- 01What Is an AI Bubble?
- 02Why Do People Think AI Is a Bubble?
- 03Why Others Say It's Not a Bubble (Yet)
- 04How the AI Boom Already Affects You
- 05What Would Happen If the AI Bubble Burst?
- 06Signs to Watch Over the Next Year
- 07What the Dot-Com Comparison Teaches Us
- 08Frequently Asked Questions
- 09The Bottom Line
“AI bubble” has gone from a niche worry to a headline. Economists are warning about stretched valuations, data-centre projects are running into trouble, and at the same time AI is making phones and laptops more expensive. If you don’t trade stocks, it’s fair to wonder whether any of this affects you. It does — just not in the way most headlines suggest. Here’s a plain-English explanation.
Some economists believe AI investment is in a bubble: Capital Economics said in September 2026 there are “plenty of signs that we are now in the late stages of a bubble in AI,” with its best guess being that it starts to burst in 2027. Others, like Dartmouth’s Kenneth French, say it’s too early to judge. Nobody can reliably time a bubble. For regular people, the practical effects are higher device prices from the AI-driven memory shortage, possible price rises or cutbacks in AI tools, and swings in stock-linked savings — not AI disappearing.
What Is an AI Bubble?
A bubble is when prices or investment race far ahead of the real value something is producing, driven by excitement that future profits will be huge. When expectations reset, prices fall — sometimes sharply. The classic comparison is the dot-com bubble around 2000: the internet was real and world-changing, but many companies were overvalued, and a crash wiped out many of them before the survivors went on to dominate.
An AI bubble would work the same way. AI can be genuinely transformative and still be overhyped in the short term. The two ideas aren’t contradictory.
Why Do People Think AI Is a Bubble?
Massive Spending
Goldman Sachs projects global capital spending on AI projects will hit about $1 trillion in 2026, including $581 billion in the US. That money is going into data centres, chips and power — and investors expect it to pay off with large future profits.
Stretched Expectations
Capital Economics’ James Reilly told CBS News that, comparing expected AI earnings growth with overall economic growth, “by many measures they look really stretched — as in this is the dot-com bubble all over again.” The firm also forecasts a stock market correction (a fall of at least 20%) in the S&P 500 next year.
Cracks in Big Projects
Some large infrastructure plans have hit trouble. In late September, TechCrunch reported that Oracle sent a force majeure notice on its New Mexico Stargate data centre and that Crusoe abandoned a $1.25 billion plan to use Boom turbines to power AI data centres. Bloomberg reported in mid-September that the risk of an AI slowdown was causing angst among tech investors.
Why Others Say It’s Not a Bubble (Yet)
Economists are genuinely split. EY-Parthenon’s Greg Daco cautioned that “every type of technological revolution tends to have a great dose of investment in the first phase,” so heavy spending alone doesn’t prove a bubble. Dartmouth’s Kenneth French told CBS that AI “is already having a huge impact on earnings” and that in five years people might look back and conclude investors were too pessimistic. His summary: “We don’t have enough information to judge if these prices are right or wrong.”
There’s also a difference between worries about AI safety and worries about AI investment returns. Daco noted that fears about missing guardrails are “slightly different than a bubble fear.”
| Argument | “It’s a bubble” | “It’s not (yet)” |
|---|---|---|
| Spending | ~$1 trillion a year is far ahead of AI revenue | Early phases of big technologies always look like this |
| Valuations | Expected earnings growth looks dot-com-like | AI is already boosting company earnings |
| Projects | Some data-centre plans are stalling | Demand for AI compute is still outstripping supply |
| Timing | Capital Economics’ best guess: bursting starts in 2027 | No one can reliably time a bubble |
How the AI Boom Already Affects You
Whether or not there’s a crash, the boom is already reaching ordinary wallets:
- More expensive phones and laptops. Memory makers have shifted production to AI data centres, pushing up the price of RAM and storage. IDC reports PC makers warning of 15–20% price increases. See whether to buy now or wait.
- AI everywhere in your apps. Companies are racing to add AI features, sometimes raising subscription prices to pay for them.
