Prof G Markets: AI is Running Up America’s Energy Costs — Who’s Footing the Bill?

Date: October 20, 2025
Hosts: Scott Galloway & Ed Elson

Episode Overview

In this episode, Scott Galloway and Ed Elson dive deep into how the frenzied expansion of AI—and the ensuing explosion in data center construction—is driving America’s energy demand, with power costs skyrocketing in key regions. The hosts question if the necessary grid buildout is realistic, who ultimately foots the bill when the electricity tab comes due, and whether technical breakthroughs or government intervention will be needed to keep AI’s dream alive (and affordable). They also discuss the rising backlash against data center expansions, the secondary effects on communities, and broader implications for the tech sector. In the latter part of the podcast, the conversation turns to tech companies’ (lackluster) steps on AI and child safety, plus sharp insights on the recent surge in publicly traded space-tech stocks.

Key Discussion Points & Insights

1. AI Data Center Boom and Soaring Energy Demands

(Segment Begins ~07:00)

• Unprecedented AI Compute Demand:
  AI’s hunger for compute is spurring a dramatic buildout of data centers. S&P Global projects grid demand from these centers will rise 22% in 2025 and triple by 2030—far outpacing the energy buildout rate.

• Staggering Cost Increases:
  Power prices near key data hub locations have soared up to 267% over five years (07:07), impacting household budgets and triggering local opposition.

  “The demand for energy because of AI is way outstripping supply…energy costs are only going to go way up.”
  —Ed (10:42)

• OpenAI’s Mega-Ambition:
  OpenAI’s aim to build out 250 gigawatts (GW) of data center capacity is roughly a quarter of the entire US electric grid, and would require 250 nuclear plants costing $12.5 trillion (10:42).

• Chronic Underinvestment in Energy:
  Despite billions committed to data center construction, grid upgrades and new power generation are lagging, causing years-long delays and grid connection bottlenecks.

2. Who Pays? — The Public, or Big Tech?

(Segment Continues ~15:00–18:00)

• Trickle-Down Costs:
  Utility bills for regular Americans are surging where data centers proliferate, but shareholders and big tech companies remain insulated—for now.

  “At some point does the government come in...do we come up with some sort of tiered pricing system where we charge a surcharge…if you’re purchasing over a certain amount [of energy]?”
  —Scott (15:04)

• Political Incentives and Capture:
  Data center projects are political win-wins for ribbon-cutting until energy bills arrive; politicians risk backlash if constituents see only costs and no benefits.

• Middle Class Squeeze:
  The hosts argue that infrastructure investments (power, housing, transit) do more for working Americans than conventional tax cuts. But in practice, grid enhancements might mostly benefit AI shareholders, not households.

3. Local Backlash Against Data Centers

(18:19–21:24)

• Community Pushback:
  $63 billion in planned U.S. data centers have been blocked by local opposition citing surging energy costs and limited job creation—data centers create far fewer permanent jobs than factories or stores. (18:19)

  “The data center is basically a giant robot…once it’s built, you just let it sit there…less than 100 permanent jobs.”
  —Ed (19:40)

• NIMBY vs. YIMBY:
  The classic “Not In My Back Yard” debate re-emerges, with communities arguing high electricity bills and minimal economic uplift outweigh the putative tech benefits.

4. Global & Policy Ramifications

(24:23–31:11)

• Gridlock in Energy Solutions:
  “Drill baby, drill” (natural gas) isn’t sufficient—max buildout rates for gas-fired capacity (<12.5 GW/yr) can’t begin to meet projected AI needs. Nuclear takes decades, and solar is seeing investment retreat due to policy changes (24:23–26:31).

  “OpenAI wants 250 gigawatts...the problem isn’t that we need energy supply, it’s that we need it now.”
  —Ed quoting team research (24:50)

• Technical Optimism and Skepticism:
  Scott floats the hope that a technological breakthrough could radically cut AI’s energy appetite or that AI’s future is overstated (26:31–30:13).

  “Sam Altman, in order to justify this valuation, is making these just enormous outlandish projections...maybe AI is more niche, doesn’t live up to the hype.”
  —Scott (26:31)

• Call for an ‘All of the Above’ Approach:
  Renewables (solar/wind) are portrayed as the quickest, cheapest new capacity, but fossil fuels aren’t vanishing. The U.S. is robustly energy independent, but China is making leaps in solar (30:13). The hosts bemoan politicization of solar in U.S. discourse.

5. Broader Tech and Societal Implications

(34:10–53:54)

Tech & Child Safety:

• Meta and OpenAI announce new (largely ineffective) parental controls and age gating, usually motivated by PR, not real safety.

