Meta Platforms is actively working to engage external partners to finance the large-scale infrastructure required for the development of artificial intelligence. On Thursday, the company announced plans to sell data center assets worth $2 billion as part of this strategy.
This move reflects a broader trend among tech giants who have traditionally funded infrastructure expansion independently, but are now seeking new sources of capital due to the rapidly rising costs of building and powering data centers critical for generative AI.
Earlier this week, Meta said it is exploring partnerships with financial firms to jointly develop infrastructure and partially fund its substantial capital expenditures for 2025.
“We’re exploring opportunities to partner with financial firms to jointly develop data centers,” said Meta CFO Susan Li during a conference call following the release of quarterly results on Wednesday.
While the company still plans to self-finance most of its spending, Li said some projects may receive significant external funding to offer greater flexibility should infrastructure needs shift.
As of now, Meta has not finalized any deals to announce publicly, but its quarterly report filed Thursday already includes concrete steps: in June, the company approved a plan to sell certain data center assets and reclassified $2.04 billion in land and construction-in-progress as “held for sale.”
The transfer of assets to third parties for joint development is expected within the next 12 months. Meta recorded no impairment loss from the reclassification, as assets were valued at the lower of either book value or fair value minus selling costs. As of June 30, the total value of held-for-sale assets stood at $3.26 billion.
Meta declined to comment further on the specifics of the process.
CEO Mark Zuckerberg has previously outlined plans to invest hundreds of billions of dollars into building AI-powered data center “superclusters,” envisioned as the backbone for developing artificial superintelligence.
“One of them alone is the size of Manhattan,” he noted.
On Wednesday, Meta raised the lower end of its annual capital expenditure forecast by $2 billion to a range of $66–72 billion. Additionally, the company reported better-than-expected ad revenues, driven by improvements in AI-powered content delivery and targeting. According to executives, these earnings are helping offset long-term infrastructure costs tied to AI development.