Waller said 50 basis point hike should be considered if core CPI hit 0.3%; ECB warns on AI valuations

By
CTOL Staff Reporter
1 min read

We expect leveraged AI suppliers to feel the squeeze while Alphabet, Amazon and Meta keep spending; our own test is at least $98.874 billion in combined property and equipment purchases on Forms 10-Q due 10 May 2027.

CoreWeave's $2.75 billion of notes due in 2031 carry a 9.75 percent coupon, an expensive starting point for expanding AI infrastructure. On 10 November 2026, Federal Reserve Governor Christopher Waller argued that a 50 basis point hike deserved consideration if October core inflation stayed around 0.3 percent or higher. The October reading was 0.3 percent. Waller's speech favoured 25 basis points as his base case. Suppliers must finance tomorrow's AI capacity at today's borrowing costs.

We would favour the cash generating buyers of AI infrastructure over leveraged capacity suppliers. We expect Alphabet, Amazon and Meta to buy at least $98.874 billion of property and equipment in January to March 2027, matching their year earlier total. The Funding Split is the difference between deciding whether a project earns enough and finding creditors willing to finance it. Both face rising rates, but a buyer can reduce investment by choice while a leveraged supplier faces a lender's decision too.

The Fed's 28 October decision left its policy range at 3¾ to 4 percent and described capital investment as robust. Waller cited September core PCE inflation of 3.1 percent, alongside price increases above 3 percent across more than 70 percent of core service components. The ten year Treasury yield rose from 4.91 percent on 2 November 2026 to 5.10 percent on 17 November 2026. Ahead of the 8 to 9 December meeting, prospective AI profits already faced a higher discount rate.

In the nine months to 30 September 2026, Alphabet generated $127.149 billion in operating cash and spent $131.953 billion on property and equipment. Amazon generated $113.841 billion and spent $175.670 billion; Meta generated $96.873 billion and spent $83.055 billion. Together that is $337.863 billion of operating cash against $390.678 billion of purchases, a $52.815 billion gap. Even strong customers must choose which commitments justify that cash strain.

Long term debt at those three companies grew from $170.939 billion at the end of 2025 to $379.149 billion at 30 September 2026, an increase of $208.210 billion. In a scenario where every dollar of that debt repriced immediately after a 50 basis point rise, $379.149 billion times 0.005 adds $1.896 billion of annual interest. That is 0.56 percent of nine month operating cash flow and an upper bound, because it assumes every dollar reprices at once. New projects must earn a higher return, even when existing debt is affordable.

At CoreWeave, $5.638 billion of nine month operating cash flow sat beside $23.558 billion of capital spending. Its debt carried a value of $47.505 billion at 30 September 2026, equivalent to 8.4 times that nine month cash flow; the buyers' $379.149 billion over $337.863 billion is 1.1 times. CoreWeave also issued $1.25 billion of 2032 notes paying 9.625 percent and arranged a $2.6 billion facility at Term SOFR plus 5.50 percent. That is expensive money for additional capacity.

CoreWeave's sensitivity at 30 September 2026 was $41 million of quarterly interest per 100 basis points on floating debt. The 50 basis point scenario is $41 million times four times 0.5, or $82 million annually. CoreWeave spent $1.754 billion on debt interest in nine months. Fresh financing terms matter more than that immediate adjustment. Microsoft, meanwhile, spent $39.367 billion on property and equipment in its September 2026 quarter while total debt declined from $40.294 billion to $35.481 billion. Microsoft added equipment while reducing its debt.

The ECB's Financial Stability Review, published on 18 November 2026, gives the counterargument its full force: "free cash flows turned negative for most US hyperscalers in 2026, as capital expenditure on AI infrastructure exceeded operating cash flows." It projects $5.5 trillion in cumulative US big tech capital spending by 2030. Higher yields discount distant earnings more harshly, while cheaper AI models threaten the revenues assumed to justify construction. A solvent buyer can cancel a project before any supplier loses access to credit. Under that mechanism, January to March 2027 purchases could fall below $98.874 billion even with healthy operating cash.

Higher yields make new AI projects harder to justify, yet these buyers increased quarterly equipment purchases in 2026. The three buyers purchased $98.874 billion of property and equipment in January to March 2026 and $162.556 billion in July to September 2026, comprising $51.355 billion at Alphabet, $77.259 billion at Amazon and $33.942 billion at Meta in the later quarter. Those purchases accelerated during the same nine months that produced the $52.815 billion cash gap. Even with pressure to reconsider returns, the buyers chose to expand. The recent expansion of that same spending line favours our March 2027 threshold over an immediate collapse in buyer investment.

The previous tightening cycle complicates that bet. The Fed moved its target range from 4¼ to 4½ percent in December 2022 to 5¼ to 5½ percent in July 2023. Amazon cut property and equipment purchases from $63.645 billion to $52.729 billion, a 17.2 percent decline, while operating cash flow rose from $46.752 billion to $84.946 billion. Meta cut purchases from $31.186 billion to $27.045 billion, down 13.3 percent. Alphabet increased purchases from $31.485 billion to $32.251 billion, up 2.4 percent. Logistics normalization and cost efficiency influenced those cuts, beyond the rate changes. Ample cash can still accompany a shrinking investment budget.

Fed Governor Michael Barr offered a limit to the alarm on 16 November 2026: "Given the still-limited amount of debt backing the current boom, I'm less concerned about direct losses." He called the debt backing the boom limited, even as the seven largest AI firms' debt is expected to exceed $800 billion before year end. The Fed's stability report nevertheless puts advanced computing related debt issuance above $680 billion by October 2026 and outstanding public debt above $1.1 trillion. The ECB worries about equity losses reaching bonds and sovereign borrowing costs; Barr highlights hedge fund leverage and fire sales. Both channels threaten financiers while buyers keep investing.

The Fed estimates more than $2 trillion of undiscounted take or pay commitments and residual value guarantees around AI infrastructure. Those obligations are contractual exposure rather than cash. They make promised customer payments valuable to lenders while leaving buyers exposed if AI returns disappoint. Customers that support borrowing gain bargaining power over terms. The Funding Split lets construction persist as financing costs rise and risks migrate through contracts. Spending resilience says little about each company's investment returns.

We favour cash generating AI buyers and demand a higher return for owning leveraged capacity suppliers, since spending can continue on punitive financing terms. We would be wrong if the Alphabet, Amazon and Meta Forms 10-Q for January to March 2027, due 10 May 2027, show combined property and equipment purchases below $98.874 billion.

You May Also Like

This article is submitted by our user under the News Submission Rules and Guidelines. The cover photo is computer generated art for illustrative purposes only; not indicative of factual content. If you believe this article infringes upon copyright rights, please do not hesitate to report it by sending an email to us. Your vigilance and cooperation are invaluable in helping us maintain a respectful and legally compliant community.

Subscribe to our Newsletter

Get the latest in enterprise business and tech with exclusive peeks at our new offerings

We use cookies on our website to enable certain functions, to provide more relevant information to you and to optimize your experience on our website. Further information can be found in our Privacy Policy and our Terms of Service . Mandatory information can be found in the legal notice