Meta's Ambitious Leap: Zuckerberg's $14 Billion Bet on AI Superintelligence

Meta Swears This Time Is Different

The tech giant has floundered on AI. Now it's going all in on a "superintelligence" team. Mark Zuckerberg was supposed to win the AI race. Eons before ChatGPT and AlphaGo, when OpenAI did not exist and Google had not yet purchased DeepMind, there was FAIR: Facebook AI Research. In 2013, Facebook tapped one of the "godfathers" of AI, the legendary computer scientist Yann LeCun, to lead its new division. That year, Zuckerberg personally traveled to one of the world's most prestigious AI conferences to announce FAIR and recruit top scientists to the lab. FAIR has since made a number of significant contributions to AI research, including in the field of computer vision. Although the division was not focused on advancing Facebook's social-networking products per se, the premise seemed to be that new AI tools could eventually support the company's core businesses, perhaps by improving content moderation or image captioning. But for years, Facebook didn't develop AI as a stand-alone, consumer-facing product. Now, in the era of ChatGPT, the company lags behind. Facebook, now called Meta, trails not just OpenAI and Google but also newer firms such as Anthropic, xAI, and DeepSeek -- all of which have launched advanced generative-AI models and chatbots over the past few years. In response, Zuckerberg's company quickly launched its own flagship model, Llama, but it has struggled relative to its competitors. In April, Meta proudly rolled out a Llama 4 model that Zuckerberg called a "beast" -- but after an experimental version of the model scored second in the world on a widely used benchmarking test, the version released to the public ranked only 32nd. In the past year, every other top AI lab has released new "reasoning" models that, thanks to a new training paradigm, are generally much better than previous chatbots at advanced math and coding problems; Meta has yet to deliver its own. Read: Chatbots are cheating on their benchmark tests So, a dozen years after building FAIR, Meta is effectively starting over. Last month, Zuckerberg went on a new recruiting spree. He hired Alexandr Wang, the 28-year-old ex-head of the start-up Scale, as chief AI officer to lead yet another division -- dubbed Meta Superintelligence Labs, or MSL -- and has reportedly been personally asking top AI researchers to join. The goal of this redo, Zuckerberg wrote in an internal memo to employees, is "to build towards our vision: personal superintelligence for everyone." Meta is reportedly attempting to lure top researchers by offering upwards of $100 million in compensation. (The company has contested this reporting; for comparison, LeBron James was paid less than $50 million last year.) More than a dozen researchers from rival companies, mainly OpenAI, have joined Meta's new AI lab so far. Zuckerberg also announced that Meta plans to spend hundreds of billions of dollars to build new data centers to support its pursuit of superintelligence. FAIR will still exist but within the new superintelligence team, meaning Meta has both a chief AI "scientist" (LeCun) and a chief AI "officer" (Wang). At the same time, MSL is cloistered off from the rest of Meta in an office space near Zuckerberg himself, according to The New York Times. When I reached out to Meta to ask about its "superintelligence" overhaul, a spokesperson pointed me to Meta's most recent earnings call, in which Zuckerberg described "how AI is transforming everything we do" and said that he is "focused on building full general intelligence." I also asked about comments made by an outgoing AI researcher at Meta: "You'll be hard pressed to find someone that really believes in our AI mission," the researcher wrote in an internal memo, reported in The Information, adding that "to most, it's not even clear what our mission is." The spokesperson told me, in response to the memo, "We're excited about our recent changes, new hires in leadership and research, and continued work to create an ideal environment for revolutionary research." Meta's superintelligence group may well succeed. Small, well-funded teams have done so before: After a group of former OpenAI researchers peeled off to form Anthropic a few years ago, they quickly emerged as a top AI lab. Elon Musk's xAI was even later to the race, but its Grok chatbot is now one of the most technically impressive AI products around (egregious racism and anti-Semitism notwithstanding). And regardless of how far Meta has fallen behind in the AI race, the company has proved its ability to endure: Meta's stock reached an all-time high earlier this year, and it made more than $17 billion in profit from