Big Tech’s Bubble Could Burst
It’s September the 3rd. A notable date. It’s the date that Richard the Lionheart was crowned King of England. It’s the date England and France declared war on Germany. And it’s also my dad’s birthday. He was a fighter pilot. One of the many jets he flew was the De Havilland Vampire training aircraft. He used to train people to fly in Sale, UK.
Can you believe it?
Anyway, people are talking at the moment about Big Tech. We’ve just had the Nvidia results, and the day after, the whole Big Tech complex seemed to have slowed down a little bit. I don’t know whether it was an end-of-month sell-off in the US or a pre–long weekend pause.
But Big Tech has seemingly tipped over. Our market has a very small tech sector, but it’s off over 100 points today. There’s this feeling that the market has peaked for a minute. And the question we have—because in our Strategy Portfolio we are buried in Big Tech stocks and the US market—is: what could possibly tip the US market over?
The obvious answer is Big Tech stocks. They’ve got a lot of sentiment in the share prices, as opposed to value. Every share price is a chunk of sentiment and a chunk of value. Traditional investors have overemphasised value, but there’s a big element of sentiment we have to worry about—and sentiment is fickle.
So what could damage sentiment in the US market? Anything that damages sentiