October 8, 2026 / SemiMedia / — Broadcom is seeking more than $50 billion in financing to support custom AI chips and related computing infrastructure being developed with OpenAI, as the cost of large-scale AI deployments continues to push semiconductor companies and technology firms toward new funding structures.
Apollo Global Management and Blackstone are among the lenders in early discussions around the financing. The transaction is expected to support several gigawatts of OpenAI computing capacity, although its final size and structure remain under negotiation.
Broadcom already established an AI infrastructure financing platform with Apollo and Blackstone in June. The platform is designed to support more than 20GW of computing capacity for frontier AI companies including OpenAI and Anthropic through 2028. Its initial transaction totaled $35 billion and was led by Apollo with participation from Blackstone, primarily supporting more than 1GW of Anthropic infrastructure.
Broadcom has also expanded its financing relationship with Anthropic. Recent disclosures showed that the company agreed to provide up to $42 billion to help finance AI computing capacity, while separate financing discussions linked to Anthropic and other customers have reportedly reached more than $60 billion.
The transactions reflect a broader change in how AI infrastructure is funded. Major cloud companies historically financed most data center investment from internal cash flow, but rapidly rising spending on accelerators, networking, power systems and data center construction is increasingly bringing banks, private-credit firms and infrastructure investors into AI hardware deployments.
Oracle is separately discussing financing with Apollo and Goldman Sachs for additional AI chip purchases. Some proposed structures involve dedicated entities purchasing the hardware and leasing it back to operators, allowing companies to spread large infrastructure costs over longer periods.
For Broadcom, financing is becoming increasingly connected to its custom accelerator and networking businesses. As OpenAI, Anthropic and other large AI developers expand proprietary chip deployments, the company is moving beyond semiconductor design and connectivity into the financial infrastructure required to support multi-billion-dollar compute projects.
The trend suggests that competition in AI infrastructure is increasingly determined not only by chip performance, but also by access to manufacturing capacity, networking, power and capital.







All Comments (0)