Broadcom isn’t just supplying Anthropic’s AI infrastructure. It’s helping finance it, too.
The company has agreed to lend Anthropic as much as $42 billion to help finance its AI infrastructure, according to Anthropic’s IPO prospectus reviewed by Reuters.
The financing would take the form of convertible notes that could eventually become Anthropic shares. Broadcom can also designate a financing partner, while Anthropic said it does not expect any notes to be sold before completing its initial public offering.
The facility could cover roughly one-third of Anthropic’s $125.2 billion commitment under a five-year lease for TPU computing capacity. That gives Broadcom an unusually broad role in Anthropic’s expansion: supplying infrastructure, leasing equipment and potentially helping finance the same capacity Anthropic is committing to use.
The financing comes as Anthropic prepares for a massive expansion of its computing capacity. Anthropic has committed to long-term infrastructure agreements involving Google, Amazon, Broadcom, Microsoft, xAI and AMD. The company has said it expects demand for advanced AI systems to exceed supply and that compute availability will be a major bottleneck.
Broadcom and Google have worked together on several generations of TPUs, and Anthropic announced in April that it would gain access to multiple gigawatts of next-generation TPU capacity beginning in 2027.
Anthropic is expected to become Broadcom’s largest compute customer that year. Broadcom projects about $115 billion in AI semiconductor revenue for fiscal 2027 and $230 billion for fiscal 2028, Reuters reported.
The financing comes with risks
Anthropic itself has warned that the arrangement creates potential complications.
The company said Broadcom’s simultaneous role as hardware supplier and financing partner creates “potential conflicts of interest” that could affect Anthropic’s ability to obtain the computing power it needs. The prospectus also warns that Broadcom’s pricing and hardware decisions could affect how much infrastructure Anthropic can procure.
Certain payment or performance defaults could also cause a substantial portion of Anthropic’s lease obligations to become immediately payable while restricting its ability to use the $42 billion facility to cover those obligations.
The financing structure has also drawn comparisons with Nvidia’s use of its balance sheet to support companies building AI infrastructure around its hardware.
“Nvidia is putting in place a massive amount of its balance sheet, and Broadcom is having to follow suit,” Seaport Research analyst Jay Goldberg told Reuters.
The comparison highlights a broader shift in the AI infrastructure market: chip suppliers are increasingly tied not only to the technology their customers deploy, but also to how those deployments are financed.
What this means for channel partners
Anthropic’s financing arrangement shows how AI infrastructure deals are increasingly tying together hardware, cloud capacity, leasing and financing.
For solution providers, integrators and infrastructure partners, Anthropic’s expansion could generate demand across data-center hardware, networking, power, cooling, cloud infrastructure and managed services. But much of that spending is increasingly being anchored by a small group of hyperscalers and semiconductor companies with direct relationships to the AI developer.
The structure also shows why partner relationships could become more important as AI companies scale. Anthropic’s ability to expand will depend partly on whether its infrastructure suppliers can deliver the computing capacity it has committed to purchase or lease.
Any disruption involving pricing, hardware availability, financing or contractual obligations could therefore affect multiple companies connected to its AI infrastructure chain.
For the channel, the larger takeaway is that the AI infrastructure opportunity is becoming as much about financing and supplier relationships as compute capacity itself. Partners that understand those dependencies may be better positioned to identify where new infrastructure spending is likely to flow — and where concentration creates additional risk.
Other news: Dell, JERA, and RHAELM are starting with a $15 billion AI data center in Japan, while the broader multi-site infrastructure buildout could eventually reach $140 billion.