What are the key takeaways from this Odd Lots episode?

Key insights include: **Supply-Demand Mismatch:** Data centers currently consume 45 gigawatts and forecasts predict 95 gigawatts by 2030, requiring 50 gigawatts of new capacity. However, utilities have already committed to 140 gigawatts of firm projects (110 gigawatts after PUE adjustment), creating potential massive oversupply by decade's end. This disconnect suggests utilities may be overbuilding infrastructure based on inflated demand projections.; **Existing Grid Capacity:** Texas operates an 87 gigawatt peak market that could immediately absorb 10 gigawatts of new data center load (equivalent to every NVIDIA chip produced in one year) using existing infrastructure for all but 40-50 hours annually. Ramping up underutilized coal and gas plants proves more economical than spending billions connecting distant renewable farms for rare peak demand periods.; **Ratepayer Protection Models:** Northern Indiana Public Service established the gold standard by creating a separate generation company for Amazon data centers that returns one billion dollars over fifteen years to residential customers, approximately 67 million dollars annually. This structure prevents ratepayers from subsidizing tech giants while utilities still benefit from earnings growth, though many jurisdictions lack similar protections.

What did Andy Devries discuss on Odd Lots?

Andy Devries, utilities analyst at Credit Science, challenges the data center energy demand narrative with simple math: utilities have committed to 110 gigawatts of new capacity, but third-party forecasts only project 50 gigawatts of additional demand by 2030. This potential oversupply threatens utility shareholders and ratepayers. Key topics include: **Supply-Demand Mismatch:** Data centers currently consume 45 gigawatts and forecasts predict 95 gigawatts by 2030, requiring 50 gigawatts of new capacity. However, utilities have already committed to 140 gigawatts of firm projects (110 gigawatts after PUE adjustment), creating potential massive oversupply by decade's end. This disconnect suggests utilities may be overbuilding infrastructure based on inflated demand projections.; **Existing Grid Capacity:** Texas operates an 87 gigawatt peak market that could immediately absorb 10 gigawatts of new data center load (equivalent to every NVIDIA chip produced in one year) using existing infrastructure for all but 40-50 hours annually. Ramping up underutilized coal and gas plants proves more economical than spending billions connecting distant renewable farms for rare peak demand periods..

How long is this episode of Odd Lots?

This episode is 46 minutes long. SignalCast provides an AI-generated summary so you can get the key insights in about 3 minutes.

Odd Lots

The Utilities Analyst Who Says the Data Center Demand Story Doesn't Add Up

February 2, 2026

46 min episode · 2 min read

Andy Devries

Episode

46 min

Read time

2 min

Topics

Investing, Fundraising & VC, Science & Discovery

AI-Generated Summary

Published Feb 2, 2026

Key Takeaways

✓Supply-Demand Mismatch: Data centers currently consume 45 gigawatts and forecasts predict 95 gigawatts by 2030, requiring 50 gigawatts of new capacity. However, utilities have already committed to 140 gigawatts of firm projects (110 gigawatts after PUE adjustment), creating potential massive oversupply by decade's end. This disconnect suggests utilities may be overbuilding infrastructure based on inflated demand projections.
✓Existing Grid Capacity: Texas operates an 87 gigawatt peak market that could immediately absorb 10 gigawatts of new data center load (equivalent to every NVIDIA chip produced in one year) using existing infrastructure for all but 40-50 hours annually. Ramping up underutilized coal and gas plants proves more economical than spending billions connecting distant renewable farms for rare peak demand periods.
✓Ratepayer Protection Models: Northern Indiana Public Service established the gold standard by creating a separate generation company for Amazon data centers that returns one billion dollars over fifteen years to residential customers, approximately 67 million dollars annually. This structure prevents ratepayers from subsidizing tech giants while utilities still benefit from earnings growth, though many jurisdictions lack similar protections.
✓Private Credit Pricing Signals: Meta paid 220 basis points over treasuries for 25 billion dollars in off-balance-sheet data center financing through PIMCO, immediately trading at 140 basis points and generating two billion dollars profit. Meanwhile, CoreWeave bonds yield 10 percent to 2030, suggesting bond markets price significant default risk despite 50 billion dollar market capitalization, reflecting skepticism about 2030 oversupply scenarios.
✓Power Construction Economics: New combined-cycle gas plants cost 3,000 dollars per kilowatt to build (up from 1,200 dollars a decade ago), but data centers themselves cost 40,000 dollars per kilowatt. Big tech pays 95 dollars per megawatt hour for power versus 55-60 dollar forward curves, making the three dollar per kilowatt premium for dedicated generation economically trivial relative to total data center investment.

What It Covers

Andy Devries, utilities analyst at Credit Science, challenges the data center energy demand narrative with simple math: utilities have committed to 110 gigawatts of new capacity, but third-party forecasts only project 50 gigawatts of additional demand by 2030. This potential oversupply threatens utility shareholders and ratepayers.

