Speakers compare AI boom to dot-com bubble, citing high costs and circular investment dynamics
Summary
Speaker_0 introduces the debate on whether the current AI enthusiasm constitutes a bubble, noting massive investments by Google, Microsoft, and others. The speaker defines bubbles as price rises driven by herd behavior rather than intrinsic value, drawing parallels to the 2008 housing crisis and early 2000s dot-com bubble. Key metrics cited include ChatGPT reaching 100 million users in two months and OpenAI's reported $11.5 billion quarterly loss in 2025. The speaker highlights high operational costs for GPUs and training, estimating GPT-4 training at over $100 million. A 'circular money glitch' is described where companies like Nvidia, OpenAI, and Oracle invest in and buy from each other, potentially inflating revenue figures. Finally, the speaker contrasts current profitable AI leaders with unprofitable dot-com era companies, suggesting that while hype may fade, viable business models will survive.
Key takeaways
Episode sections
- 0:00 – 1:17
Introduction to AI Hype and Bubble Concerns
Speaker_0 introduces the widespread use of AI tools like ChatGPT and Gemini, noting massive investment by companies like Google, Microsoft, and Elon Musk's ventures. The speaker highlights Nvidia's stock performance and poses the central question of whether the current AI enthusiasm constitutes a bubble.
- 1:17 – 3:18
Definition and History of Economic Bubbles
The speaker defines a bubble as a scenario where price rises due to herd behavior rather than intrinsic value, eventually popping when reality sets in. Historical examples cited include the 2008 housing crisis, Japan's economic stagnation, and the early 2000s dot-com bubble.
- 3:18 – 4:18
Reasons for AI Bubble Comparisons
The speaker explains that AI is compared to the dot-com era due to similar hype cycles and rapid adoption metrics. ChatGPT reached 100 million users in two months, significantly faster than TikTok, Instagram, or Facebook. However, the speaker notes that many AI companies, particularly OpenAI, are not yet profitable due to high operational costs.
- 4:18 – 6:22
High Costs of AI Operations
The speaker details the expensive nature of AI, citing costs for GPUs, training models, electricity, and cooling. It is claimed that training GPT-4 cost tens of millions to over $100 million. Additionally, OpenAI reportedly lost around $11.5 billion in a single quarter in 2025, raising concerns about sustainability.
- 6:22 – 9:14
Circular Investment Dynamics
The speaker describes a 'circular money glitch' where AI companies invest in each other and buy from each other, inflating revenue figures. An example is given where Nvidia invests in OpenAI, which spends on cloud services from Oracle, which then buys GPUs from Nvidia. The speaker questions how much revenue comes from external users versus internal circular flows.
- 9:14 – 11:02
Comparison to the Dot-Com Bubble
The speaker compares the current AI landscape to the early 2000s dot-com bubble, noting similarities in hype and investment. While the NASDAQ jumped over 600% during the dot-com peak with over 400 internet IPOs, current AI leaders like Nvidia and Microsoft are profitable, unlike many dot-com era companies.
- 11:02 – 14:05
Potential Bubble Burst and Market Impact
The speaker suggests that if an AI bubble bursts, it would primarily affect hype and investment flows rather than the technology itself. Unprofitable companies would struggle, while profitable survivors would endure. The speaker cites Amazon, eBay, and Cisco as dot-com survivors that became giants, advising investors to distinguish between hype and viable business models.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| NVIDIANVDA | equity | neutral | You might also see AI-related stocks like Nvidia going crazy over the past few years. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
ChatGPT reached 100 million users in two months.
Unidentified speaker (speaker_0) ·3:57
How to verify: Check OpenAI's official user milestone announcements or press releases from early 2026. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
TikTok took about nine months to reach 100 million users.
Unidentified speaker (speaker_0) ·3:57
How to verify: Review TikTok's historical user growth reports or third-party analytics from 2026-2026. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Instagram took two and a half years to reach 100 million users.
Unidentified speaker (speaker_0) ·3:57
How to verify: Check Instagram's historical growth data or Meta's investor presentations from 2026-2026. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Facebook took more than four years to reach 100 million users.
Unidentified speaker (speaker_0) ·3:57
How to verify: Review Facebook's historical user growth data or Meta's investor presentations from 2026-2026. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Training GPT-4 may have cost tens of millions of dollars and possibly over $100 million.
Unidentified speaker (speaker_0) ·5:15
How to verify: Search for industry reports or estimates from tech analysts regarding GPT-4 training costs; note that OpenAI has not publicly disclosed exact figures. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
OpenAI lost around $11.5 billion in a single quarter in 2025.
Unidentified speaker (speaker_0) ·6:02
How to verify: Check OpenAI's financial disclosures or reports from verified financial news sources for Q1 2025. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Over 400 internet companies went public during the dot-com bubble peak.
Unidentified speaker (UNKNOWN) ·9:40
How to verify: Review historical IPO data from the NASDAQ or SEC EDGAR filings for the 2026-2026 period. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Primary sourceThe NASDAQ jumped by over 600% in total during the dot-com bubble.
Unidentified speaker (UNKNOWN) ·9:40
How to verify: Check historical NASDAQ Composite index data from 2026 to 2026. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Amazon, eBay, and Cisco were dot-com bubble survivors.
Unidentified speaker (speaker_0) ·12:27
How to verify: Verify current market status and historical stock performance of these companies post-2026. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Pets.com earned hundreds of thousands but spent millions and failed.
Unidentified speaker (speaker_0) ·12:27
How to verify: Review SEC filings or historical news reports on Pets.com's financials and bankruptcy in 2026. For recent events, verify 2026-08-09 through 2026-11-09. Use another period only when the speaker explicitly names it.
Primary sourceDo your own research
- Verify OpenAI's 2025 financial statements for the reported $11.5 billion quarterly loss.
- Check Nvidia's investor relations page for details on investments in OpenAI and other AI firms.
- Review Oracle's earnings reports to confirm revenue from AI cloud services and GPU purchases.
- Analyze ChatGPT's user growth metrics against historical data for TikTok, Instagram, and Facebook.
- Compare current AI company profitability ratios with dot-com era companies using SEC filings.
- Research the definition and historical performance of the NASDAQ during the 2000-2002 dot-com bubble.
- Investigate the operational costs of GPU training and electricity consumption for major AI models.
- Look for counter-arguments regarding the sustainability of AI revenue models from independent analysts.
Full transcriptexpand
- 0:00
- Unidentified speaker (speaker_0): Hey there! I believe most of you have heard of AI or even used it, like ChatGPT, Gemini, and others. Maybe you use it to help with your tasks, ask questions, make pictures, create videos, and some people are even getting married to
- 0:17
- Unidentified speaker (UNKNOWN): it.
