30%, which translates into more advertising and therefore more income, specifically they are expected to reach the 50,000 million only with Reels. Goal Advantage+. It is the suite with AI that Meta offers advertisers. In addition to offering the platform to advertise, Meta also provides a ton of tools ranging from advertising actions to the creation of the ads themselves with generative audio, text and video AI. According to the brand’s results, revenue from video generation reached $10 billion in the last quarter of 2025. It was seen coming. It is not something that happened overnight, but rather The change has been in the works for years.. The displacement of searches was moving to other specialized platforms such as Amazon, Instagram or TikTok. With the emergence of AI, the landscape has become even more fragmented: with chatbots that provide answers to many user queries without us going through the classic search engine. Google is no longer the ‘default’ when doing a search, especially for younger generations who prefer audiovisual content. OpenAI enters the business. A few days ago we talked about OpenAI’s ambitious plans for its newly launched advertising business. The company hopes that, by 2030, they will have generated $100 billion with ads on ChatGPT, that’s nothing. It is still a much smaller amount than those managed by Meta or Google in a year, but it is enough for the impact to be noticeable. With social networks the exodus of searches began and perhaps we are facing the second great displacement. Time will tell. Image | Xataka, with Gemini In Xataka | The US has just opened a new wound in the Google empire: the justice system declares part of its advertising business illegal

The US has appointed executives of Meta, Palantir and OpenAI as lieutenant colonels. We have many questions

by

On June 13, 2025, four executives from some of the world’s largest technology companies donned the uniform of the United States Army at Myer-Henderson Barracks, a ten-minute drive from the Pentagon. After taking the oath, They were appointed lieutenant colonels of the Reserve. The appointment was controversial, but it was made on the occasion of the launch of Detachment 201, a very special army body dedicated exclusively to military innovation. Technological military with a wink. The four new reserve lieutenant colonels are Shyam Sankar, CTO of Palantir, Andrew Bosworth, CTO of Meta, Kevin Weil, CPO of OpenAI and Bob McGrew, advisor to Thinking Machines Labs (Mira Murati’s startup) and former head of research at OpenAI. The name of Detachment 201 is a wink to Silicon Valley, because an HTTP 201 status code on the web means that a resource was successfully created. All four will continue in their current positions while serving as reservists. Sankar’s thesis. Palantir’s CTO has already become a reference in the discourse on how to apply technology to military institutions after publishing on its website 18theses.com the document “Defense reform”. In it he talked about how “warriors fight with weapons and with git.” He criticized the Department of Defense (DoD) for treating technology as “expensive and unaffordable,” and proposed using AI to make military assets work more efficiently and quickly. The germ. The project was conceived by Brynt Oameter, who was responsible for talent management at the Pentagon. His idea was to attract technology experts so that they could take up positions in the Army when necessary. He met Sankar at a conference in early 2024 and began discussing the idea, which ended up crystallizing into a project that Donald Trump promoted. Finger designations. A curiosity: among that group of chosen ones there were no Anthropic executives even though the company was the one that ended up being the chosen one in July 2025 to integrate its AI model, Claude, into Pentagon systems. Then, how do we know, things changed. On Wired they explain how Sankar was the one who volunteered to be part of the project, but also recommended the three people who would end up forming that group with him. What will these four managers do in the Army?. The official mission of these experts is to integrate specialized knowledge in AI, software and data analysis into the Pentagon’s strategy. parameter gave an example: The commander of the Indo-Pacific region is evaluating threats in the Far East for the next ten years and has asked Detachment 201 to explain how AI can affect security in that context. These new lieutenant colonels can also operate more tactically, advising on how soldiers can use the new tools at their disposal. Or what is the same: they will act as consultants to the US Army, but in uniform and having taken the oath, something important because the relationship with the soldiers changes. The inevitable conflict of interest. The Army affirms that there is no conflict of interest because the members of Detachment 201 will not have a vote in the contracts signed with the private sector. The chronology of events tells us otherwise, however: A month before Bosworth took office, Meta announced an agreement with Anduril to develop military augmented reality products. A few months before OpenAI announced an alliance with Anduril in air defense systems. Palantir, Sankar’s company, signed a contract with the Army worth 480 million dollars in December 2024. That doesn’t prove anything, but suspicions are inevitable, because even if they don’t have a vote, they will be able to obtain internal knowledge and data that inevitably benefits their employer companies. But weren’t there going to be limits on AI in the army? Another of the thorny questions that arise from this Detachment 201 is how the recommendations of these experts will be applied on the battlefield. OpenAI theoretically has policies prohibiting its AI models from causing harm or developing military weaponry. However, the explicit mission of this body is to make the US Army be “more lethal”. That contradicts OpenAI statementswhich after allying itself with the Pentagon recently stressed again and again that its models would be used within limits… which is exactly why the Pentagon ended up wanting to turn Anthropic into a pariah company. Two weeks to get the rank. A conventional lieutenant colonel reaches that rank after between fifteen and twenty years of active military career. The members of Detachment 201 received that same rank after two weeks of partially online training that included physical conditioning, shooting as a diagnosis and basic notions of military protocol such as the rank structure and the use of the uniform. They did not complete basic training and have the flexibility to fulfill part of their 120 annual hours of service from home, something not offered to other reservists. All of this has generated reviews within the Army and also comments of all kinds on social networks. Image | DVIDS In Xataka | Anthropic and OpenAI have developed AI. The US Pentagon is showing you who really owns it

