The AI data center financing market is undergoing a notable credit tightening. Bond investors are demanding larger discounts and more favorable terms, some major banks have become more selective about project loans, and combined with macro interest rate pressure and the AI industry's own risks, the cost of financing this construction boom is rising rapidly.

According to the tech media outlet The Information, the latest landmark case comes from CleanSpark. The bitcoin miner, which is developing data centers for Meta Platforms, issued $2.3 billion in bonds earlier this month at a price of 98.5 cents — one of the largest discounts of the past year. The 7.875% coupon is already elevated, and the issuer also pledged installment repayment of principal to reduce refinancing risk. According to Morgan Stanley data, all four high-yield data center bond deals since July were sold at some form of discount, whereas of the previous ten deals over the prior 12 months, only three included a discount.

The bank loan market is also showing signs of contraction. According to reports, people familiar with the matter said that institutions including Societe Generale, Sumitomo Mitsui Banking Corporation, and Mitsubishi Bank have become more cautious about data center project loans. At the same time, Oracle's New Mexico project recently issued a force majeure notice, prompting lenders to scrutinize contract terms and loan covenants even more strictly. These developments together point to the same risk: the financing chain for AI infrastructure construction is under pressure, and the expansion plans of some projects face a substantive threat.

Discounts Become the Norm, Investor Appetite Is Not What It Used to Be

The high-yield bond market has already sold about $55 billion in AI-related bonds this year, a massive supply, while investors' risk appetite is quietly shifting.

Connor Minnaar, a fixed income portfolio manager at Manulife Investment Management, said discounts are "the new development in the current market." He noted that "at the start of the year, investors were much more accepting of deal structures," and that this tolerance has now clearly narrowed.

The special feature of the CleanSpark case is that its end user, Meta Platforms, is an investment-grade company, and this kind of endorsement used to be enough to give investors ample confidence. But this time investors still demanded extra compensation, reflecting that market concerns about construction timeline risk and project execution risk have surpassed reliance on tenant credit.

In mid-August, another deal also confirmed this trend. Zenith Arc LLC, a developer backed by a venture capital institution under Coatue Management and infrastructure startup Fluidstack, issued bonds at a discount price of 99.5 cents to finance a data center that will be leased to trading giant Jane Street. After the bonds began trading, yields rose further, with investors demanding compensation even beyond that of older bonds with equivalent ratings.

Bank Loan Market: Syndicates Narrow, Oracle Incident Sounds the Alarm

The tightening in the bank loan market also cannot be ignored. According to two people familiar with the matter, Societe Generale, Sumitomo Mitsui Banking Corporation, and Mitsubishi Bank have all become more cautious about participating in data center project loans. All three banks had previously been deeply involved in landmark AI infrastructure financings: Societe Generale led the $7.1 billion debt financing for the first site of the OpenAI and Oracle "Stargate" project; Mitsubishi and JPMorgan jointly led $38 billion in financing for two other Oracle projects; and Sumitomo Mitsui participated in leading the $18 billion financing for Oracle's New Mexico project.

The recent setbacks at Oracle's New Mexico project became a warning for lenders. Last week, Oracle issued a force majeure notice to the project developer, a company under Blue Owl Capital, because the project encountered power supply delays. Oracle sought to invoke contract terms to exempt or delay its contractual obligations on the grounds of events beyond its control.

According to reports, people familiar with the matter said that from the lender's perspective, the loan was "well structured," and even if the project failed to secure power supply, Oracle could not use that to exempt itself from lease payment obligations. But one banker said Oracle's force majeure claim could prompt some banks to seek stricter protective clauses or reassess loan risk, further compressing potential loan supply.

The Construction Boom Continues, but the Financing Window Is Narrowing

Market participants stressed that no major deal has been forced to withdraw so far, and no active syndication process has been halted because of weak pricing. Project developers are still absorbing higher financing costs in the high-yield bond market to keep projects moving forward.

Connor Minnaar said: "This is still a story about concession terms. For many companies, this is a race to bring as much capacity online as quickly as possible, and the level of financing cost is still secondary."

However, the normalization of discounts, narrowing bank syndicates, and an unstable IPO window — these three pressures combined — are reshaping the landscape of the AI data center financing market. For projects with lower ratings, limited development experience, or longer construction cycles, financing difficulty has risen substantially, and whether the AI construction boom can maintain its previous pace of expansion will depend increasingly on the financing market's capacity to absorb it.

The "Crowding Out" Effect of Hyperscale Tech Companies Intensifies Market Pressure

The pressure in the AI financing market does not come from a single source, but from multiple factors combined.

Hyperscale tech companies such as Amazon, Google, and Microsoft are expected to spend a combined total of about $700 billion in capital expenditures this year, and are expected to maintain that level for years to come. These companies have already issued nearly $160 billion in investment-grade bonds this year, flooding the market with large-scale supply.

The extra yield investors demand to hold hyperscale tech company bonds has risen by about 0.25 percentage points this year, while spreads in the broader investment-grade market widened by only 0.04 percentage points over the same period. Although these tech companies have enormous operating cash flow, the relative cost of their bonds has clearly risen.

At the same time, some large tech companies are shifting spending to other financing entities, which must raise funds on their own in the high-yield bond market; developers building data centers for AI companies such as Anthropic and OpenAI are also competing for financing in the same market, further worsening the supply-demand imbalance.

Macro Rate Shock and AI Risk Form a Resonance

The rise in Treasury yields adds another variable to an already strained market. Federal Reserve Chair Warsh said last week that large-scale tech bond issuance has a "crowding out effect" on investors and is one of the reasons pushing Treasury yields higher.

For data center developers, rising Treasury yields mean riskier projects must offer higher compensation to attract capital.

According to reports, one banker involved in related deals described this shift as an "upward migration" in credit ratings and project quality — investors who were once willing to take on risk for projects with long construction cycles and inexperienced developers in exchange for yield may now instead choose to obtain similar returns from safer bonds.

The financing costs of some projects have already approached the break-even threshold. If a project requires a 12% return while borrowing costs are close to that level, the profit margin will be compressed to the point where the project becomes financially unsustainable.

IPO Market Turmoil Puts Another Financing Channel Under Pressure

Volatility in the Treasury market has also spilled over into the IPO market, clouding the outlook for another financing channel for AI companies. According to reports, Anthropic posted a net loss of $42 billion last year, and its draft IPO filing shows computing and infrastructure spending commitments exceeding $500 billion, reflecting extremely large financing needs.

SB Energy and Nscale, which are developing data center capacity for OpenAI and Anthropic, have recently publicly filed IPO applications, but the timing of their listings remains unclear. Anthropic had previously expected to disclose its IPO application as early as this month, but so far it has not acted.