Why Data Centers Will Transform the US Grid | Livestream

Why Data Centers Will Transform the US Grid | Livestream

Title history
  • Why Data Centers Will Transform the US Grid | Livestream — observed on 20261030-0851 UTC (YouTube)
  • Why Data Centers Will Transform the US Grid - But Only Use a Fraction of the Power They Request — observed on 20261028-1952 UTC (YouTube)

by Brandt Vermillion and Tim Woodruff · View the channel on PSI

Applied Sciences & Engineering Energy THEnergy technology and engineeringTHYEnergy, power generation, distribution and storage
🎙 Brandt Vermillion and Tim Woodruff 👥 28K 📅 October 28, 2026 ⏱ 61 min 👁 945 📄 expert opinion 🧭 2026-10-30
Available in: English (current) Français

Keywords

data center demandERCOTpower pricesinterconnection queuesupply chain

Summary

This livestream from Modo Energy presents a framework for forecasting data center power demand in the US. The hosts, Brandt Vermillion and Tim Woodruff, explain that system operators have received ~800 GW of interconnection requests, 16 times the current operating fleet. They filter this inflated pipeline by assessing project readiness: only ~540 GW have public evidence of progress, and ~180 GW have advanced beyond planning. They identify five factors shaping buildout: adoption, business case, efficiency, supply chains, and capital. They note that chip efficiency is improving exponentially (Nvidia 16x from 2020 to 2026) but high bandwidth memory is limited to ~25 GW/year globally. Building all 540 GW would cost $13-17 trillion, while holding current hyperscaler capex ($630B in US) yields a terminal capacity of 77-131 GW. Data centers typically use 50-70% of nameplate, as shown in ERCOT and Dominion examples. For ERCOT, the realistic pipeline is ~100 GW, leading to ~28 GW of added peak demand over six years, and wholesale prices could rise from $25-35/MWh to $150/MWh by early 2030s before moderating. The livestream ends with a Q&A on behind-the-meter generation, GPU efficiency, and price impacts.

187 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of this information is high: it provides a rigorous, evidence-based approach to cutting through the hype of data center load growth. The hosts use concrete numbers (e.g., $50k/MW queue deposits, 50% utilization rates, cost per GW campus) and logical reasoning to bound the uncertainty. They clearly separate legitimate projects from speculative ones by filtering via press releases, permits, and site works. The argumentation is solid, as they consistently tie constraints (capital, supply chains, grid capacity) to capacity outcomes, and they present both low and high scenarios. However, the presentation is an expert opinion rather than a peer-reviewed study, and some assumptions (e.g., constant capex) are simplified for bounding exercises.

Scientific Rigor, Source Quality, Title Accuracy

The rigor is commendable: the hosts draw on data from system operators, public project announcements, and their own internal models, but they do not cite external sources explicitly. The only external link provided is Modo Energy’s website, which serves as a product pitch rather than supporting data. The title is accurate, as the content directly addresses how data centers transform the grid. No formal citations or references to academic literature are given, which limits verifiability. The analysis is transparent about methodologies, but lacks peer-reviewed validation. The livestream format includes a Q&A, showing engagement with the audience, but no comments data is provided to analyze public reception.

232 words

Title / Content Match

The title accurately reflects the content: the livestream discusses how data center growth will affect the US power grid, including forecasts and constraints.

Quality & Reliability

8/10

High-quality data and analysis based on Modo Energy's own research, with detailed metrics and scenarios, but lacks peer review.

Chapters

Cited Sources

  • Modo Energy — Website for Modo Energy's long-term power market forecasts and data center research.

Contribution & Novelties

The livestream provides a novel methodology for filtering inflated interconnection queues to estimate realistic data center load growth, combining granular project tracking with macro constraints (capex, chip supply, utilization rates). It offers a clear framework that power market analysts and investors can adopt.

Pour aller plus loin :

  • Data center — Overview of data center fundamentals and energy consumption.
  • ERCOT — Details on the Texas grid operator and its unique market design.
  • High Bandwidth Memory — Explains the memory technology that limits AI compute and thus data center capacity constraints.

90 words

Radar Profile

The profile shows high scores in information quantity and reliability, with moderate technical depth. The livestream excels in presenting detailed data and structured analysis, making it a strong resource for stakeholders in energy markets and data center development.

Reliability 8/10