January through the end of March. Billions of people around the world use its social apps. The company's approach is also different from that of its rivals, which frequently describe generative AI in ideological, quasi-religious terms. Executives at OpenAI, Anthropic, and Google DeepMind are all prone to writing long blog posts or giving long interviews about the future they hope to usher in, and they harbor long-standing philosophical disagreements with one another. Zuckerberg, by comparison, does not appear interested in using AI to transform the world. In his most recent earnings call, he focused on five areas AI is influencing at Meta: advertising, social-media content, online commerce, the Meta AI assistant, and devices, notably smart glasses. The grandest future he described to investors was trapped in today's digital services and conventions: "We're all going to have an AI that we talk to throughout the day -- while we're browsing content on our phones, and eventually as we're going through our days with glasses -- and I think this will be one of the most important and valuable services that has ever been created." Zuckerberg also said that AI-based updates to content recommendations on Facebook, Instagram, and Threads have increased the amount of time that users spend on each platform. In this framework, superintelligence may just be a way to keep people hooked on Meta's legacy social-media apps and devices. Initially, it seemed that Meta would take a different path. When the company first entered the generative-AI race, a few months after the launch of ChatGPT, the firm bet big on "open source" AI software, making its Llama model free for nearly anyone to access, modify, and use. Meta touted this strategy as a way to turn its AI models into an industry standard that would enable widespread innovation and eventually improve Meta's AI offerings. Because open-source software is popular among developers, Zuckerberg claimed, this strategy would help attract top AI talent. Read: New Mark Zuckerberg dropped Whatever industry standards Zuckerberg was hoping to set, none have come to fruition. In January, the Chinese company DeepSeek released an AI model that was more capable than Llama despite having been developed with far fewer resources. Catching up to OpenAI may now require Meta to leave behind the company's original, bold, and legitimately distinguishing bet on "open" AI. According to the Times, Meta has internally discussed the possibility of stopping work on its most powerful open-source model ("Behemoth") in favor of a closed model akin to those from OpenAI, Anthropic, and Google. In his memo to employees, Zuckerberg said that Meta will continue developing Llama while also exploring "research on our next generation of models to get to the frontier in the next year or so." The Meta spokesperson pointed me to a 2024 interview in which Zuckerberg explicitly said that although the firm is generally "pro open source," he is not committed to releasing all future Meta models in this way. While Zuckerberg figures out the path forward, he will also have to contend with the basic reality that generative AI may alienate some of his users. The company rolled back an early experiment with AI characters after human users found that the bots could easily go off the rails (one such bot, a self-proclaimed "Black queer momma of 2" that talked about cooking fried chicken and celebrating Kwanzaa, tied itself in knots when a Washington Post columnist asked about its programming); the firm's stand-alone AI app released earlier this year also led many users to unwittingly share ostensibly private conversations to the entire platform. AI-generated media has overwhelmed Facebook and Instagram, turning these platforms into oceans of low-quality, meaningless content known as "AI slop." Still, with an estimated 3.4 billion daily users across its platforms, it may be impossible for Meta to fail. Zuckerberg might appear to be burning hundreds of millions of dollars on salaries and much more than that on new hardware, but it's all part of a playbook that has worked before. When Instagram and WhatsApp emerged as potential rivals, he bought them. When TikTok became dominant, Meta added a short-form-video feed to Instagram; when Elon Musk turned Twitter into a white-supremacist hub, Meta launched Threads as an alternative. Quality and innovation have not been the firm's central proposition for many, many years. Before the AI industry obsessed over scaling up its chatbots, scale was Meta's greatest and perhaps only strength: It dominated the market by spending anything to, well, dominate the market.

Alexandr Wang is now leading Meta's AI dream team. Will Mark Zuckerberg's big bet pay off?

