Blackstone Plans to Raise Another $36 Billion to Bet on Anthropic! AI Computing Power "Securitization" Surges—Who Will Pay for Potential Credit Defaults?

According to Zhihui Finance APP, Wall Street private equity and alternative asset management giant Blackstone has begun preliminary discussions with investors to gauge market interest in a second massive AI-related debt financing plan, which will be used to fund AI leader Anthropic’s use of Alphabet-owned Google’s TPU compute chips.

Some media, citing sources familiar with the matter, reported that one of the initial plans aims to raise at least $36 billion in debt financing. The sources stated that specifics such as the scale of financing, deal structure, and whether Blackstone will ultimately lead the financing are all still under discussion and could change. Because these individuals were not authorized to comment publicly, they requested anonymity.

The primary impact of this large-scale AI financing is to first strengthen order certainty in the compute power supply chain, followed by concerns about an AI bubble burst and liquidity risks. Blackstone is discussing at least a second round of financing for at least $36 billion. Adding the $35 billion AI XPV initial deal completed by Apollo and Blackstone, the essence of these transactions is to turn Anthropic’s expected future revenue from Claude over the next few years into immediate orders for Google TPU AI compute clusters, optical communication/interconnect, data center CPUs, high-performance Ethernet switching infrastructure, data center HBM/DRAM/NAND memory components, and data center power chain and other massive AI compute infrastructure-related orders.

The first round of AI XPV platform financing corresponds to more than 1 GW of computing power and is planned to support deployments exceeding 20 GW before 2028. This means that chip orders are no longer entirely dependent on the current cash flow of AI labs but can be "capital marketized" through SPVs, long-term leases, and supplier credit support. Therefore, in the short term, Broadcom’s AI compute chain, Google’s TPU ecosystem, network equipment, power, and core data center hardware suppliers will see a significant extension in order visibility.

The risk is that financial engineering does not eliminate repayment risk; it merely shifts it from the AI lab’s balance sheet to SPV creditors, chip suppliers, cloud computing platforms, and their guarantors. This round of AI boom is not destined to break, but its core pricing factor is shifting from technology penetration to whether the credit chain will be supported by actual cash flows. Castle Securities expects that by 2028, AI chip financing debt could exceed $500 billion, making “credit leverage risk brewing” a reasonable warning phrase for investment analysts.

AI Financing Race Breaks Record Again: Blackstone Plans to Raise at Least $36 Billion, Anthropic Bets on Google TPU Chips for Compute Power Expansion

If this deal is ultimately completed at a scale close to the above, it would exceed the $35 billion debt financing arranged by Wall Street asset management giant Apollo Global Management and Blackstone for Anthropic’s rental of customized AI chips exclusively built by Google about two months ago. That financing was one of the largest private credit transactions in history.

Representatives from Blackstone, Apollo, Anthropic, and Google declined to comment.

As Silicon Valley races to construct AI infrastructure, leading companies have already entered into structurally complex, and often cyclical, financing deals to secure compute resources. Google is one of Anthropic’s earliest investors, having acquired its equity assets multiple times and now increasingly providing guarantee support for large-scale financing of the AI startup’s data center construction.

This potential new round of financing follows Anthropic’s secret submission of an initial public offering (IPO) application to U.S. stock markets. The company is seeking to go public ahead of its rival OpenAI. As the developer of the Claude AI model family, Anthropic plans, with the support of previous AI debt financing arrangements and Google’s assistance, to rent high-performance computing chips in five data centers.

Global technology companies are leveraging various parts of the credit market to meet the unprecedented capital demands of artificial intelligence, forcing Wall Street asset management giants to collectively participate in designing new debt structures to keep pace with industry expansion. Because there are market concerns that AI investments may not produce the expected positive returns, some companies have recently been forced to pay high benchmark yields when issuing new debt.

Broadcom, Apollo, and Blackstone established a cooperative platform earlier this year called the “AI XPV Platform,” aimed at helping leading AI technology development firms, including Anthropic, finance AI compute infrastructure. The $35 billion AI-related debt financing completed about two months ago was the platform’s first financing transaction.

In that transaction, Broadcom provided backstop financing support for the largest senior debt portion. Earlier media reported that Morgan Stanley served as Broadcom’s advisor and assisted in arranging the deal. Broadcom representatives did not respond to requests for comment.

