Theme · Updated
What podcasters are saying about Power demand & the grid
Summary
In the past 30 days, 30 speakers on 23 finance podcasts discussed Power demand & the grid in 38 views. Chorus has followed Power demand & the grid since October 2, 2026.
Who talks about it most
- Connor Teskey · 2
- Erin Price-Wright · 2
- Chris Wright · 2
- Tyler Rosenlicht · 2
- Chamath Palihapitiya · 2
- Yasir Arafat · 2
- Ariana Salvatore · 2
- Chris Hentemann · 2
- Jason Snipe · 1
- Connor Love · 1
The latest views
Newest first. Every quote is checked against the episode's transcript.
Jason Snipe · Odyssey Capital Advisors
Power and grid buildout still has a long way to go, so the AI-driven selloff in power names is overdone.
Power and data-center power names have sold off with the broader AI drawdown. Snipes says that remains a big unfinished job on the power and grid side. He agrees with earlier comments that AI spending may have overshot, but argues the investment needed to sustain the surge is still ahead.
“But there's still a lot to be done from a power perspective. There's still a lot to be done from a grid perspective.”at 37:54
CNBC Halftime Report: Amazon in Focus as the Nasdaq Slides 10/3026 ·
Connor Teskey · Brookfield Asset Management
Power supply is structurally short; imbalance would persist even if demand plateaus or declines.
Teskey argues electricity and energy-infrastructure supply is structurally short. Demand is up while supply is only a small fraction of it. The imbalance began in the late 2010s, driven by re-industrialization and electrification, which alone would have provided decades of electricity demand growth. AI then took that imbalance and exaggerated it. Even if demand growth slowed, plateaued or declined, a huge supply-demand imbalance would remain for providers. He sees the theme lasting through the end of this decade and beyond, not just quarters or years, and calls it perhaps the greatest capital need in history with very attractive returns. Electricity has moved from something taken for granted to the single biggest bottleneck to growth for the largest companies and countries.
“Demand is up here and supply is down here. Even if demand growth was to slow, even if it was to plateau, even if it was to decline, there would still be a huge supply demand imbalance to the providers.”at 0:10
Goldman Sachs Exchanges: Brookfield’s Connor Teskey on the Real Bottleneck in the AI Buildout ·
Connor Teskey · Brookfield Asset Management
Largest power buyers increasingly consolidate on a few scale providers over many small suppliers.
Teskey argues that because electricity has become such a critical bottleneck, the largest buyers no longer simply choose the lowest-cost supplier. A hyperscaler that cannot get power for a data center on time is badly affected. Rather than dealing with 10,000 small suppliers who may or may not deliver on time or on budget, the largest users are focusing on a smaller number of scale providers with the credibility to deliver. Brookfield illustrates the shift: a decade ago it sold power mainly to utilities and governments, and now it sells most of its power directly to corporates under global framework agreements. He says certainty and delivery are worth far more than a penny on power price.
“we're increasingly seeing the largest users of electricity increasingly focus on a smaller number of scale providers who have the credibility to work through problems, deliver on time, deliver on budget”at 11:46
Goldman Sachs Exchanges: Brookfield’s Connor Teskey on the Real Bottleneck in the AI Buildout ·
Erin Price-Wright · Andreessen Horowitz (a16z)
The grid faces a major shift from mechanical systems to software-native power electronics and battery storage.
Erin argues the US is about to go through a large step change in how power is managed and transferred on the grid. The existing system was built with static, mechanical equipment deployed across the country in the 1970s and 1980s. It is being replaced by modern power electronics and battery storage architectures that are much more software-native and flexible. Because the replacement technology is completely different, she sees an opportunity to redesign the supply chain and the engineering behind it rather than reuse older components.
“we're about to go through a huge step change in how we manage and transfer power on the grid, moving from a system of static and mechanical systems... to modern power electronics, battery storage, architectures, which are much more software native and flexible.”at 35:47
a16z Podcast: The Startups Rebuilding American Industry | The AD50 ·
Connor Love · Andreessen Horowitz (a16z)
Even if US energy and defense demand undershoots forecasts, the demand cycle has started and will not reverse.
Connor takes the pessimist's case that not all new energy and defense systems will be adopted, so energy demand will not rise fourfold. He argues this does not matter. Once the demand cycle starts moving in one direction, the snowball keeps rolling, and even delivering half the required build-out will not reverse it. He adds that rethinking systems from first principles usually brings better cost and efficiency, which makes the new equipment financially sensible to buy and easier to service and manufacture. He links this to the wave of demand from data centers, new factories and new build-outs.