In the summer of 2016, Alexandr Wang was a 19-year-old building his data-labeling startup, Scale AI, in a Silicon Valley pool house with his cofounder, Lucy Guo, while the two participated in the Y Combinator startup accelerator. When not working, the two founders slept on air mattresses and pondered the fledgling business's potential. Less than a decade later, the pool house project has reset expectations and plans across the tech industry's highest levels. In June, Mark Zuckerberg handed the now 28-year-old Wang the keys to Meta's entire AI operations as part of a $14.3 billion investment in Scale AI. As Meta's first-ever chief AI officer, Wang now leads a newly formed superintelligence team packed with AI industry superstars paid like high-priced athletes, and oversees Meta's other AI product and research teams -- all under the umbrella of a new organization called Meta Superintelligence Labs. This AI dream team, and Wang's role as its captain, mark the latest chapter in a Silicon Valley narrative that's practically a cliché by now: a new wunderkind rising in tandem with the next world-changing technology. In this case, however, rather than disrupting the old guard, Wang is leaping to action to help an established tech giant, which happens to be run by the previous generation's whiz kid. Zuckerberg, the Facebook founder who was worth more than $1 billion by his 23rd birthday, is going all in on AI to maintain the dominance of his 3.4-billion-user-strong social media empire. In addition to spending tens of billions of dollars a year on infrastructure to build the data centers running its AI, the Meta CEO is writing big checks to recruit the world's most sought-after AI talent. Along- side Wang, whom Zuckerberg has called "the most impressive founder of his generation," the superintelligence team includes former GitHub CEO Nat Friedman; Daniel Gross, who had been CEO and cofounder of buzzy startup Safe Superintelligence; as well as researchers poached from OpenAI, Anthropic, Google, and Apple, with compensation packages for some rumored to be north of $100 million. Ruoming Pang, the engineer in charge of Apple's foundation models team, is reportedly pocketing a whopping $200 million over four years. "I'm focused on building the most elite and talent-dense team in the industry," Zuckerberg wrote on his Threads app in July. The goal of this extraordinary hiring spree? Most pressingly, to reverse Meta's recent struggles in the grueling week-to-week contest for AI market share and mindshare, and to lock in Meta's large language models at the front of the pack. But Wang and his team of super friends have also been tasked with accomplishing something far more ambitious, and something that will be much trickier to measure with ordinary KPIs: to attain the still entirely theoretical concept of "superintelligence," leapfrogging ChatGPT-maker OpenAI -- and all other frontier AI competitors -- in the race to define the future of artificial intelligence. If the word "superintelligence" conjures images of omniscient science fiction machines, you're not too far off the mark. There is no agreed-upon formal definition of superintelligence, though the term typically refers to an artificial intelligence that vastly surpasses human capabilities in virtually all domains, including scientific creativity, general wisdom, and social skills. Superintelligence is generally perceived as going beyond artificial general intelligence, or AGI, which, though also vague, typically refers to an AI system with human-level intelligence across a wide range of work-related tasks. That is, it can reason, plan, solve problems, understand language, and learn in a generalizable way, much like a human. Estimates for achieving AGI vary widely, ranging from a few months to a decade or more; for superintelligence, the timeline ranges from a few years to never. Among Silicon Valley's leaders, the mission is spoken about with the assurance of an airplane pilot giving passengers a flight update en route to their destination. "Developing superintelligence is coming into sight," Zuckerberg wrote in an internal memo obtained by Fortune, though he provided no specific estimate for the date. (OpenAI CEO Sam Altman, for his part, wrote in a June blog post that humanity is now "close" to building digital superintelligence.) Zuckerberg's memo claimed that Meta is "uniquely positioned to deliver superintelligence to the world," pointing to the computing power in its data centers versus smaller labs with fewer resources. In his July Threads post, Zuckerberg touted Meta's investments in massive, multi-gigawatt data centers with grandiose names ("Prometheus" and "Hyperion") and mind-boggling scale: One of the data centers, Zuck boasted, will have a footprint nearly as big as Manhattan. Meta will "invest hundreds of billions of dollars into compute to build superintelligence," Zuckerberg declared. "We have the capital from our business to do this." Of course, a few other notable tech