In this week's energy market update, we explore a major announcement from leading AI chipmaker Nvidia, software company Emerald AI, and major energy players like Constellation to power a new class of "flexible AI factories". By utilizing Nvidia's latest Vera Rubin chip and Emerald AI's conductor platform to modulate compute demand in real-time, Nvidia claims this approach could unlock up to 100 gigawatts of capacity across the US power system.

With the US grid staring at expected peak demands that existing infrastructure simply cannot accommodate in the next three to five years, flexibility is becoming critical. For energy professionals tracking massive load growth, this video unpacks what this flexible architecture actually means for the grid:

The Grid Bottleneck & Souring Costs: Why adding inflexible data centers pushes up peak demand and exacerbates supply scarcity. We look at PJM's capacity market, where prices have soared seven or eightfold, costing ratepayers an estimated $23 billion over the last three auctions.

The Economic Power of Flexibility: How modulating compute loads during grid scarcity could allow massive new demand to connect without requiring billions in new infrastructure. We highlight recent Duke University studies suggesting that avoiding just 1% to 2% of peak hours could reduce utilities' new natural gas construction costs by 10 to 15%.

Real-World Testing: A look at the limited empirical data we have so far, including a peer-reviewed test at an Emerald AI data center in Arizona that successfully reduced power consumption by 25% during peak hours. We also discuss Google's recent milestone of surpassing 1 gigawatt of data center demand response.

Regulatory Skepticism & Risk: Why PJM's Independent Market Monitor (IMM) is pushing back hard against treating data centers as paid demand response assets. We discuss the immense financial risk to ratepayers if a data center fails to curtail power during an emergency, and the argument that flexibility should simply be a mandatory precondition for interconnection.

While the economic incentives and technical concepts are incredibly promising, the industry still needs to prove that this combination of silicon and electrons can be predictably and repeatedly flexible at scale. Join us as we unpack the 100 GW claim and discuss why significant caution is still warranted

Support the show

🎙️ About Energy Future: Powering Tomorrow’s Cleaner World

Hosted by Peter Kelly-Detwiler, Energy Future explores the trends, technologies, and policies driving the global clean-energy transition — from the U.S. grid and renewable markets to advanced nuclear, fusion, and EV innovation.

💡 Stay Connected
Subscribe wherever you listen — including Spotify, Apple Podcasts, Amazon Music, and YouTube.

🌎 Learn More
Visit peterkellydetwiler.com
for weekly market insights, in-depth articles, and energy analysis.

In early March, mid-Atlantic grid operator PJM Began using Ambient Adjusted Ratings to better determine how much power can flow through its lines based on actual weather conditions. In addition, the DOE announced it will award billions for quick and effective upgrades to the transmission system.

First we have to fix the broken interconnection issue. For all projects seeking interconnection to the grid from 2008 through 2019, only 19% of the projects actually flowed power by the end of 2024. The typical project built in 2025 took 55 months to get through the queue, compared with 36 months in 2015.

But even if all of that new supply capacity could be processed through interconnection queues, there are simply not enough transmission lines to accommodate the planned resources. And few new lines are being built: less than 1,000 miles of 345 kV+ transmission lines were completed in 2024 – far less expansion than is needed, especially in the face of enormous new data center demand.

The biggest challenge is permitting for new rights-of-way, which can take well over a decade. There is a glimmer of hope that the federal government may reform the permitting process prior to the mid-terms, but it’s unlikely.

Grid-enhancing technologies, or GETs, can offer some relief by doing more with existing transmission. In addition, there is the growing potential for reconductoring.

The GETs technology with the greatest near-term is dynamic line rating, or DLR. As power lines move more power, they heat up. Lines are limited in terms of how much they can energy move by static ratings, based on worst case weather assumptions, such as 100 degrees F with no wind.

Such conditions rarely occur, but with static ratings flows cannot exceed those pre-set amounts. Most days, one could move much more power through that line, if one were

using DLRs - a combination of software and sensors. DLRs measure ambient temperatures and wind (wind wicks lots of heat away from the line, as well as how much sunshine is warming the wires. Sensors also measure how much the wire is physically sagging at any given moment. This information helps operators move more power without hitting “thermal violations.”

