https://www.youtube.com/watch?v=MsPie9ZUuM8

无作者

A bubble is what happens when the price of something rises far above what it can, earn because people buy it expecting the price to keep rising. In the context of, AI, this is spending that runs far ahead of revenue. Amazon, Alphabet, Microsoft,Meta, and SpaceX are on track to spend between 828 billion this year on data centers, chips, and other equipment.So, how big is the gap between that spending and the revenue it will bring, in? The research firm Gartner has estimated this and says data center, owners need 3.6, trillion, which leaves more than $4 trillion of, assets at risk of being written down. Today’s prices only make sense if that, gap closes. If it doesn’t close, that’s when the, bubble bursts, which is a fall in prices, and the people who bought near, the top lose money to the people who sold. Many people now expect that the.

gap will not close. But first, what would a burst actually do to someone who, has never bought an AI stock? Let’s start with the first few days. In, the World Economic Forum scenario, a burst begins as a financial market and, media event, and the early damage to the wider economy is limited. The immediate, job is to keep markets orderly, and central banks are good at supplying cash, in a crisis. Smaller banks that lent to AI companies would come under scrutiny,so the forum says regulators would need to offer support quickly.The damage that lasts would then come through business investment. A Harvard, economist estimates that AI infrastructure spending made up 92% of, US economic growth in the first half of 2025 and says AI and IT investment has, risen to about 5% of GDP. In its scenario, a stock market shock, cuts consumer spending, companies lose confidence and cut investment, and the, US tips into recession.

GDP falls 0.6% in 2027, and investment, falls more than 6%. And for jobs, economists expect losses.

at newer AI companies that only survived because investors kept funding them.Once that money stops, those companies can’t stay open. Established tech, companies have businesses beyond AI, which would limit their cuts, and few, people outside the tech industry would know someone at risk.Then there are households. Stocks made up 30% of US household wealth in 2024.Commerce Bank assumes that people cut spending by 3 cents for every dollar of, stock wealth they lose. If household stock wealth fell 44%,households would lose more than 28 trillion dollars and spend about 850, billion dollars less, which is 2.6% of GDP.One more risk is debt. Oliver Wyman points out that AI spending is, increasingly paid for with borrowed money, and that hyperscaler bond sales, topped 100 billion dollars in 6 months, more than five times the amount in the, previous 2 years. If so, losses could spread beyond stocks into credit, markets. But all of this now raises the next.

question. Has any of this even started? Partly. On October 30th, the Nasdaq 100, fell 3.5%, which took its October loss to 6.7%,its worst month since March 2025. Nasdaq 100 companies lost more than 1, trillion dollars in market value that day.The sell-off followed forecasts from Microsoft and Google that showed less AI, spending than analysts expected, and a weak outlook from Intel.It spread beyond the US as well. South Korea’s KOSPI fell 3.48%, and Japan’s Nikkei fell 2.1%. Credit markets reacted, too. Credit, default swaps, which are contracts that pay out if a borrower defaults, hit a, record price for Broadcom. That is a warning, but it is small next, to 2000, when the Nasdaq fell almost 80% over 2 and 1/2 years. A drop of 6.7% in, a month is a bad month, but it is not a collapse.It also isn’t the first sign of a burst. Fortune lists a long run of them since, ChatGPT launched in late 2022. Bears raised alarms as Nvidia passed.

each trillion-dollar milestone. The Chinese model DeepSeek caused a panic in, early 2025. Each time, earnings kept rising and, stocks went on to set new records. So, what we have is a wobble, not a, burst. But, if it is only a wobble, why are so many people confident that a real, burst is coming? The first reason is how few companies, the market depends on. A JP Morgan basket of 41 AI-linked companies is now, 49% of the S&P 500’s value. The top 15 stocks make up 46.7%, of the index against 27.2% at the 2000 peak. Chipmakers have, contributed about 70% of the S&P 500’s gains this year, and Nvidia alone is, about 8% of the index. The second reason is cash. Four of the, five biggest spenders had negative free cash flow in the third quarter, meaning, they spent more on buildings and equipment than their operations earned.Microsoft was the exception. Amazon’s was more than 25 billion of bonds, its first.

bond sale in more than 9 years, and Meta plans to sell new shares to raise up to, 30 in, Open AI and guarantee up to 50 in, September, around $20 below figures previously reported to investors.If investors doubted that Open AI can pay its commitments, Nvidia would be hit, twice through lower sales and through its own investments.Because Nvidia is the largest company in the index, that would drag the whole, market down. The chips also wear out faster than the, old fiber did. The newest AI chips have a shelf life of only a few years.

One group looks especially fragile, the neo clouds, which are companies that, rent out data centers to AI developers. Coreweave, the biggest, has $35.1, billion in debt, and its stock has lost a quarter of its value in a month.Gartner expects 80% of neo cloud firms to stop operating independently by 2030.Each of these risks is a way for trouble in one place to spread to the rest.So, does it matter if those risks play out? Yes, because so much rests on it.Nearly half of the US stock market’s value is in AI-linked companies. AI and, IT investment is about 5% of GDP. Commerce Bank says AI investment has, been a floor under US growth, which was only about 2.3% over eight quarters even, with it. The exposure also reaches people who, never chose it. Stocks are 30% of US household wealth. In Australia, nearly, 90% of households are exposed to AI stocks through the US shares in their, retirement funds. So, what happens when the AI bubble.

bursts? The technology would probably keep working as the internet did after, 2000. The loss would fall on investors, on the companies that borrowed to build,and then through investment, jobs, and household savings, on people who own no, AI stock. How hard it hits depends on one gap, which Gartner puts at more than, $4 trillion of assets at risk. Nobody can yet say whether revenue will grow, fast enough to close it, but the October sell-off shows that investors are, starting to ask. And as always, if you like these videos, on what’s happening in the AI industry, make sure you subscribe so you can stay, up-to-date.