John Bistline

Here's a link to the report: powering-intelligence.epri.com

❌ "We know how much electricity data centers use"

There is currently little public reporting of DC electricity. We cross-checked our estimates against state C&I sales data and found they couldn't be much higher and still be consistent with observed totals.

Data transparency is a major gap.

❌ "Data centers run at full capacity all the time"

Our facility data shows peak utilization of 62-80% of nameplate. And during rapid buildout, fleet average utilization can actually drop, since new capacity takes years to ramp up.

❌ "AI is the whole story"

AI workloads are estimated at 15-25% of DC electricity today. Growing fast, yes. But non-AI data center demand (streaming, cloud, enterprise) continues steady growth, too. The base is bigger than people think.

❌ "Data centers announced X GW, so the grid needs X GW of new capacity"

Announced nominal MW is a pipeline indicator, not a peak forecast. Non-IT loads, ramp-up schedules, load shapes, and onsite generation all mean realized peak is lower than the headline.

Now that our updated data center electricity analysis is out, a few things I see people often get wrong about DC load growth:

Co-authored with Geoff Blanford, Tom Wilson, and Nils Johnson. This builds on the 2024 edition that was EPRI's most-downloaded deliverable.

Happy to discuss what we found, plenty more in the full paper and site: powering-intelligence.epri.com

We also compared DC load to EV charging load. Data centers: ~184 TWh in 2024. Light-duty EV charging: ~12 TWh.

Under reference projections, EVs don't surpass DC load nationally until the mid-to-late 2030s. In heavy DC states, maybe never.

One finding I think deserves more attention: how the generation mix to serve this load depends entirely on the policy environment.

Under current policies → natural gas dominates.
Under 24/7 carbon-free energy targets → renewables, storage, nuclear.

And some states barely on the radar are emerging fast (Indiana, Louisiana, Mississippi, Ohio) driven by land availability and power access for large AI training facilities.

Meanwhile, CA and NY have significant existing capacity but relatively little projected growth.

The state-level story is where it gets really interesting.

Virginia is currently the only state where DCs exceed 20% of electricity. By 2030, seven more states could join that club: Oregon, Iowa, Nebraska, Nevada, Wyoming, Arizona, Indiana.

But a critical nuance that gets lost in headlines: announced MW ≠ peak demand.

We mapped the chain from nominal IT capacity → nameplate (add cooling/PUE) → active capacity (ramp-up lags) → realized peak load.

The peak is often meaningfully below the headline number.

The past 18 months saw record levels of data center development activity. The pipeline of projects under construction or in advanced planning grew dramatically.

New from EPRI: Our updated Powering Intelligence analysis of U.S. data center electricity demand.

Data centers could consume 9-17% of U.S. electricity by 2030, up from 4-5% today.

Our projections are ~60% higher than last year's. Here's why. 🧵

According to MethaneSAT measurements, the Powder River Basin loses 7.5% of gas production to the atmosphere.

More results from MethaneSAT (one year in orbit before failure) available here: www.methanesat.org/project-upda...

EPRI's updated Powering Intelligence data center scenarios drop tomorrow.

Here's where our projections stood in 2024. How much higher do you think the new numbers will be?

Yes, I can definitely put together state/region deep dives! Maybe next week?

In 2025, the U.S. added 53 GW of new electricity capacity, the most since 2002. In 2026, it's on track for 86 GW, mostly solar/storage. To put that in perspective, the U.S. is expected to add more capacity in the next two years than it did in an entire decade of the 2000s combined.

Weighting for net zero Nature Climate Change - Policy and planning increasingly depend on large ensembles of climate and energy scenarios, but these collections can be biased and hard to interpret. A new weighting...

Check out my piece here: rdcu.be/e5sjx

And read the original article here: www.nature.com/articles/s41...

Are you a fan of decarbonization scenarios, oblique Samuel Beckett puns, and didgeridoos? If that's you, today is your (improbably) lucky day! My new piece in Nature Climate Change touches on all three in summarizing the important new paper from Hamish Beath and coauthors.

Adaptation is millions of decentralized decisions: where people live/work, what they buy, what they build, what they insure. Policy that improves incentives/info can move those margins at scale.

Great new Science article from my colleague Delavane Diaz and others: www.science.org/doi/10.1126/...

Hot take that shouldn't be hot: if you don't handle endogeneity, you can "prove" renewables raise prices... because you've basically regressed confusion on correlation. This paper finds modest price decreases by using a clever Bartik-style instrument: www.colorado.edu/economics/26...

Renewables and retail power prices: the "they must be making bills skyrocket" story doesn't show up in the data. New paper by Alec Truax finds ~0.3% lower retail prices for each +1p.p. renewables share (U.S., 2004-2023) with causal ID.

Hot modeling take

China now consumes a third of global electricity, more than twice that of the United States.

This is from a new Joule commentary by Trieu Mai and others: www.sciencedirect.com/science/arti...

Also, if you are planning for load growth, you are likely planning for transmission expansion, which scales roughly proportionally with load.

Transmission is the "Swiss Army knife" investment: it can lower capex, lower opex, support resource adequacy, connect wind/solar... and still clears benefit-cost >1.5 in most studies.*

*If you can permit it.

Modelers when someone calls scenarios "predictions"

EIA-860M has monthly values, including the Dec. 2025 update: www.eia.gov/electricity/...

My bad... I meant to say AC in my last tweet. I think both EIA EPM and and 860M are in AC terms.