The AI Race and Meta’s Struggle to Catch Up
Mark Zuckerberg once believed he had the upper hand in the AI race. Long before ChatGPT and AlphaGo captured global attention, and when OpenAI was still a fledgling startup and Google hadn’t yet acquired DeepMind, Facebook established its own AI research division called FAIR (Facebook AI Research). In 2013, Zuckerberg personally traveled to one of the world’s most prestigious AI conferences to announce FAIR and recruit top scientists for the lab.
FAIR has made several notable contributions to AI research, particularly in the field of computer vision. While the division wasn’t focused on directly enhancing Facebook’s social networking products, it was hoped that AI tools could eventually support the company’s core businesses, such as content moderation or image captioning. However, for years, Facebook didn’t develop AI as a standalone, consumer-facing product. Now, in the era of ChatGPT, the company is lagging behind its competitors.
Meta’s Lag and the Llama Model
Meta, now known as Meta Platforms Inc., trails not just OpenAI and Google but also newer companies like Anthropic, xAI, and DeepSeek, all of which have launched advanced generative-AI models and chatbots in recent years. In response, Meta quickly launched its own flagship model, Llama, but it has struggled relative to its competitors. In April, Meta proudly introduced Llama 4, which Zuckerberg described as a “beast.” However, after an experimental version of the model scored second in a widely used benchmarking test, the public version ranked only 32nd.
In the past year, every other top AI lab has released new “reasoning” models that are significantly better at advanced math and coding problems due to a new training paradigm. Meta, however, has yet to deliver its own.
A New Push for Superintelligence
A dozen years after building FAIR, Meta is effectively starting over. Last month, Zuckerberg initiated a new recruiting spree, hiring Alexandr Wang, the 28-year-old ex-head of the start-up Scale, as chief AI officer to lead a new division called Meta Superintelligence Labs (MSL). He has reportedly been personally asking top AI researchers to join. According to an internal memo, the goal is to build towards “personal superintelligence for everyone.”
Meta is reportedly trying to lure top researchers by offering up to $100 million in compensation. Although 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.
Challenges and Internal Doubts
Despite these efforts, Meta faces significant challenges. When reached out to about its “superintelligence” overhaul, a spokesperson directed questions to the most recent earnings call, where Zuckerberg described how AI is transforming everything Meta does. He emphasized his focus on building full general intelligence.
However, internal doubts persist. An outgoing AI researcher at Meta wrote in an internal memo, reported in The Information, that “you’ll be hard pressed to find someone that really believes in our AI mission.” The spokesperson responded by stating that Meta is excited about its recent changes, new hires, and continued work to create an ideal environment for revolutionary research.
Competitors and Different Approaches
Meta’s superintelligence group may well succeed. Small, well-funded teams have done so before. After a group of former OpenAI researchers formed Anthropic, they quickly became a top AI lab. Elon Musk’s xAI, despite being later to the race, has developed a technically impressive AI product in Grok, despite issues with racism and anti-Semitism.
Regardless of how far Meta has fallen behind, the company has proven 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.
Ideological Differences and Strategic Shifts
Meta’s approach differs from its rivals, who often describe generative AI in ideological, quasi-religious terms. Executives at OpenAI, Anthropic, and Google DeepMind frequently write long blog posts or give interviews about the future they hope to usher in, with philosophical disagreements among them. Zuckerberg, by contrast, doesn’t seem interested in using AI to transform the world. Instead, he focuses on areas like advertising, social-media content, online commerce, the Meta AI assistant, and devices, notably smart glasses.
The grandest future he described was tied to today’s digital services: “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.”
Open Source Strategy and Its Limitations
Initially, Meta seemed to take a different path. When the company entered the generative-AI race, it bet big on “open source” AI software, making its Llama model free for nearly anyone to access, modify, and use. It 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.
However, this strategy has not yielded the expected results. In January, the Chinese company DeepSeek released an AI model that was more capable than Llama despite having been developed with far fewer resources. Meta has internally discussed stopping work on its most powerful open-source model in favor of a closed model similar to those from OpenAI, Anthropic, and Google.
User Concerns and Platform Challenges
While Zuckerberg figures out the path forward, he must also contend with the basic reality that generative AI may alienate some users. The company rolled back an early experiment with AI characters after human users found that the bots could easily go off the rails. Meta’s stand-alone AI app also led many users to unwittingly share private conversations on the platform.
AI-generated media has overwhelmed Facebook and Instagram, turning these platforms into oceans of low-quality, meaningless content known as “AI slop.” Despite these challenges, 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 even more on new hardware, but it’s 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 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.