“I actually don't think it matters because I think once you start the snowball of the demand cycle moving in one way, even if you deliver half of that, you know, requirement You're not going back.”at 37:12
a16z Podcast: The Startups Rebuilding American Industry | The AD50 ·
Erin Price-Wright · Andreessen Horowitz (a16z)
The US power buildout gap with China is a question of national will, not technology; capital will follow if we grow.
Asked about the widely shared chart showing US power generation flat while China's rises, Erin argues the gap is not a technology problem. She says it requires national reflection on whether Americans want projects built and whether they want the country to grow. If the answer is yes, she believes the technology and capital will follow. Connor adds that this is 'an everything problem,' not just an energy one.
“I don't think that is a technology problem... We need to do some soul searching about our will as an American people... And if we do, I think the technology and the capital will follow.”at 45:06
a16z Podcast: The Startups Rebuilding American Industry | The AD50 ·
Danny Moses · Moses Ventures
Power and grid constraints, not data center approvals, are the bigger AI bottleneck; build pauses may help.
Danny argues the bigger issue for AI is not data center approvals. Approvals are unlikely in places like Central Park, but the real constraint is the power and grid associated with the data centers. He thinks the current halting of some build-outs is not negative, and may benefit companies by giving them time to work out overall demand. He expects power to remain an issue, and says local races for governors and mayors could make it a problem. Approvals could also be used as an excuse for missed projects.
“the bigger issue here for AI isn't data center approvals at the moment... It's the power and the grid associated with it. I actually think that this is not a negative thing right now, kind of the halting of some of these data center build outs.”at 23:55
On The Tape with Danny Moses: Lessons From The Great Financial Crisis | "Clear Takes" with Vincent Daniel and Danny Moses ·
Chris Wright · U.S. Department of Energy
DOE utility loans will strengthen grid reliability and reduce winter and summer power price spikes within one to three years.
Wright describes the DOE's roughly $5 billion loan to Alliant as capital infrastructure that will take one to three years to complete. He argues the grid must add available electricity and become robust enough for winter cold and summer peaks. He says the loans, made through the Energy Dominance Financing office, are low-risk so taxpayers get repaid, and that they will lower the price spikes that come with extreme weather. He contrasts this with the previous DOE, which he says ignored affordability.
“This is utility sector, as you mentioned, but this is about math and humans. It'll lower the price spikes that come when it gets very cold in the winter or very hot in the summertime.”at 4:32
Bloomberg Talks: Energy Secretary Chris Wright Talks Diesel Prices ·
Chris Wright · U.S. Department of Energy
America's grid will keep running mainly on natural gas, nuclear and coal for decades to come.
Wright says the US electrical grid runs today on natural gas, nuclear and coal, and he expects that to remain true decades from now. He presents this alongside a role for solar, batteries, fusion and geothermal. He cites a battery storage and LNG storage project in Wisconsin as a way to handle peak demand, which he says drives price.
“America's electrical grid today runs on natural gas, nuclear, and coal, and that'll be true decades in the future as well.”at 11:57
Bloomberg Talks: Energy Secretary Chris Wright Talks Diesel Prices ·
Brett Winton · ARK Invest
Permitting, shell construction, and turbine problems leave AI chips installed but not powered.
Brett says the reported stranded compute reflects a real bottleneck. He thinks that, on a project-by-project basis, a combination of permitting, building the shell, and getting turbines working as intended leaves NVIDIA chips in inventory but not switched on. He calls this probably true to some degree.
“A combination of permitting and building the shell and getting the turbines working the way it's supposed to is causing chips to be in inventory but not turned on.”at 23:18
ARK Invest (FYI podcast, In the Know video): Anthropic’s S-1, Meta’s $90B, And Elon’s Warning Call | The Brainstorm 155 ·
Lisa Shalett · Morgan Stanley Wealth Management
AI capex bottlenecks are moving to power and electricity generation, and to politics that could slow buildout.
Shalett says that while capex dollars keep being spent, the bottlenecks are shifting from chips to power and electricity generation. Political bottlenecks are also a concern, both for whether planned capex can be built and turned on. She says the next 24 months will show how much of this actually lights up. If buildout is slower than expected, that could reinforce scarcity and pricing power across the supply chain.
“you absolutely are gonna have the picks and shovels. But the bottlenecks are moving to power, right, to electricity generation, and that's really where we're gonna have some real questions.”at 26:42
Bloomberg Surveillance: Bloomberg Surveillance TV: October 28th, 2026 ·
Tyler Rosenlicht · Cohen & Steers
Data-center reliability needs are pushing the system back toward traditional power and storage.