- 0:18
- Unidentified speaker (speaker_0): Everything feels AI hyped right now. Many companies are starting to build their own AI too. Google, Microsoft, Elon, and many others are making it, using it, and pouring billions of dollars into it. You might also see AI-related stocks like
- 0:38
- Unidentified speaker (speaker_0): Nvidia going crazy over the past few years. Investors are happy. So they pour in even more money. And then suddenly you stop and think, wait, hold on. Isn't this starting to feel like a bubble? Everyone seems way too excited and is
- 0:58
- Unidentified speaker (speaker_0): throwing billions into AI almost blindly. People have been saying this for a long time, and now you're probably wondering is it really a bubble? Well, in this video, we're going to talk about how bubbles really happen. Why so many people say AI is a
- 1:17
- Unidentified speaker (speaker_0): bubble, the infinite circular money glitch, whether this is the dot-com bubble 2.0, and finally, whether AI will really burst and crash. Alright, let's get started. Section 1 How Bubbles Really Happen. So when
- 1:37
- Unidentified speaker (speaker_0): everyone says AI is in a bubble, some of you might ask, what is a bubble? A bubble is like this. Everyone starts buying the same thing because they think it will be super valuable later. At first, only a few people buy it. Then,
- 1:57
- Unidentified speaker (speaker_0): more people copy. Soon, everyone is buying it just because everyone else is buying it. Nobody really stops to ask if it's actually worth that much right now. So the price goes higher and higher, even though the real value hasn't
- 2:16
- Unidentified speaker (speaker_0): changed much yet. And when people finally realize it's not as amazing as they imagined, they stop stopped buying. That's where. when the bubble pops. And bubbles have happened many times before. In fact, many major economic crises were also linked to
- 2:36
- Unidentified speaker (speaker_0): bubbles. For example, the 2008 housing bubble, when everyone rushed to buy houses, and it eventually popped. It played a major role in causing a global crisis. There was also Japan's economic bubble, which is one reason its economy stayed
- 2:56
- Unidentified speaker (speaker_0): weak for decades.
- 2:58
- Unidentified speaker (UNKNOWN): And
- 2:59
- Unidentified speaker (speaker_0): of course, the dot-com bubble in the early 2000s, which was driven by internet hype. So bubbles have happened many times in many countries and with many different technologies. Section 2. Why so many people say AI is a bubble?
- 3:18
- Unidentified speaker (speaker_0): Mostly because it feels very familiar. People remember what happened during the dot-com era when the internet was new and exciting. Back then, prices went up fast, excitement was crazy, Easy, and everyone believed the future had already arrived.
- 3:38
- Unidentified speaker (speaker_0): That memory never really disappeared. And because both are transformative technologies with similar hype stories, people connect AI to the dot-com bubble. For example, ChatGPT needed only two months to reach 100 million
- 3:57
- Unidentified speaker (speaker_0): users, much faster than platforms like TikTok, which needed about nine months, or Instagram, which needed two and a half years, or Facebook, which needed more than four years. AI tools are being used by professionals and regular people across many industries.
- 4:18
- Unidentified speaker (speaker_0): So you might think if platforms like Facebook, Instagram, and Google made tons of money, then AI tools like ChatGPT should make even more money, right? Well, that's the problem. Many AI companies are still not clearly profitable,
- 4:38
- Unidentified speaker (speaker_0): especially OpenAI, the company behind ChatGPT. AI is extremely expensive to run. Training models, buying chips, running servers. paying engineers, and doing constant research all cost huge amounts of money.
- 4:58
- Unidentified speaker (speaker_0): AI needs GPUs, which are special machines that handle the heavy calculations, and that is super expensive. Some estimates suggest training GPT-4 may have cost tens of millions of dollars and possibly.
- 5:15
- Unidentified speaker (speaker_0): over $100 million. And that's not even the end of the cost. Every time you use ChatGPT online, all the calculations happen on OpenAI's servers. That means electricity, cooling systems, often water-based, water based, and
- 5:35
- Unidentified speaker (speaker_0): maintenance all add to costs at
- 5:38
- Unidentified speaker (UNKNOWN): the
- 5:38
- Unidentified speaker (speaker_0): same time. time, users expect AI to be cheap or even free. For example, when many people say something simple in Chat GPT, like thank you, actually cost millions for OpenAI. Of course, one thank you costs almost nothing, but billions
- 5:58
- Unidentified speaker (speaker_0): of them add up to millions of dollars.
- 6:01
- Unidentified speaker (UNKNOWN): Some
- 6:02
- Unidentified speaker (speaker_0): reports and estimates suggest OpenAI may have lost around $11.5 billion in a single quarter in 2025. With this situation, it starts to make sense why so many people call this an AI bubble.
- 6:22
- Unidentified speaker (speaker_0): Section 3: The Infinite Circular Money Glitch.
