Broadcom is in talks with lenders to raise up to $100bn in debt to finance AI chip infrastructure for Anthropic and other artificial intelligence companies, according to Bloomberg. The deal underscores the extraordinary capital requirements of the AI buildout - and the growing financial engineering behind it.
The company is reportedly discussing a financing package of more than $60bn with a group of investors, with Blackstone and Apollo Global Management already involved in an initial €35bn agreement with the group. The structure could include about $30bn of junior debt, with Broadcom guaranteeing part of a secured senior tranche valued at between $60bn and $70bn.
Broadcom designs custom AI chips for major technology companies, a position that has pushed its market capitalization to $1,732bn, ranking it seventh in the US. The initial commitment is expected to add 1 gigawatt of computing capacity using Broadcom's custom chips and networking solutions. The broader partnership aims to provide more than 20 gigawatts of computing power to leading AI labs by 2028.
The deal structure
The new debt could be issued through a securitization vehicle and follow a structure similar to the June financing, according to Bloomberg. That earlier deal - the €35bn agreement involving Blackstone and Apollo - appears to be the template for this larger round.
The scale of the financing reflects how AI infrastructure has become a capital-intensive business that now rivals traditional utilities and data center operators. The ability to secure debt at this size also signals that lenders view AI computing capacity as an asset with predictable, long-term revenue potential.
For finance professionals tracking the AI trade, the deal raises two questions: how much debt the AI ecosystem can absorb, and whether the returns on AI infrastructure will justify the borrowing. The interdependence of chip makers, cloud providers, and AI labs means a downturn in any one could ripple through them all.
Why this matters for finance
The Broadcom deal is a useful case study in how AI infrastructure is being financed - and where the risks sit. The structure of securitized debt backed by computing capacity is relatively new, and its performance in a downturn is untested.
Finance teams should watch how these deals are priced and guaranteed, since they set a precedent for future AI infrastructure financing. The concentration of risk among a small group of lenders and chip makers also deserves scrutiny. AI Learning Path for CFOs covers the financial mechanics of AI investments like these, and AI for Finance tracks how these deals are reshaping corporate balance sheets.
Your membership also unlocks:
