Broadcom is arranging more than $50 billion in financing for the custom AI chips it’s developing with OpenAI, according to The Wall Street Journal — the latest mega-deal in a frantic week that’s turned the AI chip boom into Wall Street’s busiest credit market.

The chip giant has spent recent weeks talking to private-credit firms about the raise, the Journal reported on October 7, citing people familiar with the discussions. Apollo Global Management and Blackstone are among the lenders approached. The financing could back several gigawatts of computing capacity and might close before the end of the year, though the talks are preliminary and the final amount could still change.

Fifty billion dollars is a staggering number. In the current AI economy, it’s starting to look routine.

How the Broadcom OpenAI AI chips financing would work

Broadcom isn’t exactly writing the check itself. The financing would fund deployments of AI accelerators built on Broadcom-designed silicon — custom chips that OpenAI is developing in partnership with the company for its own data centers, cutting its dependence on Nvidia GPUs and on renting cloud capacity.

The arrangement echoes a playbook Broadcom rolled out in June: an AI financing platform with Apollo and Blackstone as anchor investors, designed to enable more than 20 gigawatts of computing capacity built on its custom chips for frontier AI labs, including Anthropic and OpenAI, through 2028. That platform launched with an initial $35 billion tranche led by Apollo, the Business Times reported, citing Bloomberg.

It’s a clever structure. Broadcom lines up the financing for the infrastructure its customers need, which locks in demand for its own chips while someone else’s capital absorbs the construction risk.

Borrowing and lending at the same time

Here’s the strange part: Broadcom is on both sides of the lending desk.

Days before the OpenAI financing report surfaced, reports emerged that Broadcom would lend Anthropic up to $42 billion to buy computing built on Broadcom-designed chips — a commitment that appeared in Anthropic’s IPO prospectus. Bloomberg also reported in August that Broadcom was talking to lenders about a separate debt package of more than $60 billion to help Anthropic and other companies pay for AI chips.

So the company is borrowing tens of billions to finance one customer’s chips while lending tens of billions to another. Either Broadcom has found a brilliant way to sit at the center of every AI hardware deal, or it’s placed an enormous bet on a boom that everyone assumes will never cool.

Everyone’s in line for chip debt

Broadcom is hardly alone in the queue. The financing wave of the past week reads like a league table of the AI buildout:

  • SpaceX is in early talks for roughly $40 billion in debt to buy Nvidia GPUs — a package that would split into bank loans and investment-grade bonds, with Apollo expected to lead, according to reports from multiple outlets.
  • Anthropic has a separate, roughly $60 billion debt package already being syndicated by banks: a $42 billion senior secured tranche led by Bank of America, Citigroup and Morgan Stanley, plus an $18 billion junior tranche led by Blackstone.
  • Oracle is in talks with Apollo and Goldman Sachs to finance a large chip purchase of its own, details undisclosed.

Add it up and the AI ecosystem has lined up on the order of $150 billion in chip debt in about a week. That’s before counting Oracle, which hasn’t put a number on its deal yet.

Why this matters

These numbers say a lot about how the AI race is actually being fought. It isn’t only about who trains the smartest model anymore. It’s about who can finance the hardware to run it.

For OpenAI, custom silicon is straightforward math. Owning the chips — rather than renting Nvidia GPUs indefinitely — lowers costs over time and reduces dependence on cloud providers. The financing Broadcom is arranging effectively lets OpenAI scale its hardware ambitions with borrowed money instead of raising another equity round, or at least with someone else’s money.

For investors, the sheer scale of these debt packages raises a harder question: what happens if AI revenue doesn’t grow fast enough to service them? OpenAI itself reportedly revised its 2026 revenue projection down by $20 billion in recent weeks, and Broadcom shares slipped about 2% in premarket trading the day after the financing report, per Invezz, as markets weighed whether financial conditions can keep supporting the buildout. Broadcom had edged up 0.19% the prior session, extending a four-session winning streak — a reminder that the stock had been running hot.

The spree also concentrates risk in a small club. Apollo and Blackstone keep showing up on these deals. If the AI boom keeps compounding, the lenders collect fees on history’s largest infrastructure bet. If it stalls, they own the collateral — chips that depreciate faster than almost anything else on earth.

That, more than any single financing round, is the real story here. The AI industry is building the future on borrowed money, and a handful of lenders are starting to own the foundation.

FAQ

Is Broadcom really raising $50 billion for OpenAI’s AI chips?

It’s in talks to do so, according to The Wall Street Journal. Apollo Global Management and Blackstone are among the lenders approached, but the talks are preliminary, nothing is signed, and the amount could still change.

Why does OpenAI need custom chips from Broadcom instead of just buying Nvidia GPUs?

Designing its own AI accelerators with Broadcom lets OpenAI cut costs and rely less on Nvidia and cloud providers. The financing helps OpenAI scale its hardware with borrowed money rather than diluting equity.

What other AI chip financing deals are happening right now?

SpaceX is reportedly seeking about $40 billion in debt to buy Nvidia GPUs, banks are syndicating a roughly $60 billion package for Anthropic’s chip purchases, and Oracle is talking to Apollo and Goldman Sachs about chip financing of its own.

Are investors worried about all this AI debt?

Some are. Broadcom shares slipped about 2% in premarket trading after the financing report, per Invezz, as markets weighed whether financial conditions can keep supporting the AI buildout. Others see the borrowing as demand validation for AI hardware.

Sources: The Wall Street Journal, Bloomberg, The Business Times, Invezz, particle.news