- 6:27
- Unidentified speaker (UNKNOWN): One
- 6:27
- Unidentified speaker (speaker_0): of the big reasons people worry about an AI bubble is something called circular investment. The idea is simple. Big AI companies invest in each other, then buy from each other, which makes growth look much bigger than it really is. Here's an
- 6:47
- Unidentified speaker (speaker_0): oversimplified example with numbers. so it's easy to see. Remember, this is an oversimplified version.
- 6:55
- Unidentified speaker (UNKNOWN): So
- 6:56
- Unidentified speaker (speaker_0): a chip company like Nvidia designs the most important chips for AI. An AI company like OpenAI needs those chips to run ChatGPT and train new models. So let's say NVIDIA invests $10 billion into OpenAI. OpenAI, either
- 7:16
- Unidentified speaker (speaker_0): directly or through long term deals and partnerships. OpenAI now has more money and more computing power, so it spends a large part of it on cloud computing for more powerful AI. Now look at where that money goes.
- 7:35
- Unidentified speaker (speaker_0): OpenAI spends $8 billion on cloud. services from companies like Oracle or other AI cloud providers. Those cloud companies use that $8 billion to buy GPUs and hardware, mostly from NVIDIA. So a
- 7:54
- Unidentified speaker (speaker_0): big part of the original $10 billion dollars flows back to NVIDIA as chip sales. So NVIDIA sees higher revenue, OpenAI shows growth, cloud companies see higher sales. On paper, everyone looks like they're winning. And
- 8:14
- Unidentified speaker (speaker_0): again, this is an oversimplified version because the real scenario looks like this. Yeah, quite different, and more complicated. But the idea is the same. From the outside, valuations go up and revenues look strong. But the
- 8:34
- Unidentified speaker (speaker_0): uncomfortable question is this: how much of this money comes from real users and businesses hanging from outside the system? And how much is just the same tens of billions of dollars moving in a circle? If one part of that loop slows down or
- 8:53
- Unidentified speaker (speaker_0): stops, the pressure spreads quickly to everyone else. Even though big companies like Nvidia, Oracle, and Microsoft are quite diversified beyond AI, the idea of money just moving around still makes people worry.
- 9:14
- Unidentified speaker (speaker_0): Section 4. Is this dot com bubble 2.0? So, as I mentioned before, in the early 2000s, there was a bubble called the dot com bubble. The internet was the new hype, and investors threw money at almost any company with dot com
- 9:34
- Unidentified speaker (speaker_0): in its name, even if they it had no profits and no real business model.
- 9:40
- Unidentified speaker (UNKNOWN): At
- 9:41
- Unidentified speaker (speaker_0): its peak, over 400 internet companies went public, and the NASDAQ, which is heavily focused on technology stocks, jumped by over 600% in total. So is this just the same scenario with a different mask?
- 9:59
- Unidentified speaker (UNKNOWN): The
- 9:59
- Unidentified speaker (speaker_0): short answer is there are indeed similar. similarities. Because of that history, people today look at AI and feel uneasy. Like the internet in the past, AI is also a new technology. AI related stocks have pushed
- 10:19
- Unidentified speaker (speaker_0): tech markets sharply higher again in recent years. Excitement Excement is high. investment is massive, and expectations are huge. While many AI leading companies like Nvidia, Microsoft, and Google are making huge profits and are
- 10:39
- Unidentified speaker (speaker_0): strong, cash rich businesses, very different from dot-com companies, where many didn't even make a single penny. But other companies like OpenAI. are still burning tons of money with unclear profit potential. This makes people wonder, how long will
- 10:59
- Unidentified speaker (speaker_0): this trend survive?
- 11:02
- Unidentified speaker (speaker_0): Section 5. Will the AI bubble burst? And what happens next?
