📊 Full opportunity report: The queue. Why the grid, not the chip, is the binding constraint on AI. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
The main bottleneck for AI infrastructure growth has shifted from semiconductor chips to the US power grid. The interconnection queue delays projects by up to 12 years, prompting private power buildouts that shift costs onto ratepayers. This development redefines where and how AI capacity expands.
Recent data shows that the US power grid’s interconnection queue has become the primary bottleneck for AI infrastructure expansion, surpassing chip shortages as the main constraint. The median wait time for connecting new generation capacity has risen to nearly five years, with some projects facing delays up to twelve years. This shift significantly impacts how AI data centers and related infrastructure are built and financed.
Over the past two years, the narrative centered on chip supply — specifically, who controls GPU manufacturing and availability. That story has shifted; now, the bottleneck is the grid interconnection process. Currently, between 2,300 and 2,600 gigawatts of generation and storage projects are stuck in US interconnection queues, exceeding the country’s total installed power capacity. The median delay from project approval to commercial operation has increased from under two years in 2008 to nearly five years today, with some data-center projects quoting timelines up to twelve years.
This demand surge is driven by the explosive growth of data centers and AI-related power needs. US data-center power demand is projected to reach approximately 76 gigawatts in 2026, up from 50 gigawatts in 2024. Globally, data-center electricity consumption could surpass 1,000 terawatt-hours annually by the early 2030s, more than doubling 2022 levels. In Texas, requests for large power interconnections increased by 700% in a single year, from 1 gigawatt to 8 gigawatts, according to CenterPoint. Utilities such as ComEd, PPL, and Oncor report more gigawatts of data-center applications than their historical peak demands.
Because of these delays, capital is increasingly bypassing the grid. Private power solutions, like behind-the-meter gas plants or co-located nuclear facilities, are being built to circumvent the long interconnection timelines. Microsoft’s deal to restart Three Mile Island Unit 1, providing 835 MW of carbon-free baseload power, exemplifies this trend. However, these bypasses shift costs onto ratepayers; for instance, PJM’s capacity auction costs surged from $2.2 billion to nearly $15 billion in one year, with billions in transmission costs passed to consumers, sparking political debates and policy responses.
The queue.Why the grid, not the chip,
is the binding constraint on AI.
more than total installed capacity
up to 12 years for data centers
vs grid access maybe 2035
ratepayers · the cost-shift, concrete
in a single year
Virginia ratepayers (2024)
across PJM consumers
The grid is the bottleneck. The private grid is the response. And the seam between them — who pays for the public infrastructure the private builders still lean on — is where the economics and politics of the AI buildout are now decided.Thorsten Meyer · The Queue · AI Energy & Infrastructure 02
Implications of the Grid Bottleneck on AI Expansion
This shift from chip shortages to grid constraints fundamentally alters the economics and geography of AI infrastructure. The interconnection queue’s delay reprices locations, favoring sites with private power solutions that can bypass the grid, often at the expense of ratepayers and taxpayers. It also accelerates the privatization of power generation, with well-capitalized firms building private grids or co-locating power sources to avoid delays, creating a bifurcated landscape of self-powered versus grid-dependent projects. Politically, the costs associated with these bypasses are central to ongoing debates about infrastructure funding, ratepayer protections, and equitable growth in AI capacity.
Overall, this development signals a structural change: the bottleneck is no longer silicon but the physical and bureaucratic constraints of the power grid, reshaping how AI infrastructure is financed, built, and distributed across regions.
APC UPS Battery Backup for Power Outages, 600VA/330W Surge Protector, 7 Outlets, USB Charging, BE600M1 Uninterruptible Power Supply for Computers, Wi-Fi Routers, and Home Office Electronics
KEEP YOUR COMPUTER, WI-FI AND ROUTER RUNNING THROUGH POWER OUTAGES: Supplies short‑term battery power during outages to maintain…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Evolution of Power Constraints in AI Infrastructure
Initially, the focus of AI infrastructure buildout was on securing semiconductor chips, with supply chain issues dominating headlines. As chip supply stabilized, attention shifted to power generation capacity and the ability to connect new projects to the grid. Over the past decade, the US has faced a growing interconnection backlog, with the queue now containing over 2,300 gigawatts of projects—more than the entire country’s installed capacity. While China has been adding roughly 430 gigawatts annually, the US’s challenge lies in the slow, bureaucratic process of connecting new capacity, which has extended project timelines from under two years to nearly five or more.
This bottleneck has led to a strategic pivot: private entities and hyperscalers are increasingly building their own power sources or co-locating with existing facilities to bypass the grid. The result is a bifurcated buildout—one driven by capital-rich private projects that can move quickly, and another dependent on the slow, congested public grid. This dynamic has profound implications for the distribution of AI infrastructure and the political economy of energy costs.
“The grid is the bottleneck; the response is a private grid; and the seam between them — who pays for the transmission and capacity the private builders still lean on — is where the politics of the AI buildout now lives.”
— Thorsten Meyer
private gas power plant for AI infrastructure
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Remaining Uncertainties About Future Grid Capacity and Costs
It is not yet clear how quickly the US grid infrastructure will adapt to the mounting demand or whether policy interventions will accelerate upgrades. The long-term impact of private power solutions on grid stability and costs remains uncertain, as does the potential for regulatory changes to address cost-shifting issues. Additionally, the pace at which new policies might limit or regulate private bypasses is still developing.
Power Grid: Europe/North America Expansion
Power Grid: Europe/North America is a pair of expansion maps for Power Grid!
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps in Addressing the Interconnection Bottleneck
Expect ongoing policy debates and potential regulatory reforms aimed at streamlining interconnection processes and sharing costs more equitably. Infrastructure investments to upgrade the grid are likely to be prioritized, but their timelines remain uncertain. Meanwhile, private power projects will continue to proliferate, potentially reshaping the energy landscape and political discourse around AI infrastructure development. Monitoring legislative and utility industry responses over the next 12-24 months will be critical to understanding how this constraint evolves.
CyberPower ST425 Standby UPS Battery Backup and Surge Protector, 425VA/260W, 8 Outlets, Compact, UL Certified
425VA/260W Standby Uninterruptible Power Supply (UPS): Uses simulated sine wave output to provide battery backup power and to…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Why is the interconnection queue now the main bottleneck for AI infrastructure?
The queue delays stem from bureaucratic, physical, and permitting challenges that slow down connecting new power capacity, with median delays rising to nearly five years, far exceeding chip supply issues.
How are private power solutions bypassing the grid constraint?
Private entities are building behind-the-meter plants or co-locating with existing facilities, reducing reliance on the public grid and avoiding long interconnection delays, but shifting costs onto ratepayers.
What are the political implications of shifting costs to ratepayers?
Cost-shifting has sparked political debates about fairness, leading to policy proposals and pledges aimed at protecting consumers from rising transmission and capacity charges.
Could grid upgrades alleviate the bottleneck?
Yes, but current timelines for infrastructure upgrades are uncertain, and regulatory or political hurdles could delay these efforts further.
What does this mean for the future of AI infrastructure buildout?
The buildout is increasingly bifurcated: capital-rich private projects bypass the grid, while public projects face long delays, potentially altering the geographic and economic landscape of AI expansion.
Source: ThorstenMeyerAI.com