giants have the capital too, and they are dead set on edging Meta out. Microsoft and Google are both deploying tens of billions in capital expenditures to build out their AI infrastructure. And OpenAI, in addition to its partnership with Microsoft, has said it intends to invest $500 billion with partners including SoftBank to build out the Stargate network of AI data centers over the next few years. In making Wang Chief AI Officer, Meta has chosen someone who is an entrepreneur, not a computer scientist. Even more notable is the fact that the man leading Meta's quest for superintelligence comes from a company that was not in the business of actually building large language models. That's not to say that Wang -- a competitive "mathlete" in high school -- is a cerebral slouch. He's "11/10 smart and ambitious" and "superspecial," says Sarah Guo, founder of VC firm Conviction and former general partner at venture capital firm Greylock, when describing Wang. Guo (no relation to Scale AI cofounder Lucy Guo, who parted ways with Scale in 2018) got to know Scale's founders when they launched their startup from the pool house of her Portola Valley, Calif., home. Will Meta's teams of AI PhDs and computer scientists be as impressed and find in Wang the inspiration that drives them to reach the land of superintelligence? People close to Wang, including current and former Scale AI employees, investors, acquaintances, and competitors, have emphasized that Wang should not be underestimated in his ability to attract talent and lead Meta's AI organization into the future. "There's probably a handful of people in the world that you bet on" to build the kind of team Zuckerberg is looking for at Meta, and Wang is on that short list, says a former Scale AI manager who recently left the company. Wang "is a great recruiter, a really savvy commercial person." Conviction's Guo echoes the sentiment: "I think people would be mistaken to underestimate Alex and Nat," she says, referring to Wang and his Meta teammate Nat Friedman. "They are very, very smart, ambitious, technical people who are going to listen to their researchers. They are compelling recruiters with a lot of compute and have [Zuckerberg's] insane force of will behind them." An early Scale AI employee, who left in 2022, said that Wang em- bodied the startup's credo of "Ambition shapes reality." Perhaps that winds back to Wang's childhood as the son of immigrant parents who were nuclear physicists at Los Alamos National Laboratory in New Mexico. Wang, whose first name is spelled without the second "e" to give it the eight characters associated with good fortune in Chinese culture, dropped out of MIT after one year to pursue the startup dream. Wang and Guo had originally planned to create tech for a doctors' concierge service, but after joining Y Combinator they saw an opportunity in data -- specifically in the behind-the-scenes, labor- intensive work of tagging and organizing data so that AI models can learn from it. Scale's pivot to data labeling, often considered the "grunt work" that powers AI's intelligence, proved to be perfectly timed and highly valuable because of how critical data is for AI models. "Lucy's phone bricked from overheating because she'd set up the phone with Twilio to get a text notification every time a new request came in, and they flooded in so fast once researchers decided they wanted the data," recalls Sarah Guo. Meta's relationship with Scale AI dates back to 2019, when the social media company began using Scale as a data provider for its AI efforts, and Meta was among the investors in Scale's $1 billion funding round in 2024. Zuckerberg and Wang began spending more time together beginning in April, when Zuckerberg reached out and expressed a desire to work more closely, according to a source familiar with the negotiations. The Meta CEO began inviting Wang (who had become the world's youngest self-made billionaire before turning 25) to meet with him at his houses in Lake Tahoe and Palo Alto, with Zuckerberg soon coming to trust Wang's opinion. Advisors say that Zuckerberg would sometimes reference Wang's views in conversations with them, The Information reported. The conversations between the two CEOs came at a time when Zuckerberg was growing frustrated with Meta's struggles keeping up with rival AI labs such as OpenAI, Anthropic, and Google DeepMind. Meta had succeeded in creating a family of successful open-source AI models, called Llama, but never seemed able to stay ahead of the pack for long. OpenAI, Anthropic, and Google DeepMind would inevitably surge past Meta with each new update. With the release of Llama 4 in April 2025, Meta's malaise became a crisis. Allegations of possibly inflated performance metrics, a rushed release, and a lack of transparency, along with indications that Meta was failing to keep pace with open-source AI rivals like China's DeepSeek, led many in the industry to proclaim Meta's latest AI model a flop. (Meta has called claims that