Broadcom seeks $60 billion debt for Anthropic AI chip financing

Broadcom is negotiating a debt package that could reach $100 billion to fund AI chip purchases for Anthropic, one of the largest financing deals in the AI infrastructure buildout.

Broadcom is negotiating with lenders to raise over $60 billion in debt for an AI chip financing deal benefiting Anthropic, with the total package potentially reaching $100 billion.

The financing structure is still being finalized, according to people familiar with the matter. The deal may include about $30 billion in subordinated debt, with Broadcom guaranteeing part of a senior secured tranche estimated at $60 billion to $70 billion.

The proposed financing reflects the capital intensity of the AI compute buildout. Broadcom's custom XPU processors, designed for hyperscale AI workloads, have become a key revenue driver. The company's software segment is expected to grow revenue 30 percent or more in its upcoming Q3 report due Sept. 2, according to a Seeking Alpha analysis.

The deal would mark one of the largest AI-related debt financings to date, potentially reshaping how AI infrastructure is funded. For Broadcom, the financing could strengthen its position in the AI chip market against Nvidia, which dominates the sector with its GPU lineup. Broadcom shares have lagged the broader semiconductor sector year-to-date despite an 80 percent-plus total return last year.

Deal Structure Points to Growing AI Capital Intensity

The proposed financing would be structured across multiple tranches. The senior secured portion, estimated at $60 billion to $70 billion, would carry a Broadcom guarantee. A separate subordinated tranche of about $30 billion would bring the total package to as much as $100 billion, the people said.

The scale of the deal reflects the enormous capital requirements of AI infrastructure. Anthropic, the AI research company backed by Amazon and Google, has committed to massive compute purchases to train and deploy its frontier models. The financing would help fund those commitments through debt rather than equity, a structure that has become increasingly common in the AI sector as companies seek to avoid diluting existing shareholders.

This debt-heavy approach mirrors a broader trend in AI infrastructure financing. Hyperscalers and AI labs are increasingly turning to structured debt products to fund GPU and custom chip purchases, spreading the cost of compute over multi-year periods. The arrangement allows AI companies to secure the compute capacity they need while preserving equity value.

Broadcom's AI Bet Faces Scrutiny

Broadcom's stock has faced pressure as investors weigh the company's AI financing exposure. The company's AI boom faces a $370 billion financing question, according to a recent MSN report, as the scale of debt needed to fund AI chip purchases grows.

Despite the concerns, some analysts see opportunity. A Seeking Alpha analyst raised their 12-month price target on Broadcom from $550 to $600, citing the underestimated growth trajectory of the XPU business and the company's strong free cash flow profile. The analyst reiterated a Strong Buy rating.

Broadcom has built a wide moat around its businesses, with a very strong free cash flow profile and a track record as one of the best dividend growth stocks in the S&P 500, the analyst noted. The company's XPU business, which designs custom AI accelerators for hyperscale customers, has been a key growth driver, though the market has been slow to fully price in its trajectory.

The financing deal, if completed, would show strong capital market appetite for AI-related debt and could benefit the broader AI and semiconductor sector. For investors, the key question is whether Broadcom's XPU growth can justify the leverage being taken on to fund the AI buildout. Nvidia, trading at a premium to Broadcom on forward earnings, remains the benchmark against which Broadcom's AI chip business is measured. If the financing closes successfully, it could set a precedent for how future AI compute deals are structured, potentially benefiting the entire semiconductor supply chain from TSMC to memory makers.

This article is for informational purposes only and does not constitute investment advice.