Is the AI Boom Heading for a Bust? What the Dot-Com Bubble Tells Us

Summary

WSJ Tech News Briefing: Is the AI Boom Heading for a Bust? What the Dot-Com Bubble Tells Us

Release Date: March 12, 2025

In the March 12th episode of WSJ Tech News Briefing, hosted by Shara Tipkin, the Wall Street Journal delves into the burgeoning impact of artificial intelligence (AI) on data centers and draws parallels between the current AI boom and the infamous dot-com bubble of the early 2000s. Through expert insights and detailed discussions, the episode explores whether today's AI investments are paving the way for a revolutionary technological advancement or heading towards an inevitable market correction.

---

AI's Insatiable Appetite for Data Storage

The episode opens with an examination of how AI’s exponential growth is driving demand in the data center industry. AI requires vast amounts of data to train its models, leading to significant investments in data storage solutions. According to consulting firm IDC, data centers spent an estimated $40 billion on storage devices last year, with projections indicating a 31% growth over the next two years.

John Keilman, a WSJ reporter, provides an in-depth analysis of the storage landscape:

“[AI] requires stupendous amounts of data to train their models. ... all of these devices that are in a data center need to be able to hold more and more data.” (01:38)

Keilman explains the ongoing competition between traditional hard disk drives (HDDs) and solid-state drives (SSDs). While HDDs remain cost-effective, SSDs offer superior speed but at a higher price point. This competition is intensifying as both types of storage solutions strive to enhance their capacities to meet AI’s demands without expanding the physical footprint of data centers.

---

Seagate’s Innovation: Staying Relevant in a Shifting Market

To maintain relevance amidst the surge in AI-driven storage needs, companies like Seagate are innovating continuously. Keilman highlights Seagate’s advancement in hard drive technology:

“Seagate has come up with something called heat assisted magnetic recording... they are so magnetically powerful that you basically have to weaken their magnetism for a tiny portion of time so that the magnet can flip it up or down.” (02:27)

Seagate has recently introduced a hard disk drive capable of holding 36 terabytes of data, with expectations to reach 60 terabytes within a couple of years and potentially 100 terabytes before exploring new technologies. These innovations underscore HDDs' adaptability and economic efficiency, positioning them as indispensable components in modern data centers.

---

Drawing Parallels: AI Boom vs. Dot-Com Bubble

The discussion transitions to a historical perspective, comparing the current AI boom to the dot-com bubble that peaked and subsequently crashed around 25 years prior. Rolf Winkler, WSJ tech reporter, provides a nuanced perspective on this comparison:

“There's a lot of really promising opportunities... we're not really sure how we're going to use them all just yet. Is there revenue behind all this investment that justifies the investment? Not yet.” (05:11)

Winkler distinguishes between good and bad bubbles. He cites historical instances where speculative bubbles, driven by revolutionary technologies like canals, railroads, and electrification, ultimately led to significant productivity enhancements despite initial overenthusiasm and eventual market corrections. In contrast, bad bubbles, such as the housing crisis, do not contribute to long-term productivity.

---

Lessons from the Dot-Com Era: Innovation Beyond Failure

Reflecting on past failures, Winkler discusses companies like Pets.com and Webvan, which despite their collapse, paved the way for successful successors like Chewy and Instacart. He emphasizes that these early ventures were "just ahead of their time," lacking the necessary infrastructure and market readiness.

Moreover, Winkler highlights General Magic, a pioneer in smartphone technology from the early '90s. Although it failed commercially, many of its alumni went on to create foundational technologies in modern smartphones, showcasing how early failures can incubate future successes:

“The alums that came out of General Magic have helped power the smartphone revolution.” (09:30)

---

Current Market Dynamics: Winners and Losers

Winkler addresses the current market landscape, identifying Nvidia as a significant winner due to its pivotal role in supplying GPUs essential for AI data centers:

“Who won in the gold rush? Levi Strauss won selling everybody jeans. And whoever sold the picks and shovels, it wasn't necessarily gold miners. Nvidia is winning because everybody has to buy its chips to put in these data centers to power their models.” (10:11)

However, he questions the productivity of these investments and whether they will yield transformative AI products in the near future or if the most impactful innovations are still a decade away.

---

Conclusion: Navigating the AI Investment Landscape

The episode concludes by underscoring the importance of distinguishing between speculative excitement and sustainable technological advancement. While the AI boom mirrors the dot-com era’s exuberance, the potential for revolutionary productivity gains suggests that, despite possible market corrections, AI could ultimately drive significant economic and technological progress.

Shara Tipkin wraps up the briefing by acknowledging the contributions of the production team and teasing the upcoming segment, leaving listeners with a comprehensive understanding of the current AI landscape and its historical parallels.