Within Meta there is a race to see which employee consumes the most AI tokens. It’s the ‘Tokenmaxxing’ of Silicon Valley

by

There is a battle within Meta: see who spends the most AI tokens. This is the basic unit that AI uses to understand the language with which we order actions. It is like the “bridge” between our words and the numbers that the machine can process and, therefore, when ChatGPT either Google They present a model, they brag about the millions of tokens they can process. But tokens are also becoming a ‘spending’ unit in AI companies. Silicon Valleyso much so that they may be generating a toxic work culture. And Meta is an example of a company where employees compete to see how many tokens they can consume to become a Token Legend. Tokenmaxxing. It is not the first time that we talked about this. A few days ago, Jensen Huang -CEO of NVIDIA and one of the main instigators of this phenomenon- commented that he would be worried if an engineer who earns $500,000 did not spend at least $250,000 a year on tokens. Because tokens cost money and NVIDIA is already considering offering tokens as part of the signing bonuses for its artificial intelligence engineers. Goals. As it could not be otherwise, Meta does not want to miss this party. The company, which changed its name when the metaverse was going to be the big thing and, after the swerveis defined as a “native AI company”, is one of those that promotes its artificial intelligence engineers to keep a count of the tokens spent during their day. There is no official data, but there are reports revealed to media such as Business Insider and The Information which point out that some of these teams have very specific objectives related to the use of tokens. For example, the company expects 65% of its engineers to write more than 75% of code using AI tools by the middle of this year. The Scalable Machine Learning division has another objective, and so on in each of the code-related departments within Meta. Legend Token. In The Information, they directly point out that there is an internal classification table created by the employees themselves to gamify the work. It shows the 250 most intensive AI users in their tasks with an easy premise: the more tokens you spend, the more you climb in the ranking. The winner of this particular competition takes the title of ‘Token Legend’, or ‘Legend of Tokens’. It is turning an expectation into a kind of internal sport. The first paragraph of this article converted to tokens crazy spending. If we put the first paragraph of 542 words in the tool ‘tokenizer‘ from OpenAI, we see that that simple phrase has already consumed 121 tokens. Well: according to The Information, in the last 30 days the total token panel usage of that internal table was more than 60 billion (of ours) tokens And even if they want to dress it for sports and competition, it is still obligatory. In late 2025, Meta launched the ‘Level Up’ program where employees who complete the most tasks using AI earn badges. And more important than this: it made the use of AI a central criterion in its employee performance evaluations. This, obviously, sets salary and promotion objectives. Doubts. But of course, beyond paying to work, there are other underlying issues. One of the criticisms of this tokenmaxxing system is that AI companies like Meta or NVIDIA encourage spending more on tokens because, in this way, their own employees become consumers of the product they are creating. An easy example that software engineering analyst Gergely Orosz exposed which is as if Tim Cook, CEO of Apple, said that if one of his employees who earns $500,000 a year did not spend $50,000 on purchases in the App Store, he would be worried. Orosz continuous stating that productivity should not be measured in tokens spent, but in the results obtained. Industry issue. In any case, Meta and NVIDIA are not the only ones that measure their employees by their consumption of AI at work. It is something that is soaking in other AI majors, turning the tokens into an extra work benefit incorporated into the engineers’ remuneration wheel along with the base salary, performance bonuses and shares. HE esteem that an OpenAI engineer can process 210 billion tokens in a week and there are Claude Code engineers who accumulate more than $150,000 in tokens in one month. Basically it is merging part of your salary into the company that pays you. And… have they said anything from Meta? Yes, it’s not about volume, but about quality, pointing that performance rewards are based on the impact of the work and not the raw use of AI. Image | ‘Wolf of Wall Street’, Meta Logo. Edited In Xataka | Google Earth shows the world. The Spanish Xoople wants AI to understand it

