Delays and local opposition to data‑center projects are beginning to unsettle the financing chains that underwrite major AI deployments. Bloomberg reported the development this morning, and Semafor confirmed it: Oracle is seeking to postpone lease payments on a New Mexico data center that is part of the Stargate AI buildout unveiled last year during President Donald Trump’s administration.
Why is the project delayed?
State officials have repeatedly denied permits for the gas pipeline needed to power the site, setting back the project by at least six months. Because infrastructure schedules and financial arrangements are tightly coordinated for these developments, a single delay can trigger knock‑on effects across lenders, equity investors and insurers.
Scale of the financial exposure
Banks have provided $18 billion in loans for the New Mexico site. Above that sits $3 billion in equity from Blue Owl. The project’s viability depends heavily on Oracle paying rent fully and on schedule. The New Mexico project is also only one part of Oracle’s broader obligations: the company carries $288 billion in lease commitments, a sixfold increase from the start of 2025.
Risks and implications
The financing is structured as a complex patchwork of loans, insurance and milestone‑triggered payments, so timing mismatches can magnify losses for backers and insurers. Joe Peiser, CEO of risk capital at insurance broker Aon, warned that potential losses from such disruptions could be far larger than the market currently anticipates. While Oracle can defer the start of a lease, it still must cover the project’s carry costs in the interim, and pushing back the lease commencement also extends the lease’s end date — producing longer cash flows for investors while creating short‑term burdens for the lessee.
Market reaction has been tangible: the report notes that Oracle lost more than $20 billion in market value today.
Historical parallel and international perspective
The article draws a parallel to 2008, when disputes over collateral timing and insurer actions helped turn problems at a few overleveraged institutions into a global crisis. That episode underscores how insurers' and lenders' responses can amplify localized disruptions.
On a broader level, Kristalina Georgieva, managing director of the International Monetary Fund, told Semafor’s Next 3 Billion event that, so far, demand for AI‑related investments has not crowded out governments’ borrowing capacity, but the rapid expansion and complicated financing structures carry inherent risks.
What to watch next
- the outcome of the New Mexico permitting process and the status of the gas‑pipeline approvals;
- further announcements from Oracle about how it will manage lease obligations and project scheduling;
- responses from lenders and insurers if further delays or cost overruns occur.
This episode highlights that the rapid buildout of AI infrastructure often rests on fragile financial constructions that can be disrupted by local regulatory and infrastructure hurdles.
