Asset-backed financing secured by chips is becoming the core funding engine for global AI expansion. In 2023, Blackstone set a precedent by providing CoreWeave with $2.3 billion in private credit; today, major players such as Broadcom and NVIDIA have entered the space, driving financing volumes toward the trillion-dollar mark. By aligning cash flows from customer contracts with special purpose vehicles (SPVs) and incorporating chip manufacturer guarantees, this model enables low-cost financing. Despite its vast market potential, risks such as concentrated exposure, limited liquidity, and uncertainties around the reliability of customer contracts remain significant.Article author and source: Wall Street Journal
Chip-backed structured financing is becoming the core funding engine behind global AI expansion and is rapidly evolving into a new asset class with potential scale exceeding $1 trillion.
The market's雏形 began in 2023 with Blackstone’s $2.3 billion private credit facility to CoreWeave, marking the first landmark transaction in chip-backed debt.
Currently, Broadcom is negotiating with Wall Street institutions to arrange $60 billion in chip financing for Anthropic, SpaceX is reportedly seeking $40 billion in financing for NVIDIA chips, and NVIDIA itself has announced a $500 billion cooperation framework with six lending institutions.
JPMorgan analysts predict that total chip spending before 2030 could exceed $3 trillion, with most of it reliant on debt financing.
However, this expansion of debt also brings significant risks. Since most transactions are private in nature, investors lack convenient exit channels; cross-exposures among chip manufacturers, cloud giants, and emerging cloud service providers are highly concentrated.
CoreWeave's corporate junk bonds have recently plunged sharply, with some falling to 80 cents on the dollar and yields rising to around 15%, reflecting deep market skepticism about the commercial returns of AI.
A financing round that pioneered a new paradigm: The revolution began with Blackstone’s Tactical Opportunities team carefully evaluating an unfamiliar concept. According to Brannin McBee, co-founder of CoreWeave, who was involved in the due diligence at the time, Blackstone executives had directly asked: “What is a GPU?” and “Is AI real?”
After achieving a satisfactory outcome, Blackstone led a $2.3 billion private credit facility for CoreWeave in 2023, setting a precedent for chip-backed debt financing.
The core logic of this structure is to establish a special-purpose vehicle (SPV) to purchase and hold the chips, with debt repayments funded by cash flows generated from target customer contracts. The debt term is aligned with the chips’ five- to six-year depreciation cycle, thereby locking in hardware depreciation risk within a predictable range.
CoreWeave’s model is derived from commodity market experience—similar to the development approach in the liquefied natural gas sector, where contracts are signed first, followed by financing and construction. The initial transaction was secured against Microsoft’s contract, at a time when financing costs reached double-digit interest rates; McBee admitted this was “the most challenging due diligence of my career.”
As CoreWeave’s client base continues to expand, its cost of financing has decreased. In March this year, the company completed an $8.5 billion debt financing backed by its contract with Meta, reducing its borrowing rate to approximately 6%—with Wall Street pricing the risk based on Meta’s high credit rating rather than CoreWeave’s junk rating.
The cost of chips far exceeds what data centers can comprehend. To understand the scale of this financing revolution, one must first grasp the magnitude of chip expenditures. Chip costs are two to four times higher than the data centers that house them, yet they depreciate completely within five to six years, creating a massive financing gap that traditional capital markets struggle to absorb on their own.
Using Nscale’s flagship project in West Virginia as an example, the accompanying power plant is expected to cost $6.5 billion, the data center and equipment $18 billion, and GPU procurement costs as high as $47 billion.
Alphabet’s Google, Microsoft, Amazon, Meta, and other investment-grade cloud giants can cover the above hardware costs using their ample cash reserves or low-cost bond financing, but OpenAI, Anthropic, and numerous emerging cloud providers lack such balance sheet strength.
OpenAI’s Chief Financial Officer, Mike Liberatore, admitted: "The scale of infrastructure construction is unprecedented. No playbook exists for how to reach artificial general intelligence from scratch using traditional methods under conditions of extreme capital intensity."
To this end, the industry has adopted mechanisms from project finance and structured finance: isolating the risk of each transaction through an SPV, raising debt based on the quality of the transaction and the borrower’s creditworthiness rather than the issuer’s credit rating, often without recording it on the company’s balance sheet.
The direct involvement and endorsement by chip manufacturers, along with the deep engagement of Broadcom and NVIDIA, have propelled this market to a new scale. Both companies have provided "backstop" commitments for the related debt transactions:
In the event of a borrower default, the chips will be re-leased or sold to repay the outstanding debt, with the chip manufacturer responsible for covering any shortfall. This mechanism allows the senior bonds issued by the SPV to be linked to the chip manufacturer’s investment-grade rating, significantly reducing financing costs.