Meta has ended up firing its developers to pay for AI

by

Mark Zuckerberg’s company is not having its best week. To the sanctions imposed by a US court for not protecting users of the addictive consequences of their platformsjoins a new round of layoffs that affects hundreds of people in five business areas. It’s not the first time so far this year, and it probably won’t be the last either. We cannot say that the measure has caught Meta employees by surprise, because a few days ago Reuters I was already ahead that the parent company of Facebook, Instagram and WhatsApp was planning to cut staff due to the increased costs of AI development. Now have materialized eliminating the departments closest to the metaverse. 700 employees on the street and a metaverse that goes out. According to published NBC Based on sources close to the company, Meta will lay off about 700 employees in this round. The cuts will affect Reality Labs, the division that for years was the flagship of Zuckerberg’s big bet on the metaverse, which just a few days ago announced the Horizon Worlds closure on Quest headsetsas well as some in the human resources departments, sales and Facebook employees, as pointed out The New York Times. Those affected are a small fraction of the nearly 78,000 employees that Meta currently has on staff, but the reason given by the company is already a classic in big tech: “Meta’s teams restructure or implement changes periodically to guarantee that they are in the best position to achieve their objectives,” said a Meta spokesperson. in a statement to which you have had access NBC. Layoffs down, bonuses up. Hours before these layoffs were announced, Meta presented a new stock compensation program for six of its senior managers. The message between the lines has not gone unnoticed. While the company cut staff with the argument of reducing costs to face the huge investments in AIwith a forecast of expenses of between 162,000 and 169,000 million dollars for 2026, the executives closest to Zuckerberg saw their compensation increased by up to 921 million dollars each for the next five years. Meta justifies the increase to its managers as a tool to retain talent in the middle of the war for the best AI profiles, but the temporal coincidence between both announcements could not have been more unfortunate. ​Layoffs without financial hardship. Historically, a company laying off its employees was a clear sign of financial problems. Instead, in the age of AI, each round of layoffs is celebrated on the financial markets with increases in the price of shares because it is a clear sign that the company is restructuring to adapt to changes in strategy for the development of AI and continue generating million-dollar income. In fact, one of the phenomena that is occurring In the latest rounds of layoffs in large technology companies, while hundreds of employees are being laid off from certain departments, new vacancies are opening up. to hire new employees with another profile more AI oriented. ​Meta is not an isolated case. What happens in Meta is part of a dynamic that is repeated throughout the sector. Amazon, Microsoft and other big tech companies have announced massive cuts in recent months, and in all cases the AI appears as the main justification for layoffs. According to data From the consulting firm Challenger, Gray & Christmas, AI has been the argument for 12,304 layoffs so far in 2026, the equivalent of 8% of all layoffs recorded in that same period.​ In Xataka | Mark Zuckerberg spent millions on a “superintelligence” team. He is dedicating it to creating a personal AI agent for you Image | Goal