Data centers cannot afford to lose power, because an outage leaves an expensive shell with no economic value. This has pushed focus back toward traditional energy, which runs continuously, and alternatives are intermittent. The system is currently trading some cleanliness for reliability. The team argues that investment in alternative battery storage and other measures is needed to make the alternative side more reliable.
“Right now, it's let's trade off a little bit of cleanliness for reliability because that's what the economy is focused on. Let's reinvest and increase investments in alternative battery storage and other things to harden the alternative space.”at 5:31
UBS On-Air: Talking Markets (Powering the AI revolution) with Tyler Rosenlicht (Cohen & Steers) ·
Tyler Rosenlicht · Cohen & Steers
Grid and utility infrastructure investment is underappreciated and still very early; the team likes builders.
The capital needed for power generation and transmission for utilities, here and around the world, is immense. The team thinks the market still underappreciates this build-out. Electrification and growth are still in very early stages. The team likes the companies in that value chain, including those that build transmission lines, lay pipelines and build utility infrastructure.
“we do think the infrastructure build out in the energy system is still underappreciated. The amount of capital, CapEx, generation transmission required for utilities here and around the world is immense”at 8:32
UBS On-Air: Talking Markets (Powering the AI revolution) with Tyler Rosenlicht (Cohen & Steers) ·
Andrew McDermott · Mission Value Partners
US power generation was flat for twenty years and must now be rebuilt to compete with China.
He says the US had flat power generation for twenty years while adding no new capacity. China is scaling its power supply up eightfold, so the US now needs to rebuild to compete. He contrasts this with Europe, which is taking its grid apart and so cannot have a serious conversation about these issues.
“America wasn't, and we did... We had flat power generation for twenty years.”at 13:04
Top Traders Unplugged: GM109: Why Buffett Finally Went All In on Japan ft. Andrew McDermott ·
Craig Fuller · FreightWaves
Without new US power generation, AI, robotics and EV growth stall; data centers fund the build-out.
Craig says utilities need dependable customers to finance new nuclear or natural gas plants, and hyperscalers have the credit to be underwritten by banks to serve as that base customer. He argues that without new US power generation there will be no AI data economy, robotics economy or EV adoption. Electrifying every US vehicle would need more power than all AI data centers consume.
“If we do not create electricity and power generation in The United States, we are not going to have not just an AI data economy. We're not gonna have a robotics economy”at 34:02
Thoughtful Money: "No Signs Of The A.I. Party Slowing Down" | Craig Fuller ·
Scott Galloway · NYU Stern School of Business
Power, not chips, is now the bottleneck for AI, so the constraint is electricity and data-centre capacity
Galloway says the scarce input has moved from chips to electricity. Amazon stockpiled GPUs, which he calls a smart move by one of the best supply-chain operators in history, but the constraint turned out to be power and data-centre capacity. He concludes that the market has too little power rather than too much supply, which is why the chips are sitting idle and why construction delays are hurting. This is the basis for his view that the buildout is delayed rather than oversupplied.
“The bottleneck has moved from chips to electrons.”at 53:18
Prof G Markets: The $3.5T Bailout: Elon's Plan To Merge Tesla And SpaceX ·
Patrick Boyle · Patrick Boyle On Finance
Data centers aren't paying their own grid costs; ratepayers absorb future PJM capacity charges.
Boyle reports that EPRI found data centers built from 2019 to 2024 lowered power costs by about 6% in the typical state, but only because they used spare grid capacity that covered fixed costs. That spare capacity is now largely used up. The PJM Independent Market Monitor estimates about $6.3 billion of future capacity costs come from planned data centers, and he says the data centers are not paying those costs themselves. Boyle adds that if expected demand never arrives, utilities recover fixed costs from fewer customers, so the remaining customers pay for transmission lines. He treats both the earlier benefit and the later burden as true.
“...the Independent Market Monitor estimates that about $6,300,000,000 of future capacity costs come from planned data center costs”at 5:56
Naveen Rao · Unconventional AI
Data centers are paying well above grid rates for power via bespoke behind-the-meter deals, creating two-tier pricing.
He sees anecdotal evidence, without hard numbers, that data centers pay far more per kilowatt-hour through bespoke behind-the-meter deals than grid users pay. Data center compute margins are high enough to absorb 30–40% higher electricity costs, so the value of power to data centers exceeds its value for general use. He cites Bloom Energy's proposed multi-gigawatt Arizona deal as a possible technology path but says it is unproven and its economics differ from the grid.