Key Questions Answered

•Supply-Demand Mismatch: Data centers currently consume 45 gigawatts and forecasts predict 95 gigawatts by 2030, requiring 50 gigawatts of new capacity. However, utilities have already committed to 140 gigawatts of firm projects (110 gigawatts after PUE adjustment), creating potential massive oversupply by decade's end. This disconnect suggests utilities may be overbuilding infrastructure based on inflated demand projections.
•Existing Grid Capacity: Texas operates an 87 gigawatt peak market that could immediately absorb 10 gigawatts of new data center load (equivalent to every NVIDIA chip produced in one year) using existing infrastructure for all but 40-50 hours annually. Ramping up underutilized coal and gas plants proves more economical than spending billions connecting distant renewable farms for rare peak demand periods.
•Ratepayer Protection Models: Northern Indiana Public Service established the gold standard by creating a separate generation company for Amazon data centers that returns one billion dollars over fifteen years to residential customers, approximately 67 million dollars annually. This structure prevents ratepayers from subsidizing tech giants while utilities still benefit from earnings growth, though many jurisdictions lack similar protections.
•Private Credit Pricing Signals: Meta paid 220 basis points over treasuries for 25 billion dollars in off-balance-sheet data center financing through PIMCO, immediately trading at 140 basis points and generating two billion dollars profit. Meanwhile, CoreWeave bonds yield 10 percent to 2030, suggesting bond markets price significant default risk despite 50 billion dollar market capitalization, reflecting skepticism about 2030 oversupply scenarios.
•Power Construction Economics: New combined-cycle gas plants cost 3,000 dollars per kilowatt to build (up from 1,200 dollars a decade ago), but data centers themselves cost 40,000 dollars per kilowatt. Big tech pays 95 dollars per megawatt hour for power versus 55-60 dollar forward curves, making the three dollar per kilowatt premium for dedicated generation economically trivial relative to total data center investment.

Notable Moment

Devries reveals that OpenAI built ChatGPT using only two gigawatts of capacity, and all major hyperscalers combined currently use approximately 15 gigawatts (assuming 60 percent capacity factor). These modest numbers cast doubt on forecasts predicting 95 gigawatts by 2030, especially given rapid efficiency improvements in chip technology and limited number of ChatGPT-scale applications.

Know someone who'd find this useful?

Episode Transcript

UKG. Their HR, pay, and workforce management tools help business leaders empower their people. Because when work works, everything works. Learn more at ukg.com/work. This podcast is brought to you by Wise, the smarter way to manage your money internationally. If you're getting a headache from juggling different currencies and different bank accounts in different countries, there's a better way to receive money in the currency you need without the slow transfer times or hidden fees. Meet Wise, the savvy way to handle your money internationally. Hold balances in up to 40 currencies with the mid market exchange rate on every conversion. Whether you're receiving payments from tenants abroad, earning as a digital nomad, or converting dividends from your international investments, the Wise multicurrency account is for you. Be smart. Get Wise. Download the Wise app today or visit wise.com. Terms and conditions apply. So have you heard the story about the prescription plan with savings automatically built in? It's where a family of any size can feel confident the cost of their medication won't hold them back. Go to cmk.co/stories to learn how CVS Caremark helps members save just by being members. That's cmk.co/stories. Bloomberg Audio Studios. Podcasts, radio, news. Hello, and welcome to another episode of the Odd Thoughts podcast. I'm Tracy Alloway. And I'm Joe Weisenthal. Joe, imagine you are a utilities analyst. Yeah. Fun. And for years, you are laboring in the utility analysis minds. Yeah. And, you know, we like talking about utilities. We like talking about energy. We find pretty much anything interesting. That's right. Equal opportunity, interest people we are. But Yes. You gotta say utilities for a while, some people would say it was a little boring. Fly. No. That's right. I mean, for most of our careers, I think if you were a utilities analyst, a really big part of your job and maybe I'm wrong, but I just in the popular discourse was, like, talking about yield relative to treasuries. Right? They were seeing a sort of bond like instrument, etcetera, maybe a little bit of growth, but roughly safe haven ish dividend plays, I guess. Totally. And since I know where we're going with this conversation, one of the themes of the last few years has been what I would say is the old industries, they were either stable or cyclical becoming secular in the way they grow. Mhmm. I think that's right. So what is happening now is if you were I don't wanna say a lowly utilities analyst, but, you know, maybe a sort of forgotten utilities analyst outside of your sector, suddenly, you are very in demand. Right? Because all you hear about nowadays is the AI build out and energy constraints on that. And so, obviously, a lot of people wanna look at it from a utilities perspective. Totally. I always think, like, what a great luck that some people have in their careers. You know, you could be an analyst and learn modeling skills …

Get the full transcript (10,186 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all Odd Lots transcripts →