US Data Centers: How To Think Through Four Capacity Questions
AIThis post was created with the assistance of artificial intelligence (AI).

TL;DR

Prime Big Deal Days · Oct 6–7Offer from Amazon

Get tech for your team delivered free — and shop member deals

  • Fast, free delivery on millions of items
  • Access to Prime Big Deal Days deals on October 6–7
  • Prime Video, Amazon Music and more included
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

Rymvard published four illustrative U.S. data center capacity scenarios on Oct. 3, 2026, covering Northern Virginia, Texas, Arizona and central Ohio. The examples show how power reservations can differ from capacity a site can use or sell, but they do not document customer deployments or verified product results.

Rymvard published four illustrative scenarios on Oct. 3, 2026, describing how grid connection delays, curtailment rules, cooling limits and utility charges can make a data center’s usable or sellable capacity differ from its reserved power, as explored in the original analysis of U.S. data center capacity. The examples cover Northern Virginia, Texas, Arizona and central Ohio and are intended to show the problem its early-access capacity ledger aims to address, not report outcomes at named customer sites.

In Northern Virginia, Rymvard points to lengthy waits for new utility connections and a separate planning question: whether existing sites draw less power than customers have reserved. The company says capacity that could be offered this year may already exist within a campus, rather than depending entirely on a new connection. The scenarios do not identify a specific campus or quantify the amount of potentially available capacity.

For Texas, the company describes curtailment obligations under Senate Bill 6, signed in June 2025. Rymvard says sites of 75 megawatts or more must accept curtailment when the grid operator sheds load. Its example raises the operational need to distinguish loads that support critical services from those that might be reduced; it does not describe a particular curtailment event or how an operator responded.

The other examples focus on different constraints. Rymvard says Arizona heat can limit cooling on the hottest afternoons. In central Ohio, it cites an AEP Ohio tariff approved by the Public Utilities Commission of Ohio that requires certain new data centers above 25 MW to pay for at least 85% of subscribed power for up to 12 years. The company says its product brings measured power, contracts, recovery reservations, cooling and demand into one ledger.

At a glance
announcementWhen: Published Oct. 3, 2026; Rymvard says it…
The developmentRymvard published four illustrative regional scenarios to explain how grid, cooling, curtailment and tariff constraints can affect data center capacity.

Why Reserved Power Can Mislead

A facility’s contracted or reserved power is not automatically the capacity it can reliably use, sell or afford. A connection delay can hold back expansion; a curtailment obligation can affect operations during grid stress; heat can constrain cooling; and a tariff can leave an operator paying for subscribed power it does not draw. Those differences can shape customer commitments, equipment plans and cost forecasts.

For utilities and grid planners, better visibility into measured demand and loads that can be reduced may help distinguish a reservation from actual consumption. Rymvard’s examples frame a practical information problem, but the announcement supplies no independent validation, quantified savings or evidence that the ledger changes grid outcomes. Organizing data may support planning; it does not itself add power, speed up connections or remove contractual obligations.

Amazon

data center power monitoring system

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Four Markets, Four Constraints

Rymvard presents these examples as local scenarios, not a national capacity forecast. Northern Virginia’s case concerns connection timing and the gap between reserved and measured demand. Texas’s centers on curtailment; Arizona’s on cooling under extreme heat; and Ohio’s on the cost of subscribed power under a regulated tariff. The different cases illustrate why a single headline capacity figure may not capture the limits facing an individual facility.

The Ohio example refers to the AEP Ohio data center tariff in Public Utilities Commission of Ohio case 24-508-EL-ATA, with an order dated July 9, 2025. Rymvard says its product is in early access, while the published screens and scenarios use an illustrative estate. The company says they are not based on a customer site or outcome, and no customer is identified.

“Rymvard joins measured power, contracts, recovery reservations, cooling and demand into one ledger.”

— Rymvard

Amazon

industrial cooling solutions for data centers

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Product Results Remain Unverified

The announcement does not name customers using the product, disclose measured results or quantify whether the ledger has improved planning, reduced costs or changed curtailment decisions. The scenarios are illustrative and do not establish how often each constraint occurs across the named markets, or the financial effects at any particular site.

Rymvard has not provided details on data inputs, integrations or verification methods, or how the ledger is used in operational decisions. Pricing is not public; the company says terms are agreed with early-access partners. It is also unclear how the product handles differences among site-specific contracts, measurements and operating conditions.

Amazon

uninterruptible power supply (UPS) for data centers

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Evidence to Watch in Early Access

Rymvard says the product is available in early access and invites interested parties to contact the company. It has not announced a broader release date, published a pricing schedule or named a customer deployment. Further reporting from the company could clarify what information the ledger uses and how operators apply it to capacity decisions.

The next meaningful test will be whether Rymvard shares customer deployments or independently verifiable outcomes, such as documented changes in capacity planning or costs. Until it does, the four scenarios are best read as demonstrations of the constraints the product seeks to organize—not proof that it has solved them.

Amazon

energy management software for data centers

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

What did Rymvard announce?

Rymvard published four illustrative data center capacity scenarios for Northern Virginia, Texas, Arizona and central Ohio on Oct. 3, 2026. The examples describe how local power, operating and tariff constraints can affect capacity.

Are the scenarios based on real customer sites?

Rymvard says the examples use an illustrative estate, not a customer site or outcome. The announcement does not name customers or report a specific facility’s results.

What does the early-access product do?

Rymvard says its ledger combines measured power, contracts, recovery reservations, cooling and demand. The company has not published independent evidence that the product improves planning or reduces costs.

Has Rymvard announced pricing or a wider release date?

No. The company says pricing is agreed with early-access partners and has not announced a public pricing schedule or broader release date.

Primary source: Rymvard · via ThorstenMeyerAI.com

FALL

Fall Picks

As an affiliate, we earn on qualifying purchases.

You May Also Like

Artificial Intelligence: Ars Notoria And The Promise Of Instant Knowledge

New AI model, Ars Notoria, claims to deliver instant knowledge retrieval, raising questions about its accuracy and implications for learning.

Inference Optimization for MiMo v2.5: Pushing Hybrid SWA Efficiency to the Limit

MiMo v2.5 introduces new inference optimization techniques, significantly improving hybrid SWA efficiency. Details are confirmed, but full performance metrics are still emerging.

Will GPT-6 Be Released By December 31, 2026?

Speculation surrounds GPT-6’s release date, with market signals suggesting high confidence but no official confirmation from OpenAI.

The Local-First Agentic Operator

A single operator using agentic AI now builds and manages multiple software products across domains, challenging traditional organizational models.