Meta reported better-than-expected financial results during Wednesday's 2025 fourth-quarter earnings call, but it was CEO Mark Zuckerberg's vision for Meta AI in 2026 that truly stood out. Zuckerberg said the company will be spending big to build personal superintelligence. It has one major edge over competitors -- troves of personal data about me, you and everyone we know.
"We're starting to see the promise of AI that understands our personal context, including our history, our interests, our content, and our relationships," said Zuckerberg. "A lot of what makes agents valuable is the unique context that they can see, and we believe that Meta will be able to provide a uniquely personal experience."
Meta's long-term AI goal is personal superintelligence, a kind of holy grail: an artificial intelligence that's smarter than humans, tailored to our individual experiences in products like smart glasses. To get there, the company expects capital expenditures to increase dramatically, from last year's $72 billion to $115 to $135 billion, attributing the increase to supporting its AI labs.
That money will be spent on research in a couple of different places. Agentic AI, which is tech that can handle tasks autonomously, is one big piece of the puzzle. The personalized component is where Meta believes it has an advantage over competitors.
The company has spent years collecting, analyzing and monetizing a wealth of information from its users on Facebook, Instagram and WhatsApp, effectively acting as a massive data broker. Its targeted advertising business model is structured around surveilling our online activity, and it's part of what made Meta the tech titan it is today.
Personal AI is what the industry sees as the next step: chatbots, agents and other products that are relatable and customized to our individual lives and needs. As a social media giant, Meta already knows plenty about us and what we may want to see from personal AI.
That will work in tandem with the company's plans to "merge" LLMs with the recommendation systems that build our social media feeds. "Soon, we'll be able to understand people's unique personal goals, and tailor feeds to show each person content that helps them improve their lives in the ways that they want," Zuckerberg said.
Even if Meta hadn't been collecting all our data for decades with its social media platforms, it could still have an edge. Meta AI is everywhere on Facebook and Instagram, and the company doesn't let you opt out of model training or turn it off. (You can mute Meta AI, though.)
YouTube and LinkedIn are almost certainly helping their parent companies, Google and Microsoft, too. Yet Meta has a first-in-class, proven track record of turning personal data into products and monetization. That's not necessarily a win for us. Meta's previous AI integrations in WhatsApp and its plans to use AI interactions for personalizing ads sparked backlash. The battle for data privacy amid the development of data-hungry AI models is an ongoing fight.
Meta's AI development in 2025 was studded with epic highs and lows. It made waves over the summer when it hired a series of top AI researchers, poaching some of them from other bigwigs like OpenAI and Apple. But reports of internal strife and conflicting strategies between the new hires and Meta's existing FAIR lab quickly followed, stalling any major public releases. Meta eventually laid off hundreds of employees from its AI units. Yann LeCun, one of the foremost pioneers in AI, left his role as chief AI scientist at the end of last year.
... continue reading