AI Burns Energy. But Could It Save Even More?

Summary

---

WSJ Tech News Briefing

Episode: "AI Burns Energy. But Could It Save Even More?"

Date: September 30, 2025
Host: Peter Ciampelli
Guests: Sebastian Herrera (WSJ reporter), Amy Myers Jaffe (NYU research professor), Belle Lynn (WSJ)

---

Overview

This episode tackles two intertwined stories in tech: the shifting Seattle tech job market amidst major layoffs from giants like Amazon and Microsoft, and the environmental implications of AI’s rising energy demands. While data centers powering artificial intelligence consume growing amounts of electricity and water, research is revealing that AI could help us save even more energy in sectors like transportation, building management, and industrial design. The episode examines whether AI can ultimately deliver an energy net-positive – with insights from Amy Myers Jaffe, a leading NYU research professor.

---

Key Discussion Points & Insights

1. Seattle's Evolving Tech Economy

Guest: Sebastian Herrera, WSJ reporter
Timestamps: 00:34 – 05:27

• Pandemic Overhiring and the AI Boom

  • Amazon, Microsoft, and other tech giants massively increased hiring during the pandemic boom, expecting continued growth.
  • As the pandemic receded, many companies realized they had overextended, leading to significant layoffs.
  • The surge in AI has further spurred job cuts, as automation replaces certain roles and resources shift toward expensive data center expansion.
    • “AI, as we know, is replacing some jobs like with software development. And also… they are spending billions of dollars on data centers for the AI boom. And this very costly endeavor means that they have to save costs elsewhere.” — Sebastian Herrera (02:28)

• Economic Ripple Effects in Seattle

  • Loss of high-income tech jobs impacts local businesses, commercial real estate, and migration patterns.
  • Notably, more former tech workers are applying for service jobs, such as baristas.
    • “She told me that recently… she started to see people with Microsoft and other tech companies on their resumes applying to become baristas.” — Sebastian Herrera (03:26)

• Seattle’s Transition, Not Decline

  • Despite anxieties, the city isn’t entering a "doom loop" but faces a period of transition, with some laid-off workers launching startups.
    • “People are really figuring it out there.” — Sebastian Herrera (03:54)

• Shifts in Corporate Headcounts & National Trends

  • Even as Microsoft adds new employees and doubles its market cap, overall headcounts aren’t expected to rise—reflecting a trend across tech, not just in Seattle.
    • “There was once a time when Amazon was thought to become bigger than Walmart in terms of headcount, and that no longer appears to be the case.” — Sebastian Herrera (04:33)

---

2. AI’s Energy Footprint: Cost and Opportunity

Guest: Amy Myers Jaffe, NYU research professor (interviewed by Belle Lynn)
Timestamps: 06:20 – 11:19

AI’s Potential as an Energy Saver

• Offsetting Power Demands
  • While AI data centers use immense electricity, AI could save energy and fuel in other sectors—potentially offsetting or exceeding its own footprint.
    • “You could actually wind up having two to three times the energy we're using to do the AI be saved by applying the AI in critical sectors like buildings, airplanes, container ships, etc.” — Amy Myers Jaffe (07:09)

Application Examples

• Transportation Optimization

  • AI-driven route planning can significantly reduce fuel waste in cars, trucks, airplanes, and ships by minimizing congestion and optimizing deliveries.
    • “Some of the research shows that when we all drive to a mall or to this grocery store or whatever, that uses a lot of fuel. But when we let an Amazon or UPS or FedEx deliver those goods… that can be optimized by a program using AI.” — Amy Myers Jaffe (08:16)
    • Leveraging AI for demand prediction means companies can better stack shipments and reduce emissions from unnecessary over-manufacturing.

• Smarter Buildings and Greener Materials

  • AI systems in buildings adjust lighting, heating, and cooling to match occupancy, saving significant energy.
    • “These buildings now have occupancy sensors and they can calibrate how much air conditioning or lighting or heat system to provide in different rooms at different times…” — Amy Myers Jaffe (09:19)
  • AI accelerates the discovery of sustainable building materials, analogous to its use in drug discovery.
    • “…if you're trying to come up with a composite that could replace cement or make it greener, again, we save a lot of emissions.” — Amy Myers Jaffe (10:01)

Downsides & Cautions

• Privacy & Data Ownership
  • Mass sensor deployment and continuous data collection create substantial privacy concerns, especially as tech companies gain unprecedented insight into daily lives.
    • “Who owns that data, what can happen with that data? That could be a very slippery slope.” — Amy Myers Jaffe (10:26)
• Reliance on AI Accuracy
  • Critical tasks in energy and engineering demand extremely precise AI outputs; small errors can have serious consequences.
    • “…one millimeter of a mistake can be the difference between safely operating infrastructure and an explosion.” — Amy Myers Jaffe (11:02)

---

Notable Quotes & Memorable Moments

• On the paradox of AI’s energy use:
  “You could actually wind up having two to three times the energy we're using to do the AI be saved by applying the AI in critical sectors like buildings, airplanes, container ships, etc.”
  — Amy Myers Jaffe (07:09)