Meta hit it big, betting everything on the metaverse. Now they have a Schrödinger metaverse

by

We often see large companies change the design of their logos. They do it to maintain consistency with the product they are promoting at that moment, but the logo is one thing and the name and the entire brand are another. Facebook fearlessly jumped into the pool in October 2021 changing its name to Metaof ‘metaverse‘. After lose tens of billions and with the metaverse buried, Meta confirmed the inevitable: it will close the Horizon Worlds platform this year. But there is a twist: after announcing the closure, they now say that they will keep it alive for a while longer. How much? Mystery. In short. One of the most iconic moments of the technology presentations was when, in a packed room, Mark Zuckerberg walked between rows of journalists wearing a Quest helmet. The metaverse had arrived, or so Zuckerberg wanted. Years later, the reality is very different from future they hoped for Facebook Goalbut the name change had already been done and had to be accepted. ‘Horizon Worlds’ was the platform on which we could lead a second life, one that nor the employees of the Meta itself they used. To the metaverse he was doing badly, extremely badand Meta tried to make it stick in every possible way taking it to mobile and integrating it with Instagram, Facebook and WhatsApp. Meta announced the plan as one to bring the metaverse everywherebut a few hours ago, Meta took the final step: he announced that on June 15 he would close ‘Horizon Worlds’ in the Quest helmets. Lowering the blind. Although the metaverse is everywhere, it is evident that the most natural way to access it is through virtual reality. However, it is clear that they are not going the way they would have liked and, in a release On Discord, the company confirmed that virtual worlds could no longer be created, published, or accessed in VR after that date. It was announced that the closure would be carried out in stages, killing applications from the Quest store starting on March 31 so that no more users can join, culminating with the definitive closure of the VR worlds on June 15. In the announcement, Meta confirmed that ‘Horizon Worlds’ will, from now on, be a mobile-only app. It seemed like the culmination of a process that, apart from burning money, has led Meta to lay off hundreds of workers and hit a 30% blow to the budget of the Reality Labs division. Schrödinger’s metaverse. But it is clear that those who are still in the metaverse did not like the news at all, to the point that Meta has had to come out to clarify the message… and back away. Although they have done it in a curious way. In a question session on Instagram, the company’s CTO, Andrew Bosworth, came to the fore to comment that the dead man is very much alive, that they have thought better of it and that ‘Horizon Worlds’ will remain in VR for “the near future.” The statements are as follows: “The ‘Horizon Unity Runtime’ games do not work on mobile, only in VR, and we will not be introducing new games. Again, most of our strategy is aimed at mobile, but people who already have games they like will be able to download the ‘Horizon Worlds’ app and use it in VR in the near future.” In the units of measurement, “near future” is not that it is very concrete, but at least it seems that they do not kill it yet due to, according to Bosworth, showing support “to the fans who have contacted us.” He has commented on it in stories (something that, conveniently, will be deleted), but here is the video: Other type of glasses. Come on, where I said I say, I say Diego and the Quest metaverse will continue to live for a while longer. As a user, I wouldn’t throw my hat in the ring because it’s clear that they are planning a closure sooner rather than later, but it is always good news that they have listened to those who continue using the platform. Because the Quest, beyond ‘Horizon World’, is an extremely interesting headset for playing and consuming content, but Meta has been betting on another type of glasses for some time. There are the Ray-Ban Meta, “normal” glasses that are used not to consume, but to create. Already in 2024 we said that Meta was giving a flip from your VR glasses to your everyday glasses because the Ray-Ban Meta is not only a tool for content creators and anyone who wants to record their daily lives: it is a device through which Meta can distribute its AI. And an extremely controversial one, based on what we now know about Where can the images we record end up? with those glasses. And AI, of course. Because if years ago it was the metaverse, now Meta’s obsession is AI. The company is focusing on this technology in which it is not very well positioned. They focused a lot on preparing and presenting very good models, but not very consumer friendly and it was the big loser of the AI ​​race last year. Their change in strategy seeks to gain a foothold in a segment in which they are already ChatGPT, Grok, Claude or the chineseand for this it has a double strategy. On the one handa super team of AI stars whose machine will have to start working at some point. On the other hand, a collaboration with NVIDIA and another with AMD to train the AI, as well as the development of own chips for inference. There is 135 billion dollars at stakean investment in one year that exceeds the total of the Metaverse and that indicates why it is logical for Meta to abandon anything that does not work for him in the slightest right now in order to allocate all possible resources to pursue the new objective. Images | Goal … Read more