• Companies are incentivized to maximize youth engagement—and are not naturally motivated to self-regulate.

  “If I didn’t know better, I would think there's a possibility here that these firms are more concerned with shareholder value than the wellbeing of our children.”
  —Scott (36:13)

AI and Synthetic Relationships:

• AI chatbots capable of erotic or pornographic interactions will further propel youth screen time, with possible deep negative effects on social development, motivation, and mental health.

  “We have linked shareholder value to teen depression. That is not good.”
  —Scott (41:10)

• Massive economic incentives for tech giants to pursue AI porn and synthetic relationships, which require serious regulatory intervention, age gating, and wider public debate.

Policy Solutions and Global Comparisons:

• Citing new bans in Norway (social media under 13, rising to 15) and Australia (under 16), the hosts urge robust U.S. regulation:

  • Age-gating social media
  • No phones in schools
  • Fines and loss of liability shield for algorithmic harms
  • Enforcement, not “hoping” for corporate virtue

  “If we’re waiting on the better angels of Sam Altman or Satya Nadella or Mark Zuckerberg to show up, don’t hold your breath…There have to be laws.”
  —Scott (47:21)

6. SpaceX Monopoly and the Public Space-Tech Frenzy

(55:15–64:22)

• Space as the Next Tech Boom:
  Hosts flag insane stock runs at AST Space Mobile (up 2,500% in 2 years) and Rocket Lab (up 600%), fueled by hype for “the next SpaceX”—even though only SpaceX is actually dominant.

  “To find the suppliers...the amount of capex about to go into space is only going to be—well, it’s not going to rival AI, but it’s going to be huge.”
  —Scott (58:24)

• Supply Chain Bets:
  Companies like Honeywell and Taiwanese component manufacturers are in the space (literally and financially) but haven’t had the same wild ride as the “story stocks.”

• Warning on Momentum Trading:
  Ed sounds a cautionary note on buying into public space stocks on hype alone, given their fragile fundamentals and tiny asset base compared to SpaceX (59:29–62:05).

Standout Quotes & Memorable Moments

• On AI’s Energy Appetite:

  “If you wanted to supply 250 gigawatts of power, you would need to build 250 nuclear power plants.”
  —Ed (11:22)

• On Local Data Center Pushback:

  “The data center is basically a giant robot...once it’s built, you just let it sit there and it’s only a few servicing people.”
  —Ed (19:40)

• On Synthetic Relationships and Teenage Motivation:

  “Sexual desire from young men is not a bug, it’s a feature...if it’s captured in an engine, it can fuel progress forward.”
  —Scott (40:05)

• On Tech Regulation:

  “There have to be laws. We live in a capitalist society...the CEOs of these companies will rationalize whatever increases shareholder value by a dollar every day. That is his first, second, and third priority. Quite frankly, that’s his job.”
  —Scott (47:21)

• On Politicizing Solar Energy:

  “Solar is woke. China understands the issues.”
  —Ed (30:09)

• On Stock Market Hype:

  “This stock performance is absolutely crazy...2,500% in the past two years. Rocket Lab up 600%. This is crazy town...these aren’t fully rational valuations.”
  —Ed (59:29)

Key Timestamps

| Timestamp | Topic | |-----------|-------| | 07:07 | AI compute boom and energy consumption | | 10:42 | OpenAI’s power projections & grid limitations | | 15:04 | Who foots the bill for rising energy? | | 18:19 | Local backlash against data centers | | 24:23 | Energy supply constraints—gas, nuclear, solar | | 26:31 | Potential for breakthroughs or overhype | | 30:09 | Politics of solar and renewables | | 34:10 | AI, mental health, and tech’s “child safety” efforts | | 41:10 | Shareholder value vs. teen well-being | | 47:21 | Need for law, not corporate virtue | | 55:15 | Space tech stocks: AST Space Mobile & Rocket Lab | | 59:29 | Warning on public stock “space play” mania | | 62:05 | Satellite supplier investments | | 64:22 | Week ahead predictions: CPI, Netflix, Tesla, etc. |

Summary

Scott and Ed deliver a bracing, data-rich critique of the energy realities behind AI’s meteoric rise, explaining how power-hungry data centers threaten to drive up household bills and test America’s energy infrastructure. They question market optimism, critique the lack of adequate policy or corporate responsibility—especially around child safety—and caution listeners about speculative mania in the public space sector. Ultimately, they argue that only bold policy, aggressive regulation, and urgent infrastructure investments can square the tension between tech’s ambitions and society’s needs.