“I do see evidence through anecdotes at least. I don't have specific numbers to quote around this because every one of these behind the meter deals is bespoke, and people are paying way more per kilowatt hour than they would off the standard grid.”at 9:53
20VC: 20VC: Why Robotics Will be 100x Bigger than Coding Agents | Inference Will Be a $10TRN Market: What No One Sees | Why Software Investors are F***** Today with Naveen Rao, Unconventional ·
Andy Jassy · Amazon
US underinvested in grid infrastructure; AI-driven build-out should leave power rates flat or falling after capacity comes online.
Jassy says energy rates have risen in some data center regions and fallen in others, but the real issue is that the US has not invested in its grid infrastructure as needed. Meaningful load growth was going to require that investment, and AI's rapid growth is what pushed it over the line. Once data center capacity is built, he expects rates to flatline or fall in most regions, as they have in Indiana, Georgia, and Mississippi.
“some of the energy rates have gone up. Some of the energy rates have gone down. But the real issue is that as a country, we have not invested in the infrastructure of our grids the way we've needed to”at 19:06
a16z Podcast: Amazon CEO Andy Jassy on Why Amazon Is Investing $220 Billion in AI ·
Sam Unsted · Bloomberg News
Building enough power capacity to connect AI data centres to the grid is almost impossible, causing delays.
Sam argues that the energy demands of the AI build-out are colliding with grid limits. He says it is almost impossible to build enough energy capacity to connect all these data centres to the grid. That creates delays. Those delays raise the question of whether the build-out can happen on the timescale investors expect, given the hundreds of billions being spent, and the neoclouds are likely to feel the first pushback.
“...which then, as you actually just mentioned as well, energy, just quickly as well, is then bumping up against the, you know, it's just, it is, I would say, almost impossible to build enough energy capacity to actually connect all of these things to the grid as well. So then that creates delays.”at 22:17
Merryn Talks Money: French Bond Markets and AI Neoclouds Feel The Pressure ·
Chris Harvey · CIBC Capital Markets
Utilities were upgraded as bond-like defensives and offer both offense and defense, though not a long-term secular story.
Harvey upgraded utilities two weeks ago because they are bond-like and hard to lose money on over five to ten years. They had underperformed about 30% since March, creating an entry point. He sees them as playing both offense and defense, and says they are beginning to work, though they are not a long-term secular story.
“We like the utility space. It is beginning to work, but is it a long term secular story? No, not really.”at 18:22
CNBC Fast Money: Tech Tumbles In Market Sell-Off… And Moderna’s Meltdown After Data Leak 10/22/26 ·
Josh Brown · Ritholtz Wealth Management
Power and component suppliers at the tail of the AI capex cycle are exposed to any plateau in spending.
Brown says suppliers of components and power sit at the tail end of the capex chain. They depend on more and more capex. Meta held its capex outlook rather than raising it, so the spending is levelling off rather than accelerating. Vernova and Vistra fell about 11% and 10%, which Brown ties to that dependence on continued expansion.
“if you're at the tail end of the caboose, you're the companies supplying the components or building the power and counting on more and more and more and more CapEx as far as the eye can see.”at 45:34
CNBC Halftime Report: The AI Race Deepens as Access to Capital Shrinks: Your Next Move 10/22/26 ·
Sylvia Jablonski · Defiance ETFs
Data-center and AI power demand over the next year or two is so large it equals about 20 New York cities.
Jablonski argues that the AI power and infrastructure trade is attractive because demand is huge. She says the power needed for data centers and AI infrastructure over the next year or two equals about 20 New York cities. That demand backs the power trade as a whole.
“[We] need 20 New York cities to power the amount of demand that we have for the next year or two for data centers”at 10:55
CNBC Power Lunch: Cracks in the AI Trade, Tech Earnings Preview, Surging Airfares 10/20/26 ·
Chris Grisanti · MAI Capital Management
The power trade for AI will outlast the picks-and-shovels trade, given a five-to-ten-year buildout.
Grisanti argues that power demand for data centers and AI infrastructure will last longer than the chip-and-hardware trade. His reason is that it is a five-to-ten-year build-out. He sees the power trade as the longer-lasting part of the AI infrastructure boom.
“the power trade is gonna last longer than the picks and shovels AI... Because it's a five to ten year build out.”at 11:00
CNBC Power Lunch: Cracks in the AI Trade, Tech Earnings Preview, Surging Airfares 10/20/26 ·
Paisley Nardini · Tema ETFs
US grid underinvestment is coming due; regulated utilities with contractual cash flows look constructive.
