Less Talk, More Megawatt Hours - Reading & Podcast Picks - Nov. 11th 2025

AI load is real, capital is moving, new rules for big users are landing. This issue focuses on turning headlines into uptime.

We’ve talked about policy, incentives, and “someday” plans. This week’s reading makes the shift impossible to ignore. AI is pulling forward demand, global finance is routing around delays, and rulebooks for big loads are being rewritten in real time. The center of gravity is moving from intention to execution. That’s good news for people who can plan, build, and keep systems online — the kind of leadership our audience knows well.

Below are the picks. Each section connects to one mission: stop arguing about the future and start delivering the power it needs.

The State of AI: here comes the energy crunch (Financial Times)

FT’s series with MIT Technology Review calls the shot cleanly: “In the age of AI, the biggest barrier to progress isn’t money but energy.”

“It wasn’t always like this. For about a decade before 2020, data centres were able to offset increased demand with efficiency improvements. Now, however, electricity demand is ticking up, with billions of queries to popular AI models each day. Efficiency gains aren’t keeping pace and, with too little new power capacity coming online, the strain is starting to show. Electricity bills are ballooning for people who live in places where data centres place a growing load on the grid.

If we want AI to deliver on its big promises without driving electricity prices sky high, the US needs to learn some lessons from the rest of the world on energy abundance. Just look at China.

China installed 429GW of new power generation capacity in 2024, more than six times the net capacity added in the US during that time.”

The reporting frames data-center growth as an infrastructure problem first, not a VC problem. Efficiency gains help, but they won’t outrun physics or permitting. The U.S. can’t wish its way out of this, we have to build. The piece also undercuts the idea that AI’s energy use is trivial or purely self-solving, noting the “there’s a lack of public data about AI systems’ energy needs”

A second thread: flexibility. Smarter interconnection and demand response can blunt the peak. As FT summarizes the thesis, the story is less about hype and more about the grid edge:

“Data centres that can cut their power use at times of grid stress should be the norm, not the exception. Likewise, we need more deals such as those giving cheaper electricity to data centres that let power utilities access their backup generators. Both reduce the need to build more power plants, which makes sense regardless of how much electricity AI ends up using. “

We need to treat AI as a load and design for it. Interconnection strategy, firming, storage siting, and contracts with flexible loads are now core operator disciplines.

Engie chairman: AI demand will keep propelling solar (Bloomberg)

Bloomberg’s readout is blunt:

“Policy uncertainty is deterring investment in US energy projects, though surging power demand tied to artificial intelligence will continue to support wind and solar farms”

Despite federal headwinds, deployment is still moving quickly at home and abroad. As Bloomberg notes:

“the French energy giant would like to invest more in the US, where it’s already one of the largest renewable developers, and it’s also finding opportunities in other parts of the world, Clamadieu said. That includes the Middle East and India, where the focus is “on quick development” and “trying to reduce” investor risks.”

On resource mix and timelines:

“Renewable wind and solar are key near-term sources of electricity, as new nuclear projects take years to develop and there’s a shortage of turbines for natural gas power plants.”

That’s the market separating noise from signal. Capital is following load growth even when the policy picture wobbles. For practitioners, that means fewer cycles arguing abstractions and more time scoping arrays, storage, and transmission that map to real offtake.

Build to the load. If you can translate demand curves into bankable megawatt-hours, you’ll stay busy.

Why yesterday’s power grid security won’t work tomorrow (Utility Dive)

The piece is a clear brief on where grid security actually lives now. The perimeter is gone. Every sensor and DER endpoint is a doorway. The job is to integrate cybersecurity into daily operations, not bolt it on after the fact.

“The pre-digital grid era of isolated, air-gapped control systems no longer exists. Distributed energy resources now number in the thousands… each one presenting a potential entry point for attackers.”

“It’s tough to distinguish between normal maintenance activities and cyberattacks that trigger motor failures… without greater visibility… operators won’t detect attacks until their systems reach critical failure points.”

The numbers underscore the gap. Utilities report malware, cloud vulnerabilities, and ransomware at the top of their concern list, but only about half tie spending to formal risk assessments. Confidence and preparation are not the same thing.

“About 37% of respondents believe they could bounce back from an operational technology attack in a day. Yet only half say their spending decisions come from formal risk assessments.”

Culture beats tools. The article pushes for risk-based security, unified ops and IT, and workforce training as the first line of defense. Treat physical and cyber as one fight, with one team.

“Traditional security assumed a utility could secure perimeters and trust everything inside. Modern grids have no meaningful perimeter… Smart grid security means protecting what matters most.”

“No single department can manage cybersecurity for an entire utility. Grid security must become everyone’s responsibility.”

We have to build integrated defenses. Start with a formal risk assessment, unify physical and cyber playbooks, log what matters, and train crews so protocols hold under stress. That is how you keep uptime and avoid explaining failures to regulators.

PODCAST: Energy Policy in Motion: DOE Section 403 & data-center power

A clean explainer for a fast-moving change. DOE invoked Federal Power Act Section 403 to push FERC toward a national framework for large-load interconnections. The episode summarizes the shift crisply: it “calls for a sea-change in the way large energy users, especially data centers, interconnect to the power grid.”

Read that alongside the FT series and the Bloomberg/Engie signal and you get the week’s picture: regulatory architecture is being adapted to the Big-Load Era. That means new queue dynamics, different risk allocation, and more value on teams that understand capacity markets, transmission planning, and construction sequencing.

Learn the new rulebook and use it to clear the path for real projects.

If there’s a single throughline, it’s this: execution over intention. AI-driven load is here, global finance is lining up behind projects that can actually be built, and the rulemakers are moving to accommodate large interconnections. We don’t need louder arguments; we need steadier hands.

The people who will matter most in the next 24 months will do simple things well: design to the load, harden the weakest links, work with communities first, close the last-mile logistics, and publish schedules they intend to keep. That’s how you turn policy into power and ambition into uptime.

If these picks help your week, pass them to someone who builds for a living.