---

Notable Quotes:

• John Keilman (01:38): “[AI] requires stupendous amounts of data to train their models. ... all of these devices that are in a data center need to be able to hold more and more data.”

• John Keilman (02:27): “Seagate has come up with something called heat assisted magnetic recording... they are so magnetically powerful that you basically have to weaken their magnetism for a tiny portion of time so that the magnet can flip it up or down.”

• Rolf Winkler (05:11): “Is there revenue behind all this investment that justifies the investment? Not yet.”

• Rolf Winkler (09:30): “The alums that came out of General Magic have helped power the smartphone revolution.”

• Rolf Winkler (10:11): “Nvidia is winning because everybody has to buy its chips to put in these data centers to power their models.”

---

This comprehensive summary encapsulates the key discussions and insights from the episode, offering readers a clear understanding of the current state of AI investments, storage technology advancements, and historical lessons from the dot-com bubble that inform today's tech landscape.

Transcript

[Comcast Business]

With leading networking and connectivity, advanced cybersecurity and expert partnership, Comcast Business helps turn today's enterprises into engines of modern business. Powering the engine of modern business Powering possibilities. Restrictions apply.

[Shara Tipkin]

Welcome to Tech News briefing. It's Wednesday, March 12th. I'm Shara Tipkin for the Wall Street Journal. Artificial intelligence is fueling a boom in data centers. But some traditional storage companies are scrambling to adap to AI's needs. How do hard drive makers keep up? Then we explore what the dot com bus can teach us about today's AI boom. WSJ reporter Rolf Winkler tells us all about bubbles. But first, AI requires a lot of data to work, and that information has to be stored somewhere. Data centers last year spent an estimated $40 billion on storage devices, according to the consulting firm IDC. That's expected to grow by 31% over the next two years. Storage companies like Seagate hope data centers choose their hard disk drives, but the industry is facing an uncertain future as competing solid state drives get faster and cheaper. Here to explain is WSJ reporter John Keilman. John hard drives have limitations like moving parts, but they're also fairly cheap. Solid state drives, or SSDs are faster, but they cost more. What's fueling this push to have better storage options?

[John Keilman]

Well, the push is coming from the demands of AI, which require stupendous amounts of data to train their models. And at the same time, all of these companies that have invested in data centers do not want to add to their footprint to their square footage to increase their ability to store data, which means all of these devices that are in a data center need to be able to hold more and more data. So both on the hard disk drive side and on the solid state drive side, both of them are trying to increase the capacity of their individual devices as much as they can.

[Shara Tipkin]

You spent a lot of time talking with Seagate. What is it doing to make sure hard drives stay relevant?

[John Keilman]

So Seagate has spent more than 20 years, and not just them, other folks in the hard drive industry as well, but they are sort of first to market with it. What they have done is come up with something called heat assisted magnetic recording. And what that does is it uses a laser that is smaller than a grain of salt to apply a nanosecond of heat to a bit. And the reason they need to do that is because these new materials that they're using that they can make smaller and smaller and smaller, they are so magnetically powerful that you basically have to weaken their magnetism for a tiny portion of time so that the magnet can flip it up or down. And so with that, Seagate just started shipping a single hard disk drive that can fit in the palm of your hand and it holds 36 terabytes of data. And they think they can get to 60 terabytes within the next couple years. And they think upwards of 100 terabytes in a single drive is possible before they need to start perhaps moving into different types of technologies.

[Shara Tipkin]

How will these changes impact data storage?

[John Keilman]

What these innovations mean for the hard drive industry is that they're still in the game. And the more that they can store at a sort of economically efficient manner, the more sort of indispensable they make themselves. And they can sort of put off that day if it ever arrives in which they become obsolete.

[Shara Tipkin]

That was our reporter John Keilman. Coming up, today's tech market is drawing some comparisons to the.com bust. What can we learn from the early 2000s? That's after the break.

[Comcast Business]

With leading networking and connectivity, advanced cybersecurity and expert partnership, Comcast business helps turn today's enterprises into engines of modern business. Powering the engine of modern business. Powering possibilities. Restrictions apply.

[Shara Tipkin]

25 years ago this week, the dot com era hit its peak and then it collapsed. Pets.com, webvan and other new Internet companies went bust. Today's AI boom is drawing some comparisons to the dot com bubble. WSJ tech reporter Rolf Winkler is here now to talk about the similarities and the differences. Rolf, why are people comparing today's market to the dot com bust?