A 2024 case study showed static ratings could be exceeded 100% of the time, with average capacity increases of 81%. In summer, one could exceed the static ratings 94% of the time, with average increases of 27%.

A less capital-intensive approach that doesn’t require physical sensors and uses weather data, but also fails to measure the impact of wind, is called Ambient Adjusted Rating or AAR. AARs automatically predict transmission line capacity on an hourly basis.

The Federal Energy Commission’s 2021 Order 881

Support the show

🎙️ About Energy Future: Powering Tomorrow’s Cleaner World

Hosted by Peter Kelly-Detwiler, Energy Future explores the trends, technologies, and policies driving the global clean-energy transition — from the U.S. grid and renewable markets to advanced nuclear, fusion, and EV innovation.

💡 Stay Connected
Subscribe wherever you listen — including Spotify, Apple Podcasts, Amazon Music, and YouTube.

🌎 Learn More
Visit peterkellydetwiler.com
for weekly market insights, in-depth articles, and energy analysis.

The U.S. solar industry installed 43.1 gigawatts-direct current (GWdc) of capacity in 2025, down 14% from 2024. GWdc is the nameplate rating of projects before they connect to the grid through inverters, which convert direct current (DC) to the alternating current (AC) our grid uses.

Two elements lower DC ratings to AC ratings. First, inverter losses account for around 4%.

More importantly, solar panels have specific output duration curves; there’s only a very small period when they produce maximum output, or even 80–90%. It’s uneconomical to buy an inverter that rarely hits full MW ratings, so developers resort to “solar clipping.” A 100 MWdc solar array might use inverters delivering a maximum of 80 MW of AC power to the grid. Typical DC/AC ratios are 1.1 to 1.25. You lose only a bit of energy on an MWh basis, but with significantly lower inverter costs. Therefore, MWdc numbers must be translated to the real-world MWac of the grid.

However, all capacity is not the same: a MW of solar capacity has two factors differentiating it from, say, a MW of gas-fired generation.

First, solar operates at a different capacity factor (a resource operating at 100% output all year would have a 100% capacity factor). An average panel capacity factor is 25%, compared to 60% for a combined-cycle gas plant. Because of this, it’s best to think in terms of energy generated. Location also matters; the capacity factor in Massachusetts is 16.5%, while in Arizona it is 29%.

One way to compare these is by energy output. Solar is now approaching 10% of total energy contributed to the grid. Additionally, solar arrays can be deployed faster than new turbines. With rising data center demand, we need all the electricity we can get.

(Source: https://www.eia.gov/todayinenergy/detail.php?id=67005)

Furthermore, solar is not dispatchable. It only generates power when the sun shines, while a gas plant can be called upon at any time, except during certain extreme weather events. In 2024, the mid-Atlantic grid operator PJM down-rated combined-cycle turbines from 96% to 79% in terms of their ability to meet peak demand during the worst hour of the worst day, and recently lowered that rating further to 74%. By comparison, PJM rates solar at only 7%.

When you hear about solar in terms of MWdc, it helps to reframe those values using the information above. Nonetheless, solar has grown considerably. In 2009, about 1 GW (1,000 MW) of solar was added in the U.S. That cumulative total is now 279 GWdc, and analyst Wood Mackenzie forecasts an increase of 490 GWdc over the next decade.

Support the show

🎙️ About Energy Future: Powering Tomorrow’s Cleaner World

Hosted by Peter Kelly-Detwiler, Energy Future explores the trends, technologies, and policies driving the global clean-energy transition — from the U.S. grid and renewable markets to advanced nuclear, fusion, and EV innovation.

💡 Stay Connected
Subscribe wherever you listen — including Spotify, Apple Podcasts, Amazon Music, and YouTube.

🌎 Learn More
Visit peterkellydetwiler.com
for weekly market insights, in-depth articles, and energy analysis.