Yazılar

Meta Strikes $27 Billion Financing Deal With Blue Owl for Massive Louisiana AI Data Center

Meta (META.O) has finalized a $27 billion financing partnership with Blue Owl Capital (OWL.N) to fund its largest data center project to date — a massive AI computing hub in Louisiana designed to supercharge the company’s artificial intelligence ambitions.

The agreement, Meta’s biggest-ever private capital deal, gives Blue Owl-managed funds a majority ownership stake in the joint venture, while Meta retains 20% equity. Blue Owl contributed about $7 billion in cash, and Meta will receive a $3 billion one-time payout, according to Tuesday’s announcement.

The planned Hyperion Data Center in Richland Parish, Louisiana, will deliver over 2 gigawatts of computing capacity, a figure that underscores the escalating global demand for infrastructure to train large language models such as ChatGPT and Google Gemini. Blue Owl co-CEOs Doug Ostrover and Marc Lipschultz called the project “an ambitious step toward powering the next generation of AI infrastructure.”

The move comes amid a historic wave of investment in AI-related data centers. According to Morgan Stanley, leading tech giants — including Alphabet, Amazon, Meta, Microsoft, and CoreWeave — are collectively set to spend $400 billion this year building AI infrastructure.

Meta CFO Susan Li described the partnership as “a bold step forward,” noting that the project will create more than 500 jobs and help the company diversify its financing strategy while reducing exposure to debt.

Industry analysts say the deal enables Meta to offload capital risk while maintaining operational control of a strategic AI asset. “It allows Meta to finance expansion without taking on heavy debt — a smart hedge if the AI market overheats,” said Alvin Nguyen, senior analyst at Forrester.

The Hyperion facility is expected to go online within four years, with Meta holding lease options to extend. Once operational, it will stand among the largest data centers in the world, symbolizing the scale of investment driving the AI revolution.

Internal Meta Study Finds Instagram Shows More “Eating Disorder-Adjacent” Content to Vulnerable Teens

An internal Meta study reviewed by Reuters has revealed that teenagers who report feeling bad about their bodies after using Instagram are shown significantly more “eating disorder-adjacent” content than their peers. The internal document, marked “Do not distribute,” highlights serious concerns about how Instagram’s recommendation system interacts with vulnerable teens.

Meta researchers surveyed 1,149 teenagers throughout the 2023–2024 school year and analyzed the posts appearing in their feeds. Among the 223 teens who said Instagram regularly made them feel worse about their bodies, 10.5% of their feed contained body-focused or disordered-eating-related content — three times higher than the 3.3% seen by other teens. The flagged posts prominently displayed body parts, expressed judgment about physical appearance, or included material associated with negative body image.

Researchers also found that these same teens encountered more “mature” and “provocative” content overall — material involving risk-taking, suffering, and cruelty — which made up 27% of their feed compared to 13.6% for others. While the study could not prove Instagram directly worsens self-esteem, the correlation raised alarms among Meta’s internal experts.

Meta spokesperson Andy Stone said the findings show the company’s commitment to making platforms safer for young people. However, the report revealed that Meta’s moderation tools failed to detect 98.5% of sensitive content potentially inappropriate for teens. Pediatric experts like Jenny Radesky from the University of Michigan called the results “deeply disturbing,” warning that Instagram’s algorithm may be “profiling vulnerable teens and feeding them more harmful content.”

The findings come as Meta faces ongoing lawsuits and investigations in the United States over its alleged failure to protect minors and the mental health risks tied to Instagram’s design.

Meta to invest $1.5 billion in massive AI data center in Texas

Meta Platforms announced plans to invest $1.5 billion in a new data center in El Paso, Texas, marking its 29th global facility and third in the state, as part of its ongoing expansion to support artificial intelligence workloads. The new site will be one of the largest in the U.S., capable of scaling to 1 gigawatt of power — enough to supply a city the size of San Francisco for a full day.

Expected to go online in 2028, the El Paso data center will create around 100 permanent jobs and employ more than 1,800 construction workers at peak development. Meta said it chose El Paso for its strong electrical grid, skilled workforce, and access to renewable energy.

The project is part of a broader race among tech giants to build infrastructure for AI. Amazon, Alphabet, Microsoft, and Meta together are projected to spend more than $360 billion in 2025 on cloud and AI infrastructure, most of it directed toward data centers.

Meta has already invested over $10 billion in Texas and employs more than 2,500 people statewide. The company said the new facility will be powered by 100% renewable energy and use a closed-loop, liquid-cooling system to recycle water continuously. It also pledged to restore twice the water consumed by the site to local watersheds, exceeding its 2030 sustainability goal of becoming “water-positive.”

Jon Barela, CEO of the Borderplex Alliance, which helped facilitate the project, described Meta as “the fastest gazelle in the industry,” noting that the deal stemmed from discussions with Texas Governor Greg Abbott’s office four years ago and was supported by a package of local tax incentives.