🤖 AI Summary
Meta has launched a $30 billion bond sale — split into six tranches maturing between 2030 and as late as 2065 — to help finance a major buildout of AI datacenter capacity. The filing, arranged with Citigroup and Morgan Stanley, accompanies a joint venture with Blue Owl Capital that will fund $27 billion of work on the Hyperion campus (Blue Owl 80%, Meta 20%), including buildings, power, cooling and connectivity; some JV capital will come from a private debt offering. Meta reported Q3 revenue of $51.2B and said quarterly capex was $19.4B; it now expects 2025 capex of $70–72B (up from prior guidance) versus $39.23B in 2024, underscoring the scale of spending behind its AI roadmap.
The move is significant because it crystallizes how AI’s infrastructure needs are reshaping corporate finance and industry strategy: companies are taking long‑dated debt and forming specialized funding vehicles to underwrite multidecade investments in power, cooling and network capacity. It also intensifies the datacenter arms race—Google, Microsoft and Oracle have likewise ramped capex and borrowing—raising questions about grid capacity, site-level engineering (power and cooling), and credit risk as firms lock in capacity for an uncertain AI demand curve. Long maturities and JV structures shift risk, free capital for model and product development, and signal that hyperscalers view AI infrastructure as a multi‑decade strategic asset.
Loading comments...
login to comment
loading comments...
no comments yet