WSJ What’s News – "How The AI Race is Driving an Energy Wild West"

Date: October 16, 2025
Host: Caitlin McCabe
Featured Reporter: Jennifer Hiller

Episode Overview

This episode explores the mounting energy crisis driven by the explosive growth of AI and data centers in the United States. As tech companies scramble for more power, grid capacity and energy infrastructure can’t keep pace—leading companies to take unprecedented steps to create their own sources of electricity. The episode dives into how these energy challenges are reshaping the industry, introducing an “energy wild west,” and what this means for the future of AI, energy markets, and infrastructure policy.

Key Discussion Points & Insights

1. Headlines and Market Movers

• Brief mention of global political and economic news, including U.S.-Venezuela tensions and oil trade negotiations impacting global markets.

2. The AI Boom's Insatiable Energy Appetite

[08:26]

• Host Caitlin McCabe: Highlights that the demand for electricity from data centers is set to reach a staggering 12% of U.S. usage by 2028—up from less than 2% pre-2020.
• Key Issues:
  • U.S. is not building power plants or transmission lines fast enough.
  • Supply chain disruptions and permitting delays further bottleneck expansion.
  • Urgent need for infrastructure threatens to stifle the AI boom.

3. Supply Squeeze and Infrastructure Lag

[09:14]

• Jennifer Hiller:
  • The U.S. needs to build 80 GW of new power capacity annually, but is 15 GW short each year.
  • Retiring old plants adds pressure; current construction isn't closing the gap.
  • Equipment shortages, particularly transformers and gas turbines, slow down all projects.
  • The permitting process for new power lines is notoriously slow due to legal hurdles.

“We need definitely a lot of new power and a lot of new power infrastructure to be able to accommodate all of the AI growth.”
– Jennifer Hiller [09:14]

4. The "Energy Wild West": Tech Companies Take Control

[10:25]

• Companies are building their own on-site power plants as a stopgap, sometimes considering permanent off-grid operations.
• Case Studies:
  • OpenAI’s Stargate project in West Texas uses a combination of on-site and grid power.
  • Elon Musk’s Memphis initiatives started entirely off-grid, with possible future grid connections.
  • Traditional backup generators at data centers are now considered for primary use during early operations.

“What some tech companies and data center developers are doing is essentially building their own power on site and just providing their own power plants.”
– Jennifer Hiller [10:36]

• These solutions are deemed “bridges” until the grid can catch up, but some may remain fully independent if the grid doesn’t expand.

5. The Scale of New Data Centers

[12:14]

• The energy needs of a modern data center are enormous; single sites can require as much electricity as 1,000 Walmart stores.
• Construction of substations, transmission infrastructure, and transformers to support this demand is slow and logistically challenging.
• The U.S. has ample generation capacity, but the concentration and instantaneous requirements of new data centers overwhelm local infrastructure.

“Some of them use, for instance, the power of a thousand Walmart stores... it takes a while to build that out.”
– Jennifer Hiller [12:14]

6. Outlook: Will Supply Catch Up?

• Near-term: Expect energy crunches and power races over the next five years as demand massively outpaces supply.
• Mid-term: Potential transition to oversupply once infrastructure projects catch up, though not all markets or companies may benefit equally.
• Ongoing: The race to secure power will likely become the new battleground for AI dominance.

Notable Quotes & Memorable Moments

• On urgency:
  “People are really seeing this as a big crunch, maybe in the next five years or so, and that things may start to ease up after that.”
  – Jennifer Hiller [12:14]

• On solutions:
  “We’re seeing a lot of different creative solutions out there, but what some tech companies and data center developers are doing is essentially building their own power on site and just providing their own power plants.”
  – Jennifer Hiller [10:36]

• On the scale:
  “The size of these data centers is so overwhelming, it can be difficult to connect them in one location and provide that much power in one spot...”
  – Jennifer Hiller [12:14]

Important Segment Timestamps

• AI and energy demand introduction: [08:26]
• Explanation of grid and supply challenges: [09:14]
• Tech companies building their own power plants: [10:36]
• Data center scale and infrastructure needs: [12:14]

Takeaways

• The explosive growth in AI is creating a massive, immediate need for electricity that the U.S. grid is struggling to meet.
• Tech giants are increasingly building their own energy infrastructure, with some looking to permanently bypass public utilities.
• These emergency measures expose deep systemic planning failures in U.S. infrastructure and could have lasting impacts on energy policy, the environment, and the economy.
• The situation is fluid, with the potential for infrastructure to catch up in the long run—but the next five years will be marked by volatility and innovation as companies scramble to power the AI revolution.

Episode Tone:
Urgent, insightful, and pragmatic, with a focus on systemic challenges and the creative, sometimes desperate, solutions emerging in response to AI’s runaway growth.