AI infrastructure now needs its own financing industry
Broadcom (NASDAQ: AVGO) is preparing to do something much bigger than sell Anthropic chips.
The semiconductor company has agreed to provide Anthropic with up to $42 billion of financing as the Claude developer races to secure enough computing power for the next generation of its AI models.
The size of the arrangement is remarkable on its own. The financing could cover roughly one-third of a $125.2 billion, five-year lease for TPU computing capacity, according to details disclosed in Anthropic’s IPO prospectus.
Broadcom sits directly inside that spending. It designs Google’s Tensor Processing Units and is helping supply the infrastructure Anthropic intends to use beginning in 2027. Anthropic is expected to spend enough across the relationship to become Broadcom’s largest chip-design customer that year.
This is where the AI boom starts looking less like a normal semiconductor cycle and more like an infrastructure buildout.
The constraint is no longer simply whether enough advanced chips can be manufactured. Someone also has to finance the enormous data centers, equipment and compute contracts needed to put them to work.
Anthropic is locking up compute years in advance
Anthropic has been unusually aggressive about securing capacity.
Earlier this year, the company expanded its relationship with Google and Broadcom for multiple gigawatts of next-generation TPU compute beginning in 2027. Broadcom has said approximately 3.5 gigawatts of that capacity is expected to flow through infrastructure tied to its technology.
Anthropic is also building across several different chip ecosystems rather than betting everything on one supplier. Its infrastructure plans include Amazon’s Trainium processors, Google TPUs and Nvidia GPUs.
That tells us something important about where frontier AI is heading.
The largest model developers increasingly need so much compute that they cannot rely on one cloud provider, one chip architecture or one data-center network. They are effectively assembling global computing portfolios years before all of the capacity is needed.
Broadcom has already started building the financial machinery around that demand. In June, the company teamed up with Apollo and Blackstone on an AI infrastructure financing platform designed to support more than 20 gigawatts of deployments, beginning with a $35 billion transaction tied to more than one gigawatt of Anthropic-related infrastructure.
Broadcom is positioning itself to make money not only from the silicon inside AI systems, but from helping those systems get financed and built.
The circular-financing debate is getting harder to ignore
There is an unusual loop developing across the AI industry.
Technology companies invest in or lend money to AI labs. Those labs then spend enormous amounts with the same companies on chips, cloud capacity and infrastructure.
Anthropic has versions of that relationship with several major technology partners. Broadcom’s potential $42 billion facility simply makes the scale much harder to miss.
The structure itself is not necessarily a problem. Railroads, telecommunications networks and energy infrastructure have all required massive amounts of financing before demand fully matured. AI may be entering a similar period where suppliers and capital providers have to fund capacity years ahead of the revenue it eventually generates.
But investors will increasingly need to separate organic AI demand from demand being accelerated by financing.
If Claude continues gaining enterprise customers and Anthropic can turn that usage into enormous recurring revenue, today’s infrastructure commitments could look like companies racing to secure scarce capacity before it becomes even more valuable.
If AI revenue grows more slowly than the capital being committed, the financial relationships between suppliers and customers become much more important.
That makes Anthropic one of the most important companies to watch as the sector matures. It is not only a major AI developer. It is quickly becoming one of the largest buyers of computing infrastructure in the world.
Broadcom may be building a second AI advantage
Nvidia became the defining company of the first stage of the AI boom by controlling the most sought-after accelerators.
Broadcom’s opportunity looks different.
Its position in custom AI chips, networking and Google’s TPU ecosystem already gives it exposure to the shift toward specialized compute. Financing adds another layer.
A frontier AI company deciding where to place billions of dollars of workloads may increasingly care about more than performance. It may also care about who can guarantee supply, assemble the infrastructure and help finance the deployment.
Broadcom is moving toward offering all three.
The $42 billion Anthropic facility is therefore more interesting than another giant AI spending number. It shows how quickly the economics of the industry are changing.
AI’s next competitive advantage may not belong solely to the company with the best chip. It may belong to the company capable of financing an entire computing ecosystem around it.
Disclaimer
Wall Street Access is an independent financial publisher. Wall Street Access is not a financial advisor. No company mentioned has compensated Wall Street Access for the preparation, publication or distribution of this article. This content is provided for informational and educational purposes only and is not investment advice or a recommendation to buy, sell or hold any security. Investing involves risk, including the possible loss of principal.
Prices, market capitalizations, financial figures and other market data are based on information available as of the date and time of publication and may change without notice. Information is drawn from public sources believed to be reliable but is not guaranteed to be complete or accurate. Opinions and expectations expressed are our own and may change as new information becomes available.
Readers should conduct their own due diligence and contact a licensed financial professional before making any investment decision. Past performance does not guarantee future results. Any material ownership interest or conflict of interest, where applicable, will be disclosed separately.
