Broadcom is in early talks to arrange financing that could help OpenAI purchase custom AI chips it is developing with the semiconductor giant. If the discussions move forward, the deal could involve around $30 billion in debt, highlighting just how much capital is now needed to build AI infrastructure at scale.
The discussions are still at an early stage. No formal financing process has started, and the plans could change, according to people familiar with the matter.
But the potential size of the deal is significant. It also adds to a growing trend in the AI industry: companies are increasingly turning to debt markets and large institutional investors to finance the enormous cost of chips, data centers and computing capacity.
Another multibillion-dollar AI financing deal takes shape
Broadcom has become increasingly involved in financing the infrastructure behind the AI boom.
The company has already been working on a major debt package connected to Anthropic’s AI infrastructure expansion, while also developing custom chips with OpenAI.
Now, Broadcom is exploring financing that could support OpenAI’s purchase of those chips.
According to people familiar with the discussions:
- Around $30 billion in debt could potentially be raised.
- The financing would help OpenAI purchase custom AI chips developed with Broadcom.
- Discussions are still preliminary.
- No formal financing process has begun yet.
- The final structure and size of the deal could change.
That distinction matters. This is not yet a completed financing deal. It is an early indication of how the companies may be thinking about funding their next phase of AI infrastructure.
Why does OpenAI need so much financing?
Building advanced AI systems is becoming a capital-intensive business.
Training and running large AI models requires enormous amounts of computing power. That means companies need access to:
- Advanced AI chips
- Large data centers
- Networking equipment
- Power infrastructure
- Cooling systems
- Long-term computing capacity
The cost quickly adds up.
For companies operating at the scale of OpenAI, relying entirely on operating cash flow to fund this infrastructure can be difficult. Debt financing offers another way to fund large infrastructure purchases while spreading the cost over time.
This is why investors are increasingly seeing AI infrastructure deals that look more like traditional large-scale infrastructure financing than conventional technology investments.
Broadcom is becoming more than a chip supplier
Broadcom’s role in the AI boom goes beyond selling semiconductor components.
The company is increasingly positioning itself as a key partner for companies building their own AI computing infrastructure.
In June, Broadcom said it had created a financing platform with Apollo Global Management and Blackstone as anchor investors. The platform is intended to help finance more than 20 gigawatts of AI computing capacity through 2028, supporting developers including Anthropic and OpenAI.
The scale of that ambition is enormous.
Broadcom has said fulfilling the broader agreement will require hundreds of billions of dollars.
That puts the potential OpenAI financing into a much larger story about how the AI industry plans to pay for its infrastructure build-out.
The Anthropic deal shows the scale of the financing wave
Broadcom is already involved in another major AI financing transaction.
Banks involved in a Broadcom-backed debt package linked to Anthropic recently began syndicating a $42 billion Class A senior-secured tranche.
Blackstone is leading an additional $18 billion Class B junior debt tranche, with the investment firm committing $9 billion from various funds.
The numbers show how quickly AI infrastructure financing is moving into territory traditionally associated with large infrastructure projects.
And Broadcom is not alone.
Oracle and SpaceX are also looking for billions
The push to finance AI computing is spreading across the technology industry.
Oracle is reportedly in discussions with lenders to arrange financing for chip purchases, while SpaceX has begun talks with banks and investors to raise around $40 billion to purchase Nvidia chips.
SpaceX’s potential financing is not expected to close until 2027, but the direction is clear.
Companies are looking for new ways to finance the hardware needed to support the next phase of AI growth.
The common factor is the same: AI demand is creating an infrastructure bill that is becoming too large to ignore.
What this means for Broadcom
For Broadcom, the potential OpenAI financing could reinforce its position as one of the major beneficiaries of AI infrastructure spending.
The company is already exposed to AI through its semiconductor business, particularly custom accelerators and networking technology.
A deeper relationship with OpenAI could strengthen that position.
But there is another important angle for investors.
If financing becomes an increasingly important part of AI infrastructure spending, companies involved in the hardware supply chain could benefit from a longer investment cycle.
That could mean sustained demand for:
- Custom AI accelerators
- Networking chips
- Data center infrastructure
- High-speed connectivity
- Power and cooling systems
Broadcom is positioned across several of these areas.
The bigger question: How much can AI infrastructure absorb?
The rapid increase in financing needs also raises an important question for investors.
How much capital can the AI ecosystem continue to attract before investors start demanding stronger returns on that spending?
The AI industry is making enormous investments based on expectations of future demand.
Companies are building computing capacity today because they expect AI usage, software revenue and enterprise adoption to keep expanding.
That creates an important investment debate.
If AI revenues continue to grow rapidly, these massive infrastructure investments could support a powerful long-term expansion.
But if demand grows more slowly than expected, companies could find themselves carrying significant financing obligations while trying to generate returns from expensive infrastructure.
What investors should watch
For investors tracking Broadcom, OpenAI and the wider AI trade, several developments will be worth watching.
1. Will the $30 billion financing move forward?
The discussions are preliminary, so investors should watch for any formal financing announcement.
2. Who will provide the capital?
The involvement of major banks, private credit firms or institutional investors could provide clues about how the market is pricing AI infrastructure risk.
3. How will OpenAI use the chips?
The scale and timing of OpenAI’s computing requirements will be important for understanding future demand.
4. Will more AI companies turn to debt?
If companies increasingly use debt to fund chips and data centers, it could become a major feature of the AI investment cycle.
5. Can AI revenues justify the infrastructure spending?
This may ultimately be the biggest question for investors.
The AI boom is becoming an infrastructure story
The conversation around artificial intelligence has increasingly moved beyond models and applications.
It is now about who builds the infrastructure, who finances it and who ultimately earns a return on the capital being deployed.
Broadcom’s potential $30 billion financing for OpenAI is another example of that shift.
The deal is not final, and there is no formal financing process yet. But even the early discussions underline the scale of the capital required to support the next generation of AI.
For investors, the AI story may no longer be just about which company builds the best model. It is also about which companies supply the chips, build the infrastructure and finance the massive computing expansion behind it.