Data Center Ghost Demand Changes Which Equipment Bottlenecks Matter

The grid queue is not an equipment order

This distinction matters because infrastructure demand travels through several stages before it becomes supplier revenue.

A developer can reserve land and request hundreds of megawatts without having a tenant, financing, permits or firm equipment orders. Reuters reported that Pennsylvania had more than 100 proposed data centers, but only 20 had applied for required permits, and many proposals had not secured either power or a customer.[1]

An interconnection request therefore tells investors that someone wants optionality. It does not establish a booked relationship between a data center developer and Eaton, Siemens Energy, GE Vernova, Schneider Electric, Vertiv, Cummins or any other equipment supplier.

Actual procurement is a much stronger signal. Once utilities and developers place transformer orders, reserve switchgear production slots, sign long-term supply agreements or prepay for equipment, the relationship has moved from speculative infrastructure demand into contracted economic demand.

That is the more useful way to map this boom. A 500 MW project in a utility queue can disappear without producing meaningful supplier revenue. A transformer already ordered for a funded substation is a different exposure entirely.

The hardest bottlenecks sit closest to the grid

The clearest evidence of real scarcity remains in large electrical equipment.

Generator step-up transformer lead times exceeded 160 weeks in the first quarter of 2026, according to data cited by Reuters. High-voltage circuit breaker lead times reached 125 weeks in the second half of 2025. Utilities are responding by ordering equipment years earlier, signing longer supply agreements and sourcing internationally. One California utility told Reuters it was buying equipment as much as five years ahead.[2]

Those lead times cannot be explained by speculative AI projects alone. Transformers, high-voltage breakers and switchgear also serve utilities, renewable generation, transmission projects, factories and broader electrification. Data centers are adding pressure to a supply chain that was already difficult to expand.

That makes companies exposed to the grid interface fundamentally different from companies whose growth assumptions require every announced data center campus to be completed.

Eaton provides a useful public example. In its second-quarter 2026 results, the company reported that twelve-month rolling orders in Electrical Americas were up 41%, while total Electrical sector backlog was up 43% from a year earlier. Management also emphasized broad end-market strength rather than attributing the entire expansion to data centers.[3]

Backlog is not risk-free. Orders can be delayed or cancelled. But it is economically closer to realized demand than an anonymous request for hundreds of megawatts sitting in a utility planning queue.

For investors, that difference is critical.

Ghost demand becomes more important downstream

The further equipment moves toward the individual data center, the more project cancellation can matter.

Medium-voltage distribution systems, UPS equipment, busway, rack power, chillers, liquid-cooling systems, backup generators and prefabricated infrastructure ultimately require physical buildings that proceed to construction. Vertiv, Schneider Electric, Eaton, Caterpillar, Cummins and Generac all participate in different portions of this layer.

Demand for these products can still be extremely strong. But a supplier expanding factories because its customers have issued purchase orders is in a different position from one expanding because industry forecasts show thousands of planned data centers.

This is where "ghost demand" should change the analysis.

A speculative project can occupy grid planning capacity years before it needs chillers or rack-level power distribution. If regulators remove duplicate or underfunded requests, the first visible correction may appear in utility demand forecasts without immediately damaging manufacturers whose production is already allocated to financed projects.

Later, however, weaker project conversion can matter for equipment categories purchased closer to commissioning. Suppliers whose growth depends disproportionately on future campus build-outs rather than existing backlog deserve more scrutiny.

The reverse is true for equipment with long replacement cycles and multiple end markets. A cancelled data center does not eliminate an aging utility transformer that still needs replacement, a new generation project that needs interconnection equipment or an industrial facility that needs switchgear.

The investment signal is demand quality, not announced megawatts

The emerging data center correction therefore creates a useful hierarchy.

Utility connection requests are the weakest evidence because they can be duplicated, unfunded or deliberately oversized. Permitted and financed projects are stronger. Customer deposits, equipment reservations and signed contracts are stronger again. Supplier orders and backlog are closest to observable economic demand.

This distinction also changes how supply-chain dependence should be interpreted. The most interesting relationship is not necessarily the supplier attached to the largest announced data center. It is the supplier whose scarce production capacity has already been claimed by real customers, especially when those customers span utilities, industrial projects and funded data center builds.

That favors analysis of companies such as Eaton, Siemens Energy, GE Vernova and other electrical-equipment manufacturers through their actual order books and capacity constraints rather than through aggregate AI power forecasts. It also argues for greater caution when valuing data center equipment capacity that has been built primarily against an industry pipeline whose underlying projects have not crossed the same financial thresholds.

Conclusion

Data center ghost demand does not mean the infrastructure shortage disappears. It means the shortage needs to be located more precisely.

The strongest evidence of genuine capacity constraint remains in transformers, high-voltage breakers, switchgear and other equipment already facing multi-year lead times and serving several expanding markets. The more speculative exposure sits where manufacturing capacity depends on announced data center projects converting into funded construction.

For investors, the useful question is no longer how many gigawatts developers have requested. It is how much of that demand has become a real buyer-supplier relationship.

Scope and limitations

This article distinguishes utility queue demand from procurement demand using public reporting and company disclosures. It does not estimate a cancellation probability for individual data center projects or assign a probability that any named supplier will realize revenue from a particular announced campus.

The 700+ GW figure represents large-load electricity requests reported by Reuters across parts of the United States, not a forecast of completed data center capacity. The Exelon and AEP Ohio reductions reflect changes in their own pipelines after stricter requirements and should not be generalized mechanically to every utility territory.

Transformer and breaker lead times vary by equipment type, voltage class, supplier, geography, and order timing. Eaton's order and backlog growth figures are company-wide electrical indicators and are not presented here as data-center-only metrics.

No Altsets relationship percentages are used because the supplied article does not provide a quantified company relationship that would improve the analysis without introducing unrelated data.

For evidence limits and relationship methodology, see the Altsets methodology.

Sources

1. Reuters, "Texas' halt on powering data centers reflects US reckoning over 'ghost' demand," September 1, 2026. https://www.reuters.com/business/texas-halt-powering-data-centers-reflects-us-reckoning-over-ghost-demand-2026-09-01/

2. Reuters, "US power companies scramble to secure equipment as surging data center demand strains supplies," July 9, 2026. https://www.reuters.com/business/energy/us-power-companies-scramble-secure-equipment-surging-data-center-demand-strains-2026-07-09/

3. Eaton, "Eaton Reports Record Second Quarter 2026 Results, with Strong Organic Growth, Accelerating Orders and Backlog, and Raises Organic Growth Guidance," 2026. https://www.eaton.com/us/en-us/company/news-insights/news-releases/2026/eaton-reports-record-second-quarter-2026-results.html

How to Cite This

For research inquiries or data access: press@altsets.com

Go Deeper