Apollo and Blackstone recently completed a $35 billion "Big Sky" transaction to support Anthropic. This is the first closed deal on the "AI XPV" platform jointly established by Broadcom, Apollo, and Blackstone—U.S. Bank strategists estimate that if the platform achieves a financing capacity of 20 gigawatts, it may require $235 billion in backed debt by mid-2029.
In this Anthropic transaction, the two senior tranches backed by Broadcom received investment-grade ratings, with the highest tranche priced at 5.75%; the riskier $4.5 billion junior tranche was priced at 8.5%—still significantly below Anthropic’s standalone financing costs as an unrated private company.
In August this year, NVIDIA brought together Apollo, Blackstone, BlackRock, Brookfield, KKR, and Goldman Sachs to provide financing for AI infrastructure, with NVIDIA itself committing to guarantee up to 25% of the total amount.
To achieve true scale from private placements to the syndicated market, chip financing must transcend the boundaries of the private market and enter syndicated loan and even investment-grade public bond markets—among the largest and lowest-cost capital pools globally.
CoreWeave successfully completed the first chip-backed syndicated leveraged loan this year, with both transactions receiving strong institutional investor demand and trading near par in the market—standing in stark contrast to the company’s corporate bonds.
Rich Myers, Global Head of Structured Products and Private Capital Markets at Morgan Stanley, said: "These financings are directly tied to the most valuable assets in the market—the computing power generated by chips. Overwhelming evidence indicates that demand for computing power will continue to grow and outstrip supply."
Chris Edson, Head of Global Asset Allocation at Apollo, stated that the traditional public debt market alone cannot meet the financing needs of digital infrastructure, requiring participation from various sources of capital such as project financing, private placements, insurance funds, and commercial real estate. Obtaining an investment-grade rating is the key to unlocking access to the largest capital pools, including insurance funds.
Risk accumulation cannot be ignored, as risks are simultaneously building alongside the accelerated inflow of capital.
Fitch Ratings Senior Director Jason Pompeii noted that current vendor financing levels are unprecedented and risks are highly concentrated. "This only increases the likelihood that if AI ultimately disappoints, the entire supply chain will be impacted."
The private nature of the transaction structure raises liquidity concerns. Since most chip financing deals are not traded on public markets, investors face difficulties in exiting conveniently, while the transactions deeply link a small number of chip manufacturers, cloud giants, and emerging cloud service providers, resulting in highly concentrated cross-exposures. Dave Snyderman, Managing Director of Magnetar’s Credit and Fixed Income Alternatives business, warned: "The giants will only get bigger. Only they have access to financing channels."
The legal enforceability of customer contracts is also in question. Akhil Bansal, Head of Asset-Backed Financing at Carlyle, bluntly stated: "Everyone assumes these contracts are airtight, but I’m focused on the practical reality—when only seven institutions provide all the credit support for GPUs, who really holds the cards?"
Laura Parrott, Head of Private Fixed Income at Nuveen, said she is currently holding off on participating in large investment-grade chip deals until the market structure becomes more mature. "This is a completely new structure, and we don't yet know how it would evolve if faced with some kind of shock."
Additionally, macro-level variables persist: local opposition in Texas has led to a moratorium on new projects, Oracle’s project in New Mexico has been delayed due to regulatory hurdles, and OpenAI and Anthropic are discussing slowing down deployment due to safety concerns.
Moving toward the "computing power asset class," market participants remain optimistic about its long-term potential, despite existing challenges.
John Watson, Senior Managing Director at Blackstone Tactical Opportunities, compares the current moment to the early days of the automotive industry: just as those who financed the Ford Model T faced doubts about whether people would really drive cars, today automotive finance has grown into a vast and highly mature market.
On the evolution of the chip market, early research and real-world usage data indicate that the actual useful life of chips is exceeding the periods currently reflected in accounting practices, thereby improving their financiability. Karen Fang, Global Head of Banking Infrastructure and Sustainable Finance at Bank of America, expects that semiconductor valuation tools, similar to used car price guides, will gradually emerge, fostering a more active chip leasing and hedging market.
Watson’s ultimate vision is to transform a unit of computing power into an asset class that can be traded and financed globally—“We understand the economic value of computing power, and therefore we finance it.”
Achieving this goal still depends on the actual realization of market expectations for AI commercialization.
Risk Disclaimer and Liability Waiver
The market carries risks; invest with caution. This article does not constitute personal investment advice and has not taken into account individual users’ specific investment objectives, financial situations, or needs. Users should consider whether any opinions, views, or conclusions presented herein are suitable for their particular circumstances. Investment decisions based on this content are made at your own risk.