Bank of America, Citi, Morgan Stanley and Blackstone announced a $63 billion financing structure for AI computing capacity. A newly formed special purpose vehicle will buy Broadcom XPU accelerators and networking equipment and lease them to Anthropic. The three banks are arranging first-lien financing, Blackstone is providing committed second-lien financing and Broadcom offers residual value support. Deliveries are expected to start in 2027.
What it means: The structure shows how AI financing is changing. Instead of the model company buying the hardware itself, a separate vehicle borrows, buys the chips and rents them out, much as airlines have leased aircraft for decades. The capital burden moves to banks and investment funds, while the chipmaker partly guarantees what the equipment is worth when the lease ends. That speeds up the build-out but also concentrates risk: if AI demand or chip values fall faster than expected, losses would be shared across a wide network of financial institutions. Supervisors such as the Bank of England and the ECB have already flagged how concentrated the market is around a few providers. For a Greek business, the takeaway is that the availability and cost of AI models now also depend on conditions in credit markets. Anyone signing long-term contracts for AI services should check price change and service continuity clauses, while banks and fintechs assessing such exposures need a clear view of how quickly the hardware loses value.
Related service: Custom software
Related stories
- OpenAI introduces GPT-6 Cyber for aligned cybersecurity research
- NVIDIA Cosmos offers open tools and checkpoints for interactive world simulation
- Mistral Studio expands with managed agents, workflows and enterprise apps
- Google launches Gemini Omni 1.2 Flash with 15-second video and 30% lower price
- Dropbox plans to buy Tigris and launches DBX Agent Cloud for AI storage