Loading...
Loading...
The Information reported that OpenAI and Broadcom’s previously announced plan to build 10 gigawatts of custom AI chips was presented as a near-final partnership, but the companies had not worked out how OpenAI would finance it. In a post citing reporter Anissa Gardizy, Gary Marcus highlighted the line that OpenAI “has made a habit of announcing landmark partnerships without ironing out the details,” and argued that many AI-industry tie-ups may hinge on financing that could still fall through. Th
Private equity firms Apollo Global Management and Blackstone are reportedly in talks to join a roughly $35 billion private credit financing package for Broadcom. The deal would fund Broadcom’s development of AI task-specific chips, reflecting investor interest in supporting semiconductor capacity tied to generative AI demand. Negotiations are ongoing and terms could change, according to people familiar with the matter. The involvement of large alternative asset managers signals significant institutional backing for chipmakers pivoting to AI hardware, and underscores the scale of capital being mobilized to secure supply and accelerate specialized AI silicon development.
Bloomberg : Sources: Apollo Global and Blackstone are among private credit lenders in talks with Broadcom over a ~$35B financing deal to fund the development of AI chips — Apollo Global Management Inc. and Blackstone Inc. are among private credit lenders involved in talks with chipmaker Broadcom Inc …
OpenAI’s in-house chip effort has hit a funding snag after Broadcom reportedly demanded Microsoft commit to buying 40% of initial production capacity. The two companies had announced a joint custom AI chip project last year, but internal memos and sources say Broadcom is pushing to finance the first production phase—requiring about 1.3 GW of data-center compute and roughly $18 billion—while the full “Nexus” program would scale to 10 GW and an estimated $180 billion just for chip production. The dispute raises questions about capital allocation, required off-take guarantees, and whether major cloud partners like Microsoft will underwrite the enormous upfront costs for bespoke AI silicon.
OpenAI’s in-house chip effort with Broadcom has stalled after Broadcom conditioned first-phase financing on Microsoft agreeing to buy about 40% of the initial wafer capacity. The project, codenamed “Nexus,” would scale to 10 GW and carry chip-production costs potentially as high as $180 billion for the first phase and $1.8 trillion overall, aimed at cutting OpenAI’s reliance on NVIDIA. Internal memos and sources say Broadcom wants Microsoft’s purchase commitment to de-risk its investment; Microsoft has reserved some datacenter space but hasn’t committed to purchases. If Microsoft’s buys fall short, OpenAI would need to find third-party buyers, posing a major financing and execution risk for its chip-and-datacenter strategy.
The Information reported that OpenAI and Broadcom’s previously announced plan to build 10 gigawatts of custom AI chips was presented as a near-final partnership, but the companies had not worked out how OpenAI would finance it. In a post citing reporter Anissa Gardizy, Gary Marcus highlighted the line that OpenAI “has made a habit of announcing landmark partnerships without ironing out the details,” and argued that many AI-industry tie-ups may hinge on financing that could still fall through. The post also notes that widely shared “circular financing” charts of AI deals are quickly becoming outdated and may omit major relationships, including the OpenAI–Broadcom effort. The report matters because it raises questions about the feasibility and timelines of large-scale AI infrastructure commitments.
Anissa Gardizy / The Information : Sources: OpenAI and Broadcom discuss terms for Broadcom to finance initial custom chip production for ~$18B, conditioned on Microsoft buying ~40% of the chips — When OpenAI and chip designer Broadcom announced last fall that they would make custom artificial intelligence chips together, they positioned it as a done deal.