• On shifting tech job markets:
  “There is a lot of anxiety among tech workers there… but people are really figuring it out.”
  — Sebastian Herrera (03:54)

• On the privacy tradeoff:
  “That could be a very slippery slope… Who owns that data, what can happen with that data?”
  — Amy Myers Jaffe (10:26)

• On the future of tech jobs:
  “No other city in the US rests on two players like Seattle does with Amazon and Microsoft.”
  — Sebastian Herrera (04:50)

---

Key Timestamps

• 00:34 — Host introduction; Seattle tech layoff context
• 02:04–05:27 — Sebastian Herrera: reasons for layoffs, city impact, changing workforce trends
• 06:20 — Introducing energy and AI with Amy Myers Jaffe
• 07:09–08:53 — AI efficiencies in transportation and distribution
• 09:09–10:16 — AI in smarter buildings and sustainable materials
• 10:21–11:19 — Privacy concerns and risk of over-reliance on AI in critical infrastructure

---

Episode Tone & Language

The discussion is nuanced but pragmatic: optimistic about AI’s potential for good while clear-eyed about economic transitions and technological risks. Sources speak candidly about the challenges facing Seattle and the energy sector, mixing local anecdotes with broader trends and research insights.

---

Summary Takeaways

• Seattle stands as a microcosm for broader changes in tech employment and urban economies, shaped heavily by mega-corporations and the AI transition.
• Despite AI’s voracious appetite for energy, its broader deployment could drive massive efficiency gains—if privacy and safety issues are managed with foresight.
• The key future challenge: maximizing AI’s environmental benefits without trading away societal trust or safety.

---

Transcript

[A]

At the moment, AI can help you search for business advice, but it can't step in to help if things take a bad turn. But how about quickly finding an AI expert who can? Siemens xcelerator Ecosystem helps you connect with top industrial AI providers and find innovative solutions from a single trusted source. That's AI for real. From the global market leader in industrial AI, Siemens. Learn more on USA.Siemens.com AI.

[B]

Welcome to Tech News Briefing. It's Tuesday, September 30th. I'm Peter Ciampelli for the Wall Street Journal. A coffee shop in Seattle is getting more and more job applications from former tech workers. It's a sign of the times after some of the biggest tech employers there have cut workers and added fewer new positions. Decisions. We'll get into what that shift means for the city and other tech hubs like Seattle. Then the data centers that make AI work use a ton of energy and water and produce a lot of greenhouse gases. But research professor Amy Myers Jaffe says it could also save energy and fuel across a bunch of sectors. We'll explain how, but first, Seattle has been one of the premier tech hubs in the U.S. but the sector's been rocked by layoffs in the past few years, led by companies like Amazon and Microsoft. And there's now another change coming to Amazon's workforce. The company is the largest sponsor of H1B visas in the US with more than 14,000 of its employees approved for the visa in 2025. And earlier this month, Trump announced changes to the visa program, imposing a $100,000 fee for visa holders or their sponsors. Reporter Sebastian Herrera is here to tell us more about what layoffs by those big tech giants and the smaller local firms could mean for Seattle. Sebastian, can you explain what's happening with both Amazon and Microsoft? Why has there been such an increase in layoffs and slowed hiring?

[C]

Yeah, this really started to happen around 2022. And what happened is Amazon and Microsoft and a lot of these other tech companies, they overhire during the pandemic. They added thousands and thousands of people during this time of growth. But as the pandemic started to recede, a lot of companies started to cut back because they had overhired during the pandemic. And then after that, the AI boom came and these companies started to cut jobs because of that. Because AI, as we know, is replacing some jobs like with software development. And also both Amazon and Microsoft, they are spending billions of dollars on data centers for the AI boom. And this very costly endeavor means that they have to save costs elsewhere so these are some of the reasons why we've seen cutbacks at Amazon and Microsoft.

[B]

So how are the layoffs affecting the broader economy in Seattle?

[C]

As tech jobs went away, some of the things that we saw were, for example, popular shopping districts. There's been less spending there. Moving companies have seen less clientele in past year. Commercial real estate has really dropped in Seattle as offices sit empty. I talked to this one manager at a local coffee shop, and she told me that recently she started to see people with Microsoft and other tech companies on their resumes applying to become baristas. But Seattle is not necessarily a doom loop story. What's happening more there is like, it's a place that's really in transition. And as tech jobs go away, it's grappling with what that means and how that can move forward. Seattle has other strong industries like aerospace that are there, like with Boeing. So Seattle has plenty of positivity going on, but there is a lot of anxiety among tech workers there. There is a lot of people that have been laid off there in recent years who have had to figure out, you know, what to do next. We've talked to people that were laid off, for example, at Microsoft and have started their own companies. So people are really figuring it out there.

[B]

You mentioned in the story that Microsoft has added around 3,000 new employees since 2022, but in that same time period, they've doubled their market cap. I was wondering if you could explain what that trend could mean for the broader tech industry and if you're seeing that same trend at other companies or in other tech hubs like Seattle.