Wall Street Begins “Securitizing Compute Power”: Anthropic and Meta Monetize a Portion of Future Demand Ahead of Time, AI Bull Market Enters Credit Leverage Stage

The completed $35 billion AI XPV financing supports Anthropic’s expansion of over 1 GW of compute power, while Blackstone’s second round of financing under discussion has an initial scale of at least $36 billion. The newly proposed $36 billion round plus the already completed $35 billion AI XPV initial funds are turning a portion of Anthropic’s strong future revenue expectations into current orders for TPUs and data centers. The platform’s first round supports over 1 GW of compute power and aims to advance the deployment of more than 20 GW of AI infrastructure globally.

All of this means chip orders are no longer entirely dependent on the current cash flow of AI labs but can be “capital marketized” by Wall Street asset management giants through SPVs, long-term leasing, and supplier credit support. Therefore, in the short term, Broadcom’s AI compute chain, Google’s TPU ecosystem, network equipment, power, and core data center hardware suppliers will see a significant extension in order visibility.

The risk is that financial engineering does not eliminate repayment risk; it merely transfers it from AI lab balance sheets to SPV creditors, chip suppliers, cloud platforms, and their guarantors. Anthropic’s financing structure involves an SPV buying Google TPUs and leasing them to Anthropic, with Broadcom providing deficit or residual value support for the largest senior debt tranche, enabling core debt of around $25 billion to be priced at about 5.75% yield; the subordinated debt without equivalent guarantees yields about 8.5%. This creates a typical “supplier financing closed loop”: chip vendors facilitate customer purchases through credit guarantees, which turn into orders, and eventually become revenue and valuation bases for the chip manufacturers. If Anthropic’s income growth, compute power utilization, or IPO fundraising falls short of expectations, risk will transmit backward along the chain of “lease payment—SPV debt servicing—chip residual value—supplier guarantee.”

Facebook’s parent company, Meta, shows that such off-balance-sheet capital commitments have risen from project-level arrangements to the core financing model of tech giants. By the end of June 2026, Meta’s committed but not yet commenced leases—which therefore do not yet count as lease liabilities on its balance sheet—have reached $278.99 billion, a significant increase over previous quarters. In July, the company signed another approximately $68 billion in data center leases with terms of 18 to 20 years, as well as $349.31 billion in irrevocable contract commitments and up to $14.72 billion in contingent cloud capacity purchase obligations.

Meanwhile, Meta has raised its 2026 capital expenditure guidance to $130–145 billion. Strictly speaking, these large AI financings and debt projects are not “hidden debt” in the traditional sense, but their economic substance is long-term fixed payment obligations: when the facilities are put into use, lease liabilities are gradually recognized on balance sheets, and cash flow pressure has already been contractually locked in.

Castle Securities forecasts that by 2028, new debt issuance solely for AI chip purchases could exceed $500 billion, equivalent to more than 5% of the size of the U.S. investment grade bond index at that time. In 2028 alone, new issuance could surpass $250 billion, with most debt terms only three to five years to match the relatively short economic lifespan of the chips. The real danger is not just “too much debt,” but that short-maturity debt, rapidly depreciating assets, and long-term, highly uncertain AI revenues create a maturity mismatch: if interest rates remain high, inference prices for models fall, chip architecture upgrades rapidly depreciate old equipment, or credit markets demand higher issuance premiums, a large number of projects could face a refinancing wall around 2028, squeezing technology, media, and telecom bond allocations, and triggering credit spread expansion and private credit liquidity contraction.

Therefore, according to some experienced Wall Street analysts, investment strategies should separate the valuation of “financed orders” from “terminal cash demand”: giving priority allocation to chip and infrastructure leaders with diversified customers, ample net cash, orders with prepayments or irrevocable commitments, and sustainable free cash flow generation. For projects relying on SPVs, supplier guarantees, circular investments, and single AI lab lease repayments, risk discounts must be increased, and focus placed on monitoring credit spreads, CDS, lease-adjusted leverage ratios, chip residual value, and compute utilization. Overall, this round of AI prosperity is not destined to break, but its core pricing driver is shifting from technology penetration to whether the credit chain is actually supported by cash flows.