The $35B 'Big Sky' Shadow Finance: How Apollo, Blackstone, Google, and Broadcom Are Wiring Anthropic Into an Exclusive AI Power Nexus

In one of the largest private credit transactions ever recorded, Apollo Global Management and Blackstone have closed a $35 billion debt facility structured through a special purpose vehicle (SPV) to finance Anthropic’s acquisition and leasing of Google’s custom tensor processing units (TPUs). Dubbed Project Big Sky, the deal mirrors Meta’s earlier $27.3 billion 'Beignet' SPV with Blue Owl, using a mix of debt tranches and equity where lease payments from Anthropic back the financing. Senior notes totaling roughly $30 billion benefit from Broadcom’s credit backstop—effectively linking the credit quality to Broadcom’s investment-grade rating rather than the startup’s—allowing cheaper borrowing amid banker caution over AI-related debt and rapid chip obsolescence.[1][2]

This arrangement highlights a deeper structural shift: frontier AI infrastructure is increasingly funded outside traditional public markets and corporate balance sheets. With Morgan Stanley projecting AI could drive $400 billion in debt this year alone, scaling toward $1 trillion by 2028 to support nearly $2 trillion in cumulative capex, private credit giants are stepping into the void left by risk-averse banks. The SPV buys the TPUs, leases them to Anthropic (keeping the liability off-balance-sheet), and securitizes the revenue stream much like aircraft or data-center real estate financing. Connections often missed include the tight ecosystem lock-in: Google supplies the silicon, Broadcom provides implicit guarantees and co-development ties, while Apollo and Blackstone control the capital stack. This concentrates compute access, pricing power, and strategic leverage among a handful of interconnected players—Anthropic (fresh off a $65 billion private round and confidential IPO filing), its big-tech backers, and alternative-asset titans—bypassing broader market scrutiny.[3][4]

Anthropic’s near-simultaneous confidential S-1 filing with the SEC underscores the strategy: scale aggressively in private with structured finance, then tap public equity once the infrastructure moat is dug. Yet risks abound. Chip depreciation timelines remain uncertain as models evolve; should the AI investment cycle cool, these massive SPVs could transmit stress through private-credit funds already facing redemption pressures. What ZeroHedge flagged as a burgeoning wave is materializing faster than expected, signaling not just capital deployment but a reconfiguration of tech sovereignty where financing partners become de-facto co-owners of the AI stack.