Meta Platforms Inc. has selected Pacific Investment Management Co. (PIMCO) and Blue Owl Capital to lead a roughly $29 billion financing package to build out its planned AI-focused data-center campus in rural Louisiana, people familiar with the matter told Bloomberg and Reuters. The deal is the latest sign that Big Tech is turning to private credit markets and strategic partners to help shoulder the skyrocketing costs of building the massive compute infrastructure needed for generative-AI workloads. Pimco will offer $26 billion in debt, and Blue Owl will provide $3 billion in equity For weeks, Meta has been in talks with several private credit firms to identify those who could help secure up to $29 billion to finance its large-scale expansion of AI data centers in the US. As of late June, Meta was still in discussions with Apollo Global Management, KKR, Brookfield, Carlyle, and Pimco. The firm was also working alongside Morgan Stanley to assess ways to improve the debt’s marketability. According to reports, Meta was hoping to secure roughly $3 billion in equity and $26 billion in debt. With the latest selections, insiders say Pimco will handle the $26 billion debt financing, while Blue Owl will provide the $3 billion in equity. They added that the debt portion will probably be issued as bonds. They further noted that Apollo Global Management Inc. and KKR & Co. remained contenders to lead the deal up until the final round of negotiations. Lately, private credit players have been increasing their involvement in AI infrastructure. For instance, Blue Owl recently pledged $15 billion to a data center joint venture with OpenAI, which is still collaborating with SoftBank and Oracle on a $500 billion AI facility project. Meta is funneling more investments into AI projects Meta has been stepping up its spending on AI. In June, it revealed plans to invest $14.3 billion in Scale AI , taking a 49% stake without voting rights. The firm also recruited Scale AI’s founder, Alexandr Wang, and some other employees as part of the deal, to which the startup claimed would not affect users since Meta would not have access to business-related data. Around the same time as the Scale AI deal, OpenAI, the company behind ChatGPT, claimed Meta was luring its staff with signing bonuses of up to $100 million and even larger annual pay packages. Recently, Meta also unveiled its Llama 4 model. However, the AI community’s response to the model has been far from enthusiastic. Developers have flagged weak programming and software development performance, adding that Llama 4 trails innovative challengers such as DeepSeek. For starters, the model’s Scout architecture processed a lengthy 20,000-token document but generated what AI researcher Simon Willison labeled “complete junk output,” with multiple loops. The tech giant has also promoted its Llama 4 as “open source,” yet licensing rules limit genuine open access. Users have to sign in and agree to the license terms before downloading the models. Besides, the model’s release on a weekend sparked outrage among community members. When Meta CEO Mark Zuckerberg was asked what led to the unusual timing, he remarked, “That’s when it was ready.” Nevertheless, Meta has raised its full-year capital spending outlook to as high as $72 billion, pointing to data center buildouts and mounting infrastructure expenses. In a first for the company, it has also struck a deal to procure nuclear power for its AI operations. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free .