it gamed performance metrics "simply not true.") And so, even as Wang pushes for future superintelligence, he has the challenging near-term assignment of reinvigorating Meta's current LLM efforts. In the Llama large language models, Meta has a valuable asset and vast resources to draw upon. But as competition intensifies, particularly with upstart rivals from China, Wang will need to make some important strategic decisions, including whether it still makes sense to keep Llama open-source, or whether the changing competitive landscape requires that Meta keep the models' weights under lock and key. With Wang's firsthand experience building a successful business in the AI sector, the perspective he acquired as a neutral data provider serving the top LLM makers, and his business savvy, fans say he could be the missing ingredient to take Meta's AI efforts to the next level. Alex Ren, founding managing partner at Fellows Fund, says that Meta needed an entrepreneur to lead its AI organization -- and it found that in Wang, as well as Friedman and Gross. Meta's AI teams had previously been led by scientists and product managers. "That's wrong," says Ren, who counts several of the members of the new superintelligence team as friends. Pointing to OpenAI, and its business-oriented CEO, Sam Altman, Ren says that Meta "should really be led by founders and entrepreneurs. Alex Wang is not a researcher, he is a leader." With more than $164 billion in annual revenue, Meta knows all too well that it takes something truly unique to move the needle in its business. And in paying up for Wang and his team, Zuckerberg is betting that the ultimate value transcends any simple categories. If Wang strikes some as an unconventional choice, reckons one source close to the entrepreneur, it's because he does not fit into the typical tech-world archetypes: "Silicon Valley is good at putting people into boxes. They like to say, 'This person is a technical person, this person is a businessperson.' Alex is truly a man of one." Few companies could justify pouring billions into speculative AI research and assembling a roster of top-tier talent armed with the latest secret recipes for model success. But Meta can -- and has. Whether that bet pays off is still uncertain. Still, as one current Meta AI research scientist -- who isn't on the new superintelligence team -- told Fortune, if that group makes big leaps in frontier AI over the next six months, "everything can be justified."

Meta CEO Mark Zuckerberg Just Assembled a "Super Intelligence Avengers" Team That Could Totally Change the Game in Artificial Intelligence (AI). Here's Why That Makes Meta a "Must-Own" AI Stock. | The Motley Fool

Investors may be generally tracking the artificial intelligence wars (AI), with most of the "Magnificent Seven" companies spending hand over fist in a race to be the first to crack AI -- and all the financial benefits that come with it. But over the last couple of weeks, Meta Platforms (META 0.37%) CEO Mark Zuckerberg has made truly massive moves, committing huge amounts of dollars to both talent and computing infrastructure that dwarf even the current super-expensive standard of today's AI leaders. The implications of the moves may have been comprehended by some, but may still be underestimated by the larger investment community. Over the past month or so, Zuckerberg has: On infrastructure investments, Zuckerberg also shed light on massive upcoming projects: One might wonder what spurred this spending binge from Zuckerberg, and whether it was an offensive or defensive move. The answer, perhaps not surprisingly, is likely both. Zuckerberg now says Meta is aiming for "super intelligence," which could be somewhat akin to what was formerly referred to as artificial general intelligence (AGI). The concept of super intelligence, and whether AI is capable of reaching such a thing, has been hotly debated. However, it appears that Zuckerberg now believes super intelligence is achievable, and may be reached within the next few years. In a recent interview with tech magazine The Information, Zuckerberg said: There is this big debate in the industry today. All right, is super intelligence going to be possible in three years, five years, seven years? But I don't think anyone knows the answer. I just think that we should bet and act as if it's going to be ready in the next two to three years. Zuckerberg also believes "super intelligence" may mean different things to Meta than it does to more enterprise-oriented Mag Seven companies. Whereas, say, Microsoft might use AI to automate many enterprise functions, leading to an increase in productivity, for Meta, Zuckerberg apparently has a vision of giving consumers "super intelligence" related to their everyday lives, the media they consume, and their social connections. Zuckerberg also made an interesting note in the interview that the high salaries are worth it, since the ultimate team will likely be small, between 50 and 70 people: I think that