Meta spent a fortune on AI talent and data centers. Nine months later the result is: zero models

by

Mark Zuckerberg wanted to be the Florentino Pérez of AI. last summer began to sign galacticos in this segment and getting talent by letting go stacks of millions of dollars. He more popularOf course, it was the AI wunderkind Alexandr Wangwho became leader of its “Superintelligence” division. The funny thing is that the months go by and go by and in Meta they don’t seem to have absolutely anything to show. And that is very worrying. Delays. Despite having invested billions of dollars in that restructuring of the company to bet (practically) everything on AI, three internal sources confirm that Meta finds it very difficult to meet the planned deadlines. The race for generative AI waits for no one, and at the company headquarters nerves are on edge because the roadmap is not being met. Avocado, where are you? The new foundational AI model that Meta has been working on for months has been internally named Avocado, but at the moment it is not measuring up, something that reminds us what happened to Llama 4. Internal tests reveal that although it manages to surpass the aforementioned Llama 4 and the old Gemini 2.5, it falls short of Gemini 3.0 (and of course, the recent Gemini 3.1). Patience. Coming out with a model that is clearly worse than its rivals does not make sense, so Meta has decided to wait and delay the launch of its model. Avocado is expected to hit the market in May at the earliest. And meanwhile, Gemini. The situation is so critical that according to these sources, the leaders of the AI ​​division are considering something unthinkable: paying a license to Google to be able to use Gemini in their own products, something that for example will Apple do Siri. That would be a clear sign that for now this own model is not capable enough to power the AI ​​functions of WhatsApp, Instagram and Threads. Money does not equal speed. The company has spent billions of dollars on AI researchers, and has committed to invest 600,000 million dollars in building AI data centers. In January, Meta projected a capex of $135 billion dedicated almost entirely to these projectsalmost double the $72 billion it spent last year. Despite these investments, the company is currently missing from an area in which its competitors continue to advance. Internal tension. According to these sources, Meta is becoming a tinderbox. The “TBD Lab” (for “To Be Determined”), the unit led by Wang, is working under maximum pressure on models named after fruits (Avocado, Mango, Watermelon), but has clashed with old-school Meta managers like Chris Cox and Andrew Bossworth. The company is trying to integrate those models with Meta’s advertising business, which is what supports everything, but Wang doesn’t seem to handle that part of the business very well. Goodbye to open models. Meta stood out at the beginning of this AI race as the company whose open models —not Open Source— were above the rest. Llama became the norm in this area, but in this new stage that philosophy seems to change and China is the one that now leads that segment. Thus, there is talk that both Zuckerberg and Wang lean toward closed models, such as those of OpenAI (GPT) or Google (Gemini). This allows you to have full control over the code, a competitive advantage that Meta does not seem to want to give up. Few fruits of this tree. Despite the extraordinary deployment of resources, the current balance is poor. Meta’s only tangible product of those investments is Vibes, an application similar to Sora that has not managed to fully gel. Meanwhile, those initial talent signings have turned into abandonments: the trickle of AI researchers who leave the company to join others (or found their own projects) is increasing. In Xataka | Meta has been buying chips from NVIDIA and AMD for years. Now it also makes its own so as not to fall short

Meta has been buying chips from NVIDIA and AMD for years. Now it also makes its own so as not to fall short

by

Meta has not thrown in the towel with its MTIA (Meta Training and Inference Accelerators) chips. And although they didn’t have it all on their sidestopping depending on NVIDIA is a very juicy candy to jump to conclusions. For that very reason, They have presented a roadmap of four new chips with which the company intends to accelerate both its content recommendation systems and its generative AI capabilities. The first chip is now operational; The other three will arrive before the end of 2027. Below are all the details. Dependence. For years, Meta has relied almost entirely on NVIDIA and AMD to power its data centers. The development of our own silicon is complicated, but if it is achieved, it can be a very successful financial and strategic bet in these times. According to statements According to its vice president of engineering, Yee Jiun Song, designing its own chips allows the company to “eliminate what we don’t need,” which directly translates into cost reduction. Added to this is greater independence from possible price variations or supply restrictions. Which is exactly what you have announced. The four new chips are the MTIA 300, 400, 450 and 500. Each one has a different use: The MTIA 300 is already in production and is intended to train the algorithms that decide what content Facebook and Instagram users see. The MTIA 400 (known internally as Iris) has completed laboratory testing and is en route to data centers. Meta claims that it offers performance “competitive with leading commercial products,” according to its official statement. The MTIA 450 (Arke) will double the high-bandwidth memory compared to the 400 and is scheduled for early 2027. The MTIA 500 (Astrid), the most advanced, will arrive in mid-2027 and will incorporate, according to the company, improvements in low-precision data processing. The chips are manufactured by TSMC, the world’s largest semiconductor producer, and have been developed in collaboration with Broadcom on the RISC-V open architecture. The rhythm is the most striking thing. What’s unusual is not just that Meta makes its own chips, but the speed at which it plans to do so. The usual cycle in the industry is one or two years between generations. Meta aims to release new versions every six months. “The pace of AI evolution is so fast that we always want to have the most advanced chip available when we need it,” counted Song. This accelerated cadence is possible, according to the company, thanks to a modular design that allows components to be reused between generations. ANDthis does not replace NVIDIA. It is important not to lose sight of the context. Meta remains one of the largest buyers of GPUs on the market. just a few weeks ago signed multi-million dollar agreements with NVIDIA and AMD to supply chips for the next few years, and has also reached an agreement to rent computing capacity on Google chips, as share Wired. MTIA chips are designed for specific and internal tasks (inference and recommendation systems), not for training large language models, so this strategy is complementary to your chip plans with NVIDIA or AMD. Nor should we forget that Meta recently had to abandon its most ambitious training chip, known internally as Olympus, after the project became complicated in the design phase, according to counted The Information. Susan Li, CFO of Meta, confirmed at a Morgan Stanley event that the company still has the goal of developing processors capable of training models, but without giving more details. And now what. The real test of this bet will come when the chips are deployed at scale. The challenge at the moment is to guarantee HBM memory supply before a RAM crisis that is affecting the entire technology sector. Song himself recognized to CNBC that the company “is absolutely concerned” about it, although it stated that they have assured supply for their current plans. In the long term, we will see if Meta can achieve something similar to what Google did with its TPUs. Cover image | Mariia Shalabaieva and Goal In Xataka | OpenClaw has caused a real media earthquake in China. The Government has prevented its officials from using it

Meta wants to warn us before it’s too late

by

Social networks and messaging applications have become one of the favorite places for scammers to look for victims. It’s not a coincidence. They are platforms that we trust, where we interact with known people and where many decisions are made quickly, almost without thinking. In that scenario, a seemingly normal message, link, or request may be the first step in a scam. Goal has decided reinforce their alert systems with new tools that seek to detect suspicious signals and warn the user before it is too late. When someone tries to link your WhatsApp without you knowing. One of the novelties announced by Meta affects a fraud technique with which scammers try to link the victim’s WhatsApp account to another device. According to the company, they can try to convince the user to share their phone number and then the pairing code that appears on WhatsApp, or even ask them to scan a QR code under a false pretext. With the new alerts, the application will display warnings when it detects signs of suspicious behavior in these types of requests and will indicate where the request comes from so that the user can stop and reconsider the action before accepting it. Suspicious friend requests. Another of the functions that Meta is testing focuses on Facebook and such an everyday gesture as accepting a friend request. The platform tests warnings when it detects signs that may point to suspicious activity on an account. These signs include, for example, profiles with very few friends in common with the user or accounts that indicate a location in another country in their profile. The goal of these alerts is not to decide for the user, but rather to provide them with additional context before accepting the request, blocking it, or simply ignoring it. Messenger also receives news. Meta will expand its advanced scam detection system within this application to more countries this month. In this case, the focus is on the content of the conversation itself: when a chat with a new contact presents patterns associated with common fraud, the platform can show a warning to the user. The company mentions as an example messages that include suspicious job offers. At that point, the system can ask the user if they want to share recent messages for an AI-based tool to analyze and, if it detects signs of deception, offer additional information and suggest actions such as blocking or reporting the account. Beyond the visible notices. The firm led by Mark Zuckerberg explains that these in-app alerts are only part of its strategy against scams. In parallel, the company is strengthening its automatic detection systems with tools based on artificial intelligence capable of analyzing multiple signals at the same time, such as the text of a publication, the images used or the context in which the content appears. These systems seek to identify more complex patterns, such as accounts that impersonate celebrities, public figures or well-known brands, as well as links that redirect to pages designed to imitate legitimate sites. The figures. According to the company, during 2025 it removed more than 159 million scam-related ads worldwide for violating its policies. Of that total, around 92% were removed before any users reported them, suggesting that automatic detection systems were already acting before the fraud spread. Meta also points out that it took down 10.9 million accounts on Facebook and Instagram linked to scam centers and participated in an international operation that allowed the deactivation of more than 150,000 accounts associated with criminal networks in Southeast Asia. Images | Aman Pal | Goal In Xataka | China has turned OpenClaw into a viral phenomenon. And then it has prohibited its officials from using it

