Can Meta actually win the AI race, or is the idea just another one of Mark Zuckerberg's pipe dreams?
The author's argument: Zuckerberg's fantasies, combined with his delusions of genius, potentially overstate the role that luck and outside input have played in the rise of his social media empire, and could be blinding him from reality.
This could potentially be applied to all the leading minds in tech. Elon Musk, for example, has invested wisely, but government grants and the research breakthroughs of others have amplified his achievements. Sam Altman didn't create the technology that powers OpenAI, but he has become the face of the business, which lends more weight to his opinions.
Like every success story, there's some luck involved. It's impossible to discount the serendipity of being in the right place at the right time.
The original idea for Facebook may not have even been entirely his, yet he parlayed the platform into a trillion-dollar business through clever business strategies and smart acquisitions.
Instagram and WhatsApp.
Meta was unable to acquire Snapchat. In 2013, the BBC reported that Snapchat chief Evan Spiegel rejected Meta's $3 billion takeover offer — a rebuff that prompted significant investment in various Snapchat-like apps and formats.
| Attempt | Copied from | Outcome |
|---|---|---|
| Slingshot (2014) | Snapchat | Failed |
| Stories | Snapchat | Gained traction but cost a lot of money and time, and hasn't overpowered Snapchat |
| Bonfire | Houseparty (group live-streaming) | Failed to gain traction |
| Hotline | Clubhouse (audio chat) | Failed to gain traction |
In fact, Meta doesn't have much to show off in terms of innovation, apart from replicating or acquiring other apps and tools.
The Portal video device, drone-delivered internet for remote regions, its cryptocurrency project, and Instant Articles for publishers — all failed dramatically.
Fortunately, Meta's core ad business is so strong that there's no major business impact from these failures. These experiments make sense in the broader context of pursuing relevance and growth.
The metaverse was one of Zuckerberg's most high-profile and expensive experiments, with a major promotional push showcasing what it presented as the next generation of digital connectivity.
Then AI gained traction. Zuckerberg decided this was the actual tech development of a generation, not the metaverse, and immediately became obsessed with winning the AI race.
He lost interest in the metaverse vision he had been all-in on just a year or so before OpenAI released ChatGPT.
Meta has now invested hundreds of billions of dollars into data center projects, flashy staff hires and systematic updates, leveraging its scale to overtake AI competitors.
Meta may not even want to win this one, given the questionable profit margins that come with these expanding AI projects.
While the tech industry is obsessed with AI, practical use data isn't matching the hype. Many businesses adopting AI tools haven't seen the promised productivity, and many have not been able to reduce staff costs by outsourcing work to AI agents.
A study published earlier this year by the National Bureau of Economic Research found that among nearly 6,000 CEOs, chief financial officers and other executives, the vast majority reported seeing little operations-level impact from AI.
If the predicted gains can't be realized, Meta may be burning money on yet another expensive project.
Meta reportedly sunk more than $80 billion on the metaverse, though per Business Insider much of that development was later transferred to other projects and Meta is still working on VR. Even losing half that amount would be a massive hit.
The AI bets are riskier still.
| Item | Figure |
|---|---|
| Time to break even even at $100 billion a year in AI subscriptions | More than a decade |
| Meta total revenue, 2025 | $200.97 billion |
| Of which non-advertising | $4.8 billion |
That means the company needs to turn AI into a business in its own right, and ensure that business is at least half as profitable as one of the most profitable businesses in the world — just to recover the costs already sunk.
Is this even possible? Given Meta's track record for innovation, it's not clear the company has the capacity to make this work, unless it copies or acquires some other breakthrough AI provider.
Re-examine how much you depend on a platform roadmap. This is a company that went all-in on the metaverse and pivoted within a year. Betting budget on a platform's promoted new feature is betting on that company's attention span.
Face the Reels dependency. Almost all engagement growth comes from a TikTok clone — Meta channel strategy is effectively short-form strategy. See Building a Brand on Facebook in 2026.
Validate "AI productivity" claims yourself. The NBER finding that most of nearly 6,000 executives saw little operations-level impact sits directly against vendor claims — the same conclusion appears in Why 88% of Enterprise AI Projects Fail.
Understand that the ad business underwrites the experiments. That is why failures cost Meta little — and its current state is covered in Meta Grows Ad Revenue 27% to $59.4B. For advertisers, the bill for this investment may arrive as ad pricing.
Assess AI hardware adoption separately. Even the EssilorLuxottica-assisted glasses hinge on social acceptance rather than technology — see The AI Glasses Privacy Backlash.
Instagram and WhatsApp acquisitions succeeded, but most in-house inventions failed: Slingshot (copying Snapchat), Bonfire (Houseparty), Hotline (Clubhouse), plus Portal, drone internet, cryptocurrency and Instant Articles. Reels was copied from TikTok and Stories from Snapchat.
Hundreds of billions across data centers, hires and systems. Even at $100 billion a year in AI subscriptions it would take more than a decade to break even. Of $200.97 billion total 2025 revenue, only $4.8 billion came from non-advertising sources.
A National Bureau of Economic Research study published earlier this year found that among nearly 6,000 CEOs, CFOs and other executives, the vast majority reported seeing little operations-level impact from AI.
Reportedly more than $80 billion. Business Insider reported much of that development was transferred to other projects and VR work continues, but even losing half would be a massive hit.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
Opening an AI before a search engine is becoming the default. As the path back to original creators narrows, where does competitive advantage in the AI era actually come from?
Meta expanded its vibe coding app Pocket to all U.S. users. Describe what you want and it builds a small app or game — then that output circulates in a social feed.
Meta's CEO published a 6,500-word essay on personal superintelligence that doubles as an argument against AI regulation — while Meta cut 20% of staff earlier this year.