In This AI Moment, Meta's Zuckerberg Again Embraces Disruption, Inside and Out -- Barrons.com

Dow Jones
Oct 25

By Adam Levine

Meta Platforms CEO Mark Zuckerberg is on a mission, and it is the same one he has been on since Facebook became the dominant social media site in the 2010s: Don't be disrupted, and be willing to disrupt yourself. This past week, he continued down the path of shaking up Meta for the artificial-intelligence age by finding creative private financing to fund Meta's enormous capital commitments. For good measure, he also shook up his staff.

In a business where, as former Intel CEO Andy Grove so famously said, "only the paranoid survive," Zuckerberg continues to stand out as the most paranoid of CEOs.

Meta may still struggle to stay ahead in this AI moment, but it won't be because the company didn't try hard enough. The company is on pace to spend about $70 billion this year on new AI data centers, with much more promised. Alone among the big capital investors in AI, Meta is buying all this computing capacity for its own use, not to rent out to customers in the cloud.

And the capital expenditures are beginning to weigh on Meta's financials, though operations continue to excel, with operating profits up 38% in the second quarter. The stock is up 25% for the year, compared with 19% for the tech-heavy Nasdaq 100.

Even with operational cash flows growing quickly in 2024 and 2025 -- some $102 billion in the last 12 months -- cash flow is beginning to decline once you subtract capex. Operational cash flows alone aren't going to be enough to fund Zuckerberg's AI data-center ambitions, so Meta is tapping outside financing. The company added over $10 billion in new debt since last year. In the second quarter, its debt and lease liabilities exceeded its cash and short-term investments for the first time.

And this past week, Meta got creative. The company came up with a way to fund its massive Hyperion data center in Louisiana, which is already under construction. Meta forged a joint venture with investment manger Blue Owl Capital that will own the site, though Meta will be its only customer for at least the next four years. Meta will own only 20% of the JV, so it will not appear on the company's consolidated balance sheet. Nor will the capex hit its cash-flow statement. Blue Owl will own the remainder, adding $7 billion of its own money in bonds sold to Pimco and other lenders.

Meta will also be contributing land and building assets valued at up to $3.3 billion in June, though the company may have added other Hyperion assets since then. Meta will also receive a $3 billion payout from the JV. Neither Meta nor Blue Owl responded to a Barron's request for comment on the venture.

Ultimately, Zuckerberg believes he needs all this computing capacity to serve up AI to the 3.5 billion people who visit at least one of Meta's apps every day. Zuckerberg wants to create AI language models that can compete with the best in the business, which would power new user experiences and better ad-targeting across its products. There is already evidence that Meta is seeing success in using AI to increase ad sales.

But Zuckerberg also fears that its AI models are falling behind ones that power OpenAI's ChatGPT, still the leader. Meta released its Llama 4 model in April to little acclaim, and Zuckerberg then decided he had to disrupt his own AI engineering teams.

Beginning in June, Meta went on a hiring spree that is not yet done, reportedly offering nine-figure multiyear deals to the world's top AI researchers. The primary gambit was to buy 49% of Scale AI at a $29 billion valuation, double what it had been before after its previous fund-raising round. Scale AI's CEO, Alexandr Wang, became Chief AI Officer for the newly formed Meta Superintelligence Labs, overseeing AI research. Among the teams below him, TBD Lab houses the new pricey scientists, who will work on future AI models, and the Fundamental AI Research Team, or FAIR, which had been developing the Llama models.

Zuckerberg thus upended the organization. FAIR and longtime Chief AI Scientist Yann LeCun receded into the background, and one of the new hires, Shengjia Zhao, now also holds a chief scientist title.

With the structure a mess from new hiring and shifting responsibilities, Meta moved this week to make more sense of it all. On Wednesday, Axios reported that the company will begin to lay off 600 workers across its AI units in a reorganization effort. The layoffs are happening outside of TBD Lab -- indeed, TBD is still hiring. The New York Times estimates Meta's total AI staff at around 3,000.

All this is why Zuckerberg remains one of the best CEOs in tech. He never rests on yesterday's successes and doesn't get complacent. In the 2010s, even as Facebook became the dominant social-media site, he fretted endlessly about competition from Amazon.com's Google, Foursquare, Pinterest, and Snapchat, and that worrying is one reason none of them surpassed Meta.

"Move fast and break things" was the early Facebook corporate mantra, and now that spirit is back. No other Big Tech company has expended as much money or effort to restructure its business model for AI. But even that doesn't guarantee success, as Zuckerberg would probably be the first to say.

Write to Adam Levine at adam.levine@barrons.com

This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.

 

(END) Dow Jones Newswires

October 24, 2025 14:29 ET (18:29 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.

At the request of the copyright holder, you need to log in to view this content

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Most Discussed

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10