- Jobs. Some companies are cutting staff to fund AI investment, while others are hiring for AI roles.
- Savings and pensions. Many retirement and index funds hold large tech companies, so AI-driven market swings can show up in your balance.
What Would Happen If the AI Bubble Burst?
If history is a guide, a burst wouldn’t make AI go away. It would mean:
- Stock prices of AI-heavy companies falling, possibly sharply
- Some AI startups running out of money and shutting down
- Big companies slowing data-centre construction
- Free AI tiers getting stingier and some subscriptions getting pricier as companies focus on profit
- Eventually, perhaps, easing pressure on memory prices as data-centre demand cools
The practical takeaway: if you rely on AI tools for work, it’s sensible to keep your prompts, chats and files backed up and exportable, and to avoid depending on a single small provider. We’ll cover how to do that in detail in an upcoming guide.
We’re explaining the debate, not recommending any investment decision. Economists disagree and bubbles are notoriously hard to time. If you’re worried about your savings or pension, speak to a qualified, regulated financial adviser before making changes.
Signs to Watch Over the Next Year
- Whether AI companies’ revenues start catching up with their spending
- More data-centre projects being paused or cancelled
- AI tools raising prices or cutting free plans
- Memory and device prices stabilising — a sign AI demand is cooling
- Big tech earnings calls mentioning slower AI spending
What the Dot-Com Comparison Teaches Us
It’s worth remembering how the last big tech bubble played out. After the dot-com crash, most of the hyped companies disappeared, but the internet itself kept growing, and a handful of survivors became some of the biggest companies in the world. Infrastructure built during the boom — fibre networks, data centres — ended up powering the next decade cheaply. If AI follows a similar path, today’s overbuilt data centres and cheaper chips after a correction could make AI tools more affordable in the long run, even if some companies and investors lose badly on the way.
The lesson for regular people isn’t to pick a side in the bubble debate. It’s to avoid betting everything on any single company or tool, and to make sure your own work and savings don’t depend on one outcome.
- Some economists say AI is in a late-stage bubble; others say it’s too early to judge
- AI spending is around $1 trillion in 2026 (Goldman Sachs)
- You’re already paying for the boom through higher device prices
- A burst would hit AI stocks and startups — not make AI disappear
- This is a debate to follow, not a signal to act on without professional advice
Frequently Asked Questions
Is the AI bubble going to burst?
Nobody knows for sure. Capital Economics’ best guess is that it begins to burst in 2027, while other economists say there isn’t enough information to call it a bubble at all.
How would an AI bubble burst affect my job?
It depends on your industry. A slowdown could mean fewer AI-related hiring sprees and some layoffs at companies that over-invested, while businesses that use AI practically would likely keep doing so. Building skills in using AI tools well tends to stay valuable either way.
Will AI tools like ChatGPT disappear if the bubble bursts?
Unlikely. Large, well-funded services would probably survive, though free plans could shrink and prices could rise. Smaller AI startups are more at risk of shutting down.
Why are laptops more expensive because of AI?
AI data centres need huge amounts of high-bandwidth memory. Memory makers shifted production toward it, leaving less supply — and higher prices — for the RAM and storage in consumer devices.
Should I stop using AI tools because of the bubble talk?
No. The bubble debate is about investment and company valuations, not whether the tools work. Keep using what helps you, but avoid locking all your work into one small provider, and keep copies of anything important.
Is the AI bubble like the dot-com bubble?
Some economists make that comparison. As with the internet, a bubble doesn’t mean the technology is fake — only that expectations and prices may have run ahead of reality.
The Bottom Line
The AI bubble debate is real, and credible economists sit on both sides. For most people the lesson isn’t about predicting a crash — it’s about noticing how the AI boom already affects your costs, avoiding over-reliance on any single AI service, and getting proper advice before making money decisions. If you want to understand the memory side of the story, read how much RAM you actually need and why storage prices are rising.
Last updated: October 18, 2026. Sources: CBS News, Goldman Sachs, Bloomberg, TechCrunch, IDC.