Meta just launched managed accounts for tweens

by

WhatsApp is part of the daily lives of millions of people and, in many homes, also part of family communication. The company itself has been presenting it for some time as a common tool to talk to parents, notify that someone has arrived home or coordinate day-to-day activities. However, The platform establishes that its use is intended for people over 13 years of age.. Now Meta has decided introduce a new modality designed precisely for that terrain. The novelty. What was announced by Meta consists of introducing a new type of account within WhatsApp designed for preteens. Instead of creating a conventional profile, the minor uses an account managed by a parent or guardian that is linked to that of the adult from the moment of configuration. This allows the person responsible to monitor certain aspects of app usage, such as who can send messages, which group invitations can be accepted, or what privacy settings apply to the account. A more limited experience from the beginning. The managed account does not replicate all the usual WhatsApp functions, but rather reduces the service to the essentials. In this format, the preteen can use the application to send messages or make calls, but some of the tools that the platform has incorporated in recent years are excluded. Among them are channels, the possibility of sharing location or integration with Meta AI. The adult is in control. As we say, these managed accounts not only limit functions, they also change who makes certain decisions within the application. Once the minor’s account is linked to that of the father, mother or guardian, that person begins to manage various aspects of the use of WhatsApp. You can decide which contacts are authorized to communicate with the account, which group invitations can be accepted, and review message requests from unknown numbers. Additionally, privacy settings are protected by a parental PIN, meaning only the responsible adult can access and modify them. privacy. Although managed accounts introduce new controls for adults, WhatsApp ensures that the platform’s privacy system remains intact. Messages and calls remain protected by end-to-end encryption, so only people participating in the conversation can access their content. Step by step activation. To launch one of these managed accounts, the process begins on the child’s phone and also requires the parent or guardian’s device. WhatsApp also indicates that both devices must have the most recent version of the application and that the person managing the account must be over 18 years of age. The adult must download WhatsApp to the preteen’s phone and choose the option to create a managed account during the setup process. Download WhatsApp on the minor’s mobile Choose the option to create an account managed by a parent or guardian Register and verify the minor’s phone number Enter date of birth and confirm age Scan the QR code with the adult’s mobile phone to link the accounts Verify that the adult is of legal age Create a six-digit parental PIN to protect settings Finish setup on the child’s device The Spanish context adds another layer. In Spain, the debate about minors’ access to certain digital platforms has been ongoing for some time. At the beginning of 2026, the President of the Government, Pedro Sánchez, announced the intention to ban access to social networks for minors under 16 years of age as part of a future regulation aimed at reinforcing digital protection at those ages. In this framework, platforms such as TikTok, Instagram or YouTube appear in the debate, while WhatsApp would be left out as it is considered a messaging service and not a social network. The new function seems designed to respond to a familiar use of messaging that the company itself assumes exists. Instead of ignoring it, Meta proposes a model in which this access occurs with more limits and with the direct supervision of a responsible adult. The result is a more limited version of WhatsApp, focused on basic communication and with additional controls over contacts, groups and privacy settings. In this way, the company tries to fit the use of the application by preteens within a more controlled environment. AvailabilityWhatsApp has only confirmed that these accounts will begin to roll out gradually in the coming months. That calendar leaves open an important question in regions like the European Union. In the European Region, On April 11, 2024, the company lowered the minimum age of use from 16 to 13 years to harmonize it with the rest of the world. However, the sources consulted do not yet detail how this new modality administered for minors below that threshold will be articulated in Europe or what scope it will actually have in those markets. Images | WhatsApp In Xataka | You’ve been ‘user84721’ for years. A study just showed that AI can know who you are in minutes

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.