the physics of this is, you don't need a massive team to do this. You actually kind of want the smallest group of people who can fit the whole thing in their head. So there's just an absolute premium for the best and most talented people. This makes sense. The architecting of AI systems is very complex, and if a technician makes a wrong architectural choice along the way, that can affect the performance of the entire model. According to AI chip blog Semianalysis, Meta's recent large language model Llama 4 has been a disappointment, and the reasons were partly due to poor data labeling -- which the Scale AI acquisition should help with -- and a few poor architectural choices. Thus, it's perhaps no surprise that Zuckerberg feels investing in a smaller number of high-caliber engineers is the best path. The difference between a winning model and a disappointing model may come down to a few high-level decisions, so it makes sense that Zuckerberg would pay up for quality over quantity for Meta's new AI efforts. Another offensive aspect of this is that Meta has arguably more financial resources than its rivals, especially OpenAI, which is considered a start-up and losing tens of billions at the moment. Last year, Meta's "core" social media advertising business brought in a whopping $87.1 billion in operating income, somewhat offset by a $17.7 billion loss in its Reality Labs division. And that $87 billion is probably on track to reach close to $100 billion this year. Therefore, Meta has the ability to pay as much or more than its rivals, and by paying these types of astronomical salaries, it's raising the costs of employment for everybody -- OpenAI included. Zuckerberg continued: ... one of the benefits of reinforcement learning is it gives you a venue to, you know, potentially convert very large amounts of capital into a better and better service, and potentially a better service than other less well-funded or less bold competitors will be able to do so... I view that as a competitive advantage. If we can get this to work well, and that's why we are basically all in on this. We're building, you know, we're building multiple, multi-gigawatt data centers, and we can basically do this all funded from the cash flow of the company. While the "all-in" spending binge from Zuckerberg is exciting, investors should also be wary of a few things. First, it appears Meta's AI super intelligence dream team will be essentially starting from scratch. This is likely due to Meta's recent efforts on its Llama 4 LLM coming up short of expectations, or at least falling further behind its other competitors than Zuckerberg would like. So, it appears Meta's latest attempt at leading AI is a bit of a bust, raising questions about the need to put all its chips into the pot, so to speak, at this moment. It has also been reported that Zuckerberg wasn't able to successfully acquire all the companies and talent that he wanted. In addition to Scale AI, Zuckerberg reportedly also wanted to acquire Mira Murati's Thinking Machines and Ilya Sustkever's SSI, but was rebuffed in both cases. It was also reported Zuckerberg extended billion-dollar offers to some of OpenAI's leadership team, but was also rebuffed. So, while Meta now has perhaps the most formidable AI "dream team" around, it isn't a "full" dream team necessarily. Finally, Meta has a history of throwing money at certain far-off ventures, without immediate tangible outcomes. Look no further than the Reality Labs segment, which is basically Zuckerberg's gambit to create the "next computing platform" of virtual reality goggles or glasses. Meta even changed its name from Facebook to Meta Platforms in 2021 to show its commitment to the effort. However, in 2024, three years later, that segment lost $17.7 billion, up from a $16.1 billion loss in 2023. Finally, Zuckerberg didn't really spell out what he exactly meant by an everyday consumer "super intelligence." While both the Reality Labs division and the concept of consumer super-intelligence may one day come to fruition, it's not assured -- even with Zuckerberg assembling an AI "dream team." So while this past month's spending is exciting, look for investors to get impatient if Meta's spending goes up without a corresponding growth in revenue. If one of today's current tech leaders reaches "super intelligence" before the others, it has the potential to disrupt the balance of power among today's Magnificent Seven. That's why any young person or growth investor should have exposure to Meta and its rivals, in spite of their massive AI spending today. If and when one of these companies "cracks the code" before others, it's possible the Magnificent Seven could become the Magnificent Three, Two... or even One. With his moves over the past month, Zuckerberg is investing heavily to make sure Meta is one of the leading candidates to become that "one." Investors should keep their ears out for more information when Meta reports earnings at the end of the month on July 30.