Loading page…
Apollo and Blackstone are reportedly among prospective lenders. The talks concern a chip collaboration with a multiyear deployment target, not a completed financing.
Listen to this story
Broadcom is seeking to assemble more than $50 billion in financing for custom AI chips it is developing with OpenAI, with Apollo Global Management and Blackstone reportedly among lenders in talks. The discussions are at an early stage, and neither the financing nor those lenders’ participation is confirmed, so the headline figure is not committed capital. The package would support OpenAI’s previously announced 10-gigawatt accelerator program; the next signal to watch is whether talks produce firm lender commitments and financing terms.
OpenAI is to design the accelerators and systems, while Broadcom helps develop and deploy them.
The companies targeted rack deployments beginning in the second half of 2026, with rollout completion by the end of 2029.
The Wall Street Journal attributed the financing discussions to unidentified people familiar with them.
OpenAI’s custom-chip effort with Broadcom could gain more than $50 billion in financing—a prospective funding package behind hardware the companies are developing together. Broadcom is arranging the financing, The Wall Street Journal reported, according to Bloomberg’s October 7 account. Apollo Global Management and Blackstone are among lenders in talks to participate, not confirmed backers.
The discussions are at an early stage, according to an Investing.com article carried by Yahoo Finance that also cites the Journal. That places the headline figure at the negotiating stage: it describes the package Broadcom is seeking to assemble, rather than money already secured for the chip program.
The Journal attributed the lender discussions to unidentified people familiar with them.
The proposed package would finance the custom AI chip Broadcom is developing with OpenAI; it is not a completed transaction.
The chip collaboration predates this financing report. On October 13, 2025, OpenAI and Broadcom announced a partnership covering 10 gigawatts of custom AI accelerators—processors built for AI computing. OpenAI was to design the accelerators and systems, with Broadcom helping develop and deploy them.
The companies said they had long-standing agreements to co-develop and supply the accelerators. They also signed a term sheet to deploy racks containing the processors and Broadcom networking equipment.
The announced systems use its Ethernet and other connectivity technologies, with deployments planned across OpenAI facilities and partner data centers.
OpenAI’s rationale for designing the hardware was to carry lessons from its models and products directly into the chips and systems.
The original schedule targeted the start of rack deployments in the second half of 2026, with completion by the end of 2029.
Loading discussion...
Join the conversation
Explain what would make the investment worthwhile—or too risky.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.