Bernstein Statistics: Only About 35% of the 488GW of U.S. Data Center Plans Is Likely to Materialize

Recently, the well-known Wall Street research firm Bernstein conducted a study, and the results were quite surprising.

The cumulative planned data center capacity that has been announced in the United States is as high as 488GW. But only about 170GW can be confidently expected to materialize, accounting for just 35%.

In other words, for nearly two-thirds of the data centers planned in the United States, whether they can be built in the future remains unknown.

What is even more interesting is that the reasons these projects cannot get built are quite varied.

Some developers treat data centers directly as real estate: before the project has even broken ground, the land is already listed for sale; others manage to start construction, but the client keeps changing the design plans, pushing the construction budget up by nearly $800 million.

In the past, these problems were not a big deal. After all, AI is the hottest investment theme right now. As long as you dare to announce a few GW of data center plans, Wall Street is willing to assign a high valuation, and banks and bond investors are willing to put up money.

But now, the situation is starting to change.

On one hand, the Federal Reserve raised interest rates again in September, tightening the financing environment; on the other hand, more and more data center projects are exposing problems such as construction delays and cost overruns, prompting investors to re-examine the risks involved.

Clearly, even Wall Street is no longer buying the data center pie.

Two-thirds of US data centers are in doubt

Why are two-thirds of US data centers in doubt? The problem mainly lies in two aspects.

One is on the planning side: more and more hyperscale data centers are falling into the hands of novices with no construction experience.

BloombergNEF statistics from July show that among the 100 largest data center plans in the United States, 52 come from developers with no construction experience, totaling 87GW.

Why are these new players so bold?

The reason is simple: in the US data center business, you do not necessarily have to build the server rooms to make money.

For some land developers, as long as they acquire land in advance, push approvals forward, and secure power connections, they have a chance to package the land as a data center project and sell it at a high price to the real operators.

The La Osa project in Arizona is a typical case.

In July 2025, Vermaland, a land developer with no data center construction track record, directly announced a $33 billion megaproject, planning 59 buildings and 3GW of power demand.

How exaggerated is this scale?

As of the end of June this year, the 51 operating data centers under CoreWeave had a combined active power capacity of only 1.5GW.

That is to say, on its first foray into data centers, Vermaland planned power capacity equivalent to twice CoreWeave's existing operating scale.

But what is more interesting is that in May this year, a Vermaland representative admitted at a government hearing that the project land had already been listed for sale and was intended to be sold to the end user.

The previously announced $33 billion megaproject instead became the best advertisement for selling the land. On the LoopNet land sales page, Vermaland even used "poised to become the largest data center in the United States" as a selling point.

For this type of developer, the bigger the plan, the better the chance the land can be sold at a high price. As for whether it can actually be built in the future, that is something the buyer needs to worry about.

This phenomenon is becoming more and more common, and the industry has even coined a sarcastic term for it: Bragawatts, that is, "gigawatts that are bragged about."

But inflated plans are only one aspect; what is more troublesome is that even if construction actually starts, delivery may not happen on time.

Research by Goldman Sachs in May this year showed that the historical on-time delivery rate for US data centers is about 72%, but the expected on-time delivery ratio for planned capacity one year and two years out may fall to 60% and 50%, respectively.

Data centers require a large amount of supporting infrastructure such as power, substations, and cooling systems, and if any one link falls behind, it can slow down delivery progress.

Moreover, even after solving these problems, developers still have to face ever-expanding construction costs.

The Barber Lake project in Texas is a typical case.

This project is being developed by Cipher Digital, which transformed from a Bitcoin mining company. It has a planned total power supply capacity of 300MW, has already signed a long-term lease with AI computing company Fluidstack, and Google has provided partial guarantees.

The client is there, the contract is signed; by all accounts, this should be a high-quality project.

But during construction, the client kept raising new design requirements, causing the engineering plan to be adjusted repeatedly, and the lease commencement time for some server rooms was also postponed to the first quarter of 2027.

What is more troublesome is that the project's expected total completion cost has risen from $2.157 billion to $2.941 billion–$2.966 billion.

Before completion, the budget has already increased by nearly $800 million, an increase of more than 36%.

Yet the extra money cannot all be borne by the client.

According to the agreement, the first approximately $359 million over the original budget is borne by Cipher, and for the portion above that threshold, the client only shares 50%, to be gradually compensated through additional rent over the next 20 years.

This is quite awkward. The original idea was to lock in revenue through a 20-year long-term lease, but when it actually came to execution, before the rent could be collected, the project was bogged down by one construction payment after another.

Ultimately, the problem with US data centers is that the pace of expansion driven by the capital markets has far exceeded the real world's construction capacity.

Wall Street is also starting to recalculate

The most direct manifestation is the decline in stock prices.

According to Kpex.AI statistics, from October 6 to 20, eight AI infrastructure companies including CoreWeave, Nebius, and IREN saw their combined market value evaporate by approximately $51 billion, with a median stock price decline of 28%.

Among them, CoreWeave fell nearly 45% at one point. According to Reuters, short sellers reaped about $1.7 billion in paper profits from this company alone.

This shift has already spread to the IPO market. In October this year, Firmus, an Australian AI data center company backed by Nvidia, originally planned to raise about $5 billion in an IPO, but some investors withdrew their subscription intentions before the listing. On October 9, Firmus ultimately canceled its IPO and turned to private market financing.

An important reason was that its partner CDC Data Centres confirmed that the 1.6GW data center project the two sides had originally planned to jointly develop would no longer proceed.

SB Energy, backed by SoftBank, encountered a similar problem. The company had originally planned to move forward with an IPO in September, but later postponed it to the end of the year or even next year.

Beyond the stock market, the bond market is also changing. According to Trepp, a commercial real estate data agency, as of October 23, among the 64 data center-related mortgage-backed securities issued by the end of last year, 60 had seen their credit spreads widen.

Among them, the credit spreads on data center bonds with the highest AAA rating widened by about 0.6 percentage points during the year, while other comparable commercial real estate bonds widened by only about 0.01 percentage points. Simply put, investors are beginning to demand higher risk compensation.

What is more troublesome is that new bond issuance is also starting to cool. Research by Morgan Stanley in October shows that at the beginning of this year, the issuance yields on data center asset-backed securities with the highest AAA rating were still below 5%, but now they are close to 7%.

The experience of US data center developer EdgeCore is very representative. In October this year, the company prepared to issue $521 million in asset-backed securities, but nearly canceled the transaction because of insufficient investor subscriptions.

To complete the financing, EdgeCore had to raise its pricing, increasing the credit spread on the highest AAA-rated bonds from an initial more than 100 basis points to 200 basis points, and ultimately completed the issuance with a coupon of 5.90%.

And just last year, the $235 million in bonds issued by EdgeCore carried a coupon of only 4.45% and were oversubscribed.

In just one year, data center financing has gone from investors rushing to buy to needing higher interest rates to find buyers.

For the data center industry, which is highly dependent on debt to expand, once this negative feedback loop forms, it may be more fatal than a short-term slowdown in demand.

This article is from the WeChat public account "Silicon-Based Observation Pro", author: Silicon-Based Jun, published by 36Kr with authorization.