Nardini argues the US has underinvested in infrastructure for several decades and that reckoning has arrived. The timing coincides with AI and the power needs of data centers, which is why the issue is now getting attention. Grid-building companies are often government entities or regulated utilities, and she is more constructive on that part of the market because they have contractual cash flows from regulated industries that support the buildout. She frames the need as true irrespective of AI.
“we're actually kind of more constructive on that part of the market because you have contractual cash flows from some of these regulatory industries that are helping support and build out this power play”at 44:31
Excess Returns: The S&P 500 Is Only 50 Stocks. 70% of the Index Is in a Correction. Did the Sell-Off Already Happen? ·
Chamath Palihapitiya · Social Capital
US electricity deficit makes power the ultimate AI moat; hyperscalers will not supply startups.
Chamath argues the US is already behind and in a massive energy deficit, which he puts at about 1.4 (terawatt-hour, as stated) short by 2050, even if planned builds finish on time and budget. Load growth from AI would make it worse. As a result, hyperscalers can no longer be counted on to supply power for new AI products, and startups will be turned away for lack of capacity. He argues the best moat is owning the land, power and shell, because power is such a constrained asset.
“I'll give you the ultimate moat. I think power will be the moat. I think that you can build the most fantastical things in the world and you're gonna see people on the dry docks basically trying to hitchhike to get power.”at 38:19
All-In Podcast: CoreWeave Crash, AI Trade Cracks, Anthropic IPO Red Flags, Bill Gates Doomer ·
Yasir Arafat · Aalo Atomics
Data centers need about 125 GW by 2030, but the grid can supply only about 25 GW, a 100 GW gap.
Yasir cites Bank of America's estimate that AI data centers need about 125 GW between 2026 and 2030, and he says each gigawatt powers about a million homes. He believes the grid and other suppliers can pull together only about 25 GW, leaving a roughly 100 GW generation gap. He says this is why the data center market is so large and why reliable on-site power matters, since grid strain would make data centers a burden on communities.
“Bank of America, for example, they said, hey. We need a 125 gigawatts... from this year to 2030, we need 125 gigawatts. And right now, I believe the grid and everybody else, they can pull together about 25 gigawatt or so.”at 12:49
MacroVoices: MacroVoices #554 Yasir Arafat: Mass Manufacturing Nuclear Power Plants ·
Yasir Arafat · Aalo Atomics
Gas turbines will become obsolete once nuclear volume production drives costs down.
Yasir argues that once volume production brings costs down far enough, nuclear reactors become an almost obvious choice for power, and gas turbines will become obsolete. The argument depends on his own cost projections: first-of-a-kind power in the mid-teens cents per kWh falling below 10 cents with volume, and reaching parity with coal and gas.
“when we unlock the volume production, costs will come down so low, then this technology will become almost a no brainer to go after. At that point, gas turbines will truly become obsolete”at 25:09
MacroVoices: MacroVoices #554 Yasir Arafat: Mass Manufacturing Nuclear Power Plants ·
Trina Chen · Goldman Sachs
Chinese transformers are unlikely to enter US or European main grids because of security concerns.
Chen relays the view of her power equipment analyst. Chinese transformers have gained share in other markets, but the analyst is very skeptical they would ever reach the main grids in the US and Europe, for security reasons. This shows that some sectors face access limits beyond price and quality.
“she would be very skeptical that Chinese transformer would ever go to the main grid of power grid in US and Europe for security reason.”at 14:18
Goldman Sachs Exchanges: Where Chinese Exports Are Gaining the Most Market Share ·
Ariana Salvatore · Morgan Stanley
US data center power shortfall grows to 57 GW through 2028, up from 38 GW previously estimated.
Her colleague Steven Byrd and team updated projections and now see a potential 57 GW gap in US data center power through 2028, before accounting for new solutions that could accelerate access to power. That is up from their previous estimate of 38 GW. She cites this scale of the energy challenge as the reason the permitting proposal matters.
“They now see a potential 57 gigawatt gap through 2028, and that's before accounting for new solutions that could accelerate access to power. That's up from their previous estimate of 38 gigawatts.”at 1:12
Thoughts on the Market (Morgan Stanley): The Politics of Powering AI ·
8 older views on Power demand & the grid are on the episode pages, and all of them are a question away in your AI with Chorus.
Ask your AI about Power demand & the grid
Chorus gives Claude, ChatGPT and other assistants every view on Power demand & the grid from the market's podcasts. Ask who's worried, who isn't, who changed their mind, and why.
Chorus reports what speakers said. It isn't investment advice.