[C]

For me, what's interesting is there was once a time when Amazon was thought to become bigger than Walmart in terms of headcount, and that no longer appears to be the case. They've said that in, in the next few years, their headcount's actually going to decrease. And this is definitely a broader tech issue. Outside of Amazon and Microsoft, other large technology companies have cut back or said that their headcount's not really gonna go up. No other city in the US Rests on two players like Seattle does with Amazon and Microsoft. Amazon and Microsoft are about 40% of the tech jobs in Seattle.

[B]

And.

[C]

And we really don't see that anywhere else around America, even in San Francisco Bay Area, where tech is obviously king in Seattle. Amazon and Microsoft are such a big part of that local economy that what's playing out in Seattle is like a study in what happens when your city and economy is built on tech. And then what happens as tech jobs go away.

[B]

That was Wall Street Journal reporter Sebastian Herrera coming up. From reducing traffic to creating sustainable building materials, could AI actually end up saving more energy than it consumes? That's after the break.

[A]

At the moment, AI can help you search for business advice, but it can't step in to help if things take a bad turn. But how about quickly finding an AI expert who can? Siemens Xcelerator Ecosystem helps you connect with top industrial AI providers and find innovative solutions from a single trusted source. That's AI for real. From the global market leader in industrial AI, Siemens. Learn more on USA.siemens.com AI.

[B]

AI uses a huge amount of energy, but it also has the potential to save energy and fuel. Amy Myers Jaffe is a research professor in the Global Affairs Program at New York University's School for Professional Studies. The Wall Street Journal's Belle Lynn sat down with her to learn more.

[D]

Amy, contrary to this narrative that AI is a huge energy suck, you found in your research that AI might be able to balance out its own power demands. How is that possible?

[E]

Well, you have to balance out two things. Of course AI is going to have a burdensome electricity use, as we've all been hearing about, but it's also going to save us energy in many different ways. And not just saving us electricity, it's also going to save us fuel. Things like gasoline, diesel fuel, marine bunker fuel, airplane, jet fuel. And so when you add that all together, you could actually wind up having two to three times the energy we're using to do the AI be saved by applying the AI in critical sectors like buildings, airplanes, container ships, etc.

[D]

So speaking of doing something positive in the transportation business, you found that AI driven route planning can cut down on the amount of fuel required for vehicles, airplanes for ships. Why is AI an effective optimizer for transportation?

[E]

Well, if you think about your own life, you might be using Waze or some other kind of program when you're driving to avoid traffic. When you actually do stop and go driving in traffic, that wastes a lot of fuel. But there's also other variables as well. Some of the research shows that when we all drive to a mall or to this grocery store or whatever that uses a lot of fuel. But when we let an Amazon or UPS or FedEx deliver those goods to us or Instacart, that can be optimized by a program using AI and the companies are starting to take it to the next level, which is since I know roughly when you're going to order paper towels, then I can stack Those goods based on my projected purchasing patterns in a distribution center to minimize not only the distance that these goods have to go, but also not over manufacture, that also saves a lot of emissions.

[D]

And in other areas, you also find that AI can help buildings get smarter and raw materials get greener. What does that mean?

[E]

Let's talk about the two different things. I'm living in a smarter building. Many of us have experienced when you're going into large buildings that the lights go off when no one's in a room. So think about that. Across the entire building, these buildings now have occupancy sensors and they can calibrate how much air conditioning or lighting or heat system to provide in different rooms at different times and thereby only heat or air condition to a certain level. And then on the building materials. If you think about AI, and there's been a lot of, I think, press around using AI to speed up the process of figuring out what drug combinations would be the best way to cure diseases or the flu or whatever it is that they're applying across all the different possible combinations of medical input. You can do the same thing with building materials. So if you're trying to come up with a composite that could replace cement or make it greener again, we save a lot of emissions.

[D]

But what are the downsides of having so much AI in our daily lives?

[E]

There's some critical privacy issues. So if we start with all this sensor technology and we have these big technology companies, and they know when I'm leaving my house and they know when I'm putting on my TV and they know everything about everything I'm doing, who owns that data, what can happen with that data? That could be a very slippery slope. And then we need to understand the limitations of the AI, like what can it and can it not do accurately? Because if you're going to use AI to design a new oil well completion or to help you with a power station or something like that, I have to be really sure that the calculations that the AI is helping me do are really accurate. Because in the energy space, you know, one millimeter of a mistake can be the difference between safely operating infrastructure and an explosion.

[B]

That was NYU research professor Amy Myers. Javi. And that's it for Tech News Briefing. Today's show was produced by Julie Chang with supervising producer Chris Sinsley. I'm Peter Ciampelli for the Wall Street Journal. We'll be back later this morning with TNV Tech Minute. Thanks for listening.

[A]

At the moment, AI can help you search for business advice, but it can't step in to help if things take a bad turn, but how about quickly finding an AI expert who can? Siemens Accelerator Ecosystem helps you connect with top industrial AI providers and find innovative solutions from a single trusted source. That's AI for real. From the global market leader in industrial AI, Siemens. Learn more on USA.siemens.com AI.