- 11:08
- Unidentified speaker (UNKNOWN): So
- 11:08
- Unidentified speaker (speaker_0): the big question is, will the AI bubble actually burst? The honest answer is it might. There's a real possibility it will happen, but that doesn't mean AI will suddenly disappear. When bubbles burst, what usually breaks first is not
- 11:28
- Unidentified speaker (speaker_0): the technology itself, but the hype around it. If AI does burst, the first thing that happens is that money slows down. Investors become more careful. Companies stop throwing money at every AI project just because it
- 11:47
- Unidentified speaker (speaker_0): sounds cool. AI companies that don't make real profit will struggle, while the ones that survive will rise again. But here's the important part: AI itself doesn't go away. Just like after the dot com bubble, the internet didn't disappear.
- 12:07
- Unidentified speaker (UNKNOWN): I
- 12:07
- Unidentified speaker (speaker_0): mean, you're literally watching this video on the internet right now. What disappeared were unrealistic business models, like Pets.com, a company selling pet products that earned hundreds of thousands but spent millions. That didn't last. But companies that
- 12:27
- Unidentified speaker (speaker_0): proved their business could survive didn't die. Many people forgot that Amazon, eBay, and Cisco were all dot-com bubble survivors. Are they dead today? No, they're giants. And if you're investing in tech stocks, this is
- 12:47
- Unidentified speaker (speaker_0): why it's important to understand what you're actually investing.
- 12:51
- Unidentified speaker (speaker_0): in. Always do your own research.
- 12:55
- Unidentified speaker (UNKNOWN): If
- 12:55
- Unidentified speaker (speaker_0): you want to learn more about market crashes, you can check out my other video on what happens when the stock market crashes. The link is in the description. So, in conclusion, AI is a real and revolutionary technology that's already changing how we work and live.
- 13:15
- Unidentified speaker (speaker_0): But when a powerful new technology shows up, money often moves faster than reality. Some AI companies may fail, and some hype may disappear, but AI itself won't vanish. Just like after the dot-com bubble, the companies that
- 13:35
- Unidentified speaker (speaker_0): actually make sense are the ones that survive. So don't panic and don't blindly hype it either. If you want to invest in tech, always do your own research. And don't just follow the crowd or FOMO. Help us keep making videos like this by becoming a
- 13:55
- Unidentified speaker (speaker_0): member and unlock some exclusive perks too.
- 13:58
- Unidentified speaker (UNKNOWN): If
- 13:58
- Unidentified speaker (speaker_0): you want me to make other videos explaining these topics, please like and subscribe.
More from Explains 101
Why Nobody Can Afford Canada Anymore
Episode discussion: Speakers argue Canada's post-2008 monetary policy and immigration surge created a housing crisis leaving half the population financially strained.
Speakers’ claims: The speakers contend that Canada's economic stability post-2008 led to excessive monetary easing, fueling a real estate boom that outpaced wage growth and manufacturing. They claim that aggressive immigration policies between 2023 and 2024 exacerbated a housing supply shortage, as new arrivals significantly outnumbered new home construction. The episode highlights a disconnect between Canada's high GDP and the financial reality for households, with many living paycheck to paycheck due to high costs of living and debt. Speakers assert that the dominance of real estate has stifled innovation and other sectors like automotive manufacturing. They suggest that high housing costs distort GDP metrics and contribute to a broader economic paradox where national wealth does not translate to household financial health.
How the Rich Use Debt to Get Richer?
Episode discussion: Speakers explain how wealthy individuals use debt for tax advantages and leverage instead of viewing it as negative
Speakers’ claims: Speaker_0 argues that wealthy individuals utilize debt as a strategic tool rather than viewing it as inherently negative. The speaker distinguishes between 'good debt' for income-generating assets and 'bad debt' for consumption. Key strategies discussed include using mortgages to generate rental profit, leveraging low-interest loans for higher-yield investments, and the 'buy, borrow, die' method to avoid capital gains taxes. The episode also covers leveraged buyouts and cites Elon Musk as an example of maintaining control through debt rather than selling equity.