[Rolf Winkler]

Probably just because there's a huge amount of investment going into a technology where the return is not yet clear. People are really excited about AI. There's a lot of really promising opportunities there. And you can see a world 10 years from now, 15 years from now where it changes everything, assuming that technology continues to progress. But right now we're spending hundreds of billions of dollars on basically GPUs, graphic processing units, Nvidia chips to fill out these cavernous data centers to power AI models that are cool and advancing, but we're not really sure how we're going to use them all just yet. Is there revenue behind all this investment that justifies the investment? Not yet.

[Shara Tipkin]

When you think about bubble, like I think that's bad, but you actually mentioned that there are good bubbles that fuel rapid adoption of revolutionary tech and then there are bad bubbles. Could you explain a little more the difference between good and bad bubbles?

[Rolf Winkler]

This is just sort of a fascinating thing I have learned about in recent years Talking to experts and they'll say, look, there are revolutionary technologies that have happened in history. Canals in the 1700s, railroads in the 1800s, electrification in the 1900s and the DOT com boom. These things led to speculative bubbles where people were so excited. They're like, oh my God, electrification will change everything. It really did. But it took longer to establish itself than maybe all the early excitement. And so for all of these things, the early excitement can overshoot. But what it also can do is it can be an organizing function, it can bring a lot of people together, it can lead to a lot of capital investment and basically a whole bunch of people end up marching in a direction because they believe in a technology. And even if there's a big speculative bubble and it crashes, it's a productive technology, something that adds to the economy's productivity and in that way still takes over. It takes more years than maybe people projected at the beginning. And so you get that kind of crash, but a rebound. Now you compare that, that's a good bubble, as some experts have told me. And AI would seem to fit that. We're investing in productive assets that could deliver a very amazing future. Now there's bad bubbles, obviously, the housing crisis, housing's great, but overbuilding, real estate bubbles, that's not something that is going to lead to a productivity boom later. So that can be a bad bubble.

[Shara Tipkin]

What are some examples of companies that went bust but influence products and services we still use today?

[Rolf Winkler]

What's funny is two are the most famous dot com flameouts, pets.com and webvan. Right. Delivering pet food to your door and delivering groceries. And they were widely mocked at the time and they achieved big valuations. And look at Chewy today. Instacart. Collectively the two companies are worth 25 billion doing exactly the same thing. Pets.com and Webvan had the right idea. They were just ahead of their time.

[Shara Tipkin]

Yeah, just too early. Another company you talk about is General Magic. Tell us about that.

[Rolf Winkler]

In the early 90s you had this group of people get together that basically built the first smartphone. But the problem was in the early 90s there was no Internet. Really. Of course there was an Internet. But you know, Netscape hadn't been invented yet, so there was no web browser. People, regular people, were not using the Internet. There was not much content. We were all on very slow dial up modems and interactive touch screens had not been invented. All of these things had not yet happened to make smartphones useful for consumers. It doesn't mean they weren't really smart and ahead of their time. And so by the time Steve Jobs releases the iPhone, 13 years later, responsive touchscreens are invented. The Internet is widely used. Broadband is a thing. All of this. People are using the Internet widely.

[Shara Tipkin]

Batteries are smaller.

[Rolf Winkler]

Batteries are smaller. Flash memory is cheaper. All these things make the phone possible, obviously chips. So what's really ironic is all of the alums that came out of general magic that have helped power the smartphone revolution. A couple names. Tony Fadell goes to Apple. He helps invent the ipod and later the iPhone. Andy Rubin, who founded Android, world's largest mobile operating system. And then another one, John. John Andrea. He was working on AI agents. He now runs AI at Apple. That idea was 40 years ahead of its time.

[Shara Tipkin]

Totally. What do investors and experts think will be the kind of winners and losers of this current market that we're in right now?

[Rolf Winkler]

It's so hard to say Nvidia is a winner because picks and shovels, I mean, who won in the gold rush? Levi Strauss won selling everybody jeans. And whoever sold the picks and shovels, it wasn't necessarily gold miners. Nvidia is winning because everybody has to buy its chips to put in these data centers to power their models. But the question is, is that productive? Is that going to lead to amazing AI products? Maybe. But maybe not soon. Maybe the most amazing ones are still going to take 10 years.

[Shara Tipkin]

That was our reporter Rolf Winkler. And that's it for Tech News Briefing. Today's show was produced by Julie Chang with supervising producer Kathryn Milsop. Sharon, I'm Shara Tipkin for the Wall Street Journal. We'll be back this afternoon with TNB Tech Minute. Thanks for listening.

[Comcast Business]

With leading networking and connectivity, advanced cybersecurity and expert partnership, Comcast business helps turn today's enterprises into engines of modern business. Powering the engine of modern business powering possibilities, restrictions apply.