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 …
Context & Ripple Effects
The related coverage shows this was an early report in a financing arc that later became a $35B package for Anthropic to lease TPUs, with Broadcom backstopping payments on the largest senior debt portions. A separate report described a Broadcom-backed tranche with lower indicated yields than a riskier portion, pointing to explicit risk segmentation rather than a conventional corporate loan.
The talks also sit alongside reported Broadcom discussions to finance initial custom-chip production for OpenAI. Together, the coverage ties AI-chip expansion to financing structures involving chip suppliers, AI-model customers, and private-credit firms.
First-order effects
- Apollo and Blackstone would become prospective lenders to a large Broadcom-linked AI-chip buildout, broadening their exposure from traditional corporate credit into infrastructure-like AI hardware financing.
- Broadcom would be positioned to support chip development with external debt capital; the later coverage indicates that its backing was central to securing financing for TPU capacity leased by Anthropic.
Second-order effects
- Debt backstops and tranche structures can lower the cost of capital for the most protected lenders while leaving higher-return, higher-risk exposure for other private-credit participants.
- AI companies seeking dedicated compute capacity may increasingly obtain it through leases financed against hardware and contractual payment support, rather than funding all capacity directly from their own balance sheets.
Third-order effects
- If replicated, AI compute investment could shift toward project-finance-style structures in which chip vendors, model developers, and private-credit managers divide operating, customer-credit, and residual hardware risks.
- The pattern would make access to large-scale AI infrastructure increasingly dependent not only on chip supply but also on lenders' willingness to underwrite long-duration utilization and payment commitments.
The trend: AI compute is becoming a financeable asset class, with private credit and vendor support helping convert custom-chip demand into leased infrastructure capacity.