Mark Zuckerberg’s AI Bet, Paid for by Three Billion Users

Mark Zuckerberg

Photo: Jeff Sainlar; Social Producer and Editor, Meta / CC BY-SA 4.0, via Wikimedia Commons

Born	May 14, 1984 White Plains
Occupation	Programmer
Awards	Axel Springer Award u00b7 Big Brother Awards u00b7 California Hall of Fame

There is a question embedded in every press release Meta has issued over the past two years: what does it mean to give something away when the thing you are giving away runs on infrastructure you control, trains on data you collected, and advances a strategy that keeps you at the center of the internet’s next phase? Zuckerberg’s answer — delivered in product launches, podcast interviews, and congressional testimonies — is consistently the same: open models democratize AI, and Meta’s interests align with the public’s. The consistency of the answer is what makes it worth examining.

He grew up in a dentist’s house in Dobbs Ferry, New York, and arrived at Harvard with a reputation for software that got him into trouble before it made him famous. The early projects — a program that linked home and office computers when he was eleven, a music recommendation tool in high school that Microsoft and AOL wanted to license — established a pattern: building things that rearranged how information moved between people. Facemash, the photo-ranking site that earned him a Harvard disciplinary hearing in 2003, had the same logic. So did the social network he launched from his dormitory room the following February.

He dropped out before completing his second year, moved to Palo Alto, and spent the following decade transforming what had started as a way for Harvard students to connect into something that eventually ran on the attention of three billion people. The company’s early “move fast and break things” ethos was not a metaphor — it described a genuine tolerance for collateral damage in pursuit of growth metrics.

The platform era began in 2007, when Facebook opened to third-party developers and converted itself from a walled community into infrastructure for the early social internet. The mobile pivot came later, forced by a near-collapse of market confidence around the 2012 IPO: Facebook had built for the desktop and nearly missed the device that would dominate the decade. The company rebuilt itself around mobile advertising and acquired Instagram for a billion dollars in April 2012, before it could develop into a serious rival. The $19 billion purchase of WhatsApp in February 2014 extended Meta’s reach into markets where Facebook had limited penetration and removed another potential competitor from the board.

Between 2016 and 2021, the company’s public image was defined by what went wrong more than what went right. Cambridge Analytica’s misuse of user data — 87 million profiles harvested through a third-party personality quiz — produced a congressional hearing in April 2018 in which Zuckerberg spent two days explaining how advertising targeting works to legislators who appeared largely unfamiliar with the concept. The five-billion-dollar FTC fine that followed was enormous in absolute terms and essentially an operational expense in relative ones. He settled an eight-billion-dollar shareholder lawsuit arising from the same scandal in 2025, this time choosing payment over testimony.

The metaverse was the most expensive mistake of Zuckerberg’s career, and it did something no antitrust regulator had managed: it made him look genuinely wrong. Between 2021 and 2023, Meta spent more than eighty billion dollars chasing an immersive digital world that users were not requesting, while TikTok — a company Zuckerberg had lobbied to have banned on national security grounds — quietly eroded the young-adult user base that Meta’s advertisers most valued. The rebrand from Facebook to Meta, intended to signal a pivot toward the future, came instead to signal a failure of imagination. When Zuckerberg pivoted to artificial intelligence, the move was widely regarded as the right call — which obscured a more important fact: he had been significantly wrong about the most consequential strategic bet he had ever made.

The AI era has gone differently. Llama, Meta’s open-weight large language model family, has shipped in successive generations since 2023, each more capable than the last, each free to download and modify. Llama 5, launched in April 2026 with a five-million-token context window and capabilities Meta describes as “System 2 thinking,” was received by the developer community as one of the most capable open models available. The strategic logic is not purely altruistic: open models make it harder for any single company to build a defensible moat, which works in Meta’s favor when the companies with the deepest moats are Google and OpenAI. Meta’s AI spending guidance for 2026 — between $125 and $145 billion — represents one of the largest capital commitments to a single technology platform in corporate history.

In June 2026, testifying in a lawsuit over social media addiction in Los Angeles, Zuckerberg argued that the popularity of Instagram and Facebook is itself evidence against the addiction claim — a position that drew attention for its circular logic as much as for its legal strategy. The appointment that same month to the President’s Council of Advisors on Science and Technology suggested that, whatever the litigation’s outcome, his relationship with Washington had evolved considerably since the 2018 hearings.

His wife Priscilla Chan, a pediatrician and co-founder of the Chan Zuckerberg Initiative, has directed the family’s philanthropy toward education and biomedical research. They have three daughters. Zuckerberg has spent the past two years training seriously in Brazilian jiu-jitsu and mixed martial arts, winning medals in competition — a shift in public persona that coincided, not entirely coincidentally, with a period of deliberately more physical, less robotic self-presentation.

The Social Reckoning, Aaron Sorkin’s follow-up to The Social Network, is in production with Jeremy Strong as Zuckerberg. The original film, which Zuckerberg disliked, managed to make him sympathetic despite portraying him as ruthless. Whether the sequel will do the same depends, probably, on which version of the man Sorkin finds most interesting: the one spending $145 billion to reshape the future, or the one who cannot quite step out of the machine he started.

Tags: business, meta, ceos, business leaders, instagram, facebook

Related People

Related People

Discussion