Facebook (META) began as a digital college yearbook, connecting Harvard students face to face. Three years ago on Monday, Mark Zuckerberg rebranded his social media empire as Meta, betting billions on a future where we’d meet in virtual worlds instead.
“Today we are seen as a social media company,” Zuckerberg said on Oct. 28, 2021, from the stage at the Facebook Connect augmented and virtual reality conference. “But in our DNA we are a company that builds technology to connect people, and the metaverse is the next frontier just like social networking was when we got started.”
Zuckerberg said he hoped the metaverse would reach a billion people and conduct “hundreds of billions” of dollars worth of commerce, employing millions — all within the decade.
Since then, Meta has invested more than $63 billion in Reality Labs, its division for virtual and augmented reality technology. The results include updated Quest headsets, which power Horizon Worlds, a virtual universe where customized avatars (initially designed without legs) roam digital spaces, and a partnership with Ray-Ban to develop AR glasses. These efforts drew skepticism from investors and a lukewarm response from consumers.
But everything changed when ChatGPT exploded onto the scene in November 2022. Meta was ready: Its AI research division, led by pioneer Yann LeCun, had long been advancing the field. And while Zuckerberg hasn’t given up on his metaverse dreams, his public messaging has increasingly focused on the transformative potential of generative artificial intelligence.
Those AI bets are already paying dividends. In July, Meta reported stronger-than-expected sales, crediting AI improvements in ad targeting. The company is now rolling out AI tools to help marketers enhance their listings.
“There are all the jokes about how all the tech CEOs get on these earnings calls and just talk about AI the whole time,” Zuckerberg said on the earnings call. “It’s because it’s actually super exciting and it’s going to change all these different things over multiple time horizons.”
He added that AI will “end up affecting almost every product that we have in some way.”
The pivot proved timely. After losing almost two-thirds of its value in 2022, Meta’s stock price almost tripled last year and is up more than 60% in 2024, hitting an all-time closing high of $595.94 per share earlier this month.
For Meta, the metaverse has dimmed and been replaced by all things AI, according to Gene Munster, a managing partner at Deep Water Management.
“They haven’t given up on the metaverse,” he said. “They are doing what every big tech company is doing, which is applying AI everywhere.”
Reality Labs will likely lose $20 billion this year, but Munster noted that about 70% of that investment is going toward AI-enabled glasses and wearables rather than Quest headsets. Meta has put at least $10 billion into AI glasses development alone, which has helped put the company ahead of competitors in the space.
But Munster said that the company’s “drunken sailor” spending on metaverse projects may not be sustainable alongside the growing costs of generative AI development, which has shown clearer financial benefits. While Meta doesn’t need to completely abandon Reality Labs, Munster said the next couple of years are crucial: Either the hardware needs to advance enough to prove the opportunity is real, or the spending needs to be redirected.
“One is going to have to give,” he said.
The future of the internet is 3D
Although the name change happened in 2021, Meta’s investment in the metaverse vision began long before. The company had bought VR headset maker Oculus for $2 billion in 2014 — twice what it paid for Instagram in a blockbuster deal two years earlier. By the time of the Meta rebrand, the company had already invested an estimated $32 billion in metaverse-related technologies.
But the “metaverse” itself is often misunderstood, according to Matthew Ball, an entrepreneur and author of “The Metaverse: And How It Will Revolutionize Everything.” While many associate it purely with VR headsets and virtual worlds, Meta’s vision is actually about the evolution of the internet itself into a 3D medium spanning virtual reality, augmented reality, and traditional screens.
“The metaverse is not limited to and does not even require virtual reality,” Ball said. “Most people now believe we will use no term to describe it whatsoever — we’ll just call it the 3D internet.”
But in the meantime, the proto-3D internet suffers. Global shipments of VR and AR headsets have sank roughly 28% since last September, with growth expected in 2025, according to market researcher IDC. But Meta’s Ray-Ban (EL) smart glasses have found success, with more than 730,000 units sold in their first three quarters, according to CNBC. The second generation of the glasses, which came out in September, got updates like better image quality, improved battery life and an AI voice assistant.
Meta’s recent triumph with Ray-Ban smart glasses has paved the way for an even more ambitious project: Orion. This cutting-edge eyewear prototype, which Meta has been developing for almost 10 years, represents a significant leap forward in capability.
When Zuckerberg unveiled Orion in September, it sparked an unprecedented wave of excitement about Meta’s metaverse technology — a notable shift from the previously tepid response to the company’s metaverse initiatives. The device showcases how AI could power next-generation augmented reality, with features like real-time 3D mapping and advanced scene understanding.
“Without AI, it would be impossible to virtualize the real world, or perhaps build the fantastical ones that we’ve imagined,” Ball said.
A new shiny diversion by any other name
In October 2021, when The Wall Street Journal started publishing The Facebook Files, exposing the company’s knowledge of its platforms’ harmful effects, Zuckerberg unveiled his metaverse rebrand. This was only a few years after election interference claims around Cambridge Analytica’s data collection on American voters that had Zuckerberg making multiple appearances in front of Congress.
At the time, Sara M. Watson, a technology critic and independent industry analyst, called the name change a diversion that didn’t address what the company had already built and broken. Now she sees the same playbook at work as Meta pivots to AI amid tepid response to its metaverse projects.
Meta has been aggressive in competing to build cutting-edge AI. In July, Meta revealed Llama 3.1, the first openly available model that the company says competes and, in some cases, beats the performance on benchmarks of several closed AI models, including OpeanAI’s ChatGPT and Anthropic’s Claude. When Zuckerberg published a 2,300-word blog post called “Open Source AI Is the Path Forward,” he positioned open models as a public good — while building another ecosystem that could entrench Meta’s power with the same “move fast and break things” approach that led to congressional mea culpas and, ultimately, the name change.
If Llama becomes the go-to infrastructure for AI tools, that would give Zuckerberg the same kind of control over generative AI that he once held over social networking, despite Meta’s troubled history with platform power.
“You build a thing that everyone needs, and they build on top of that,” Watson said. “Then you become essential no matter how you decide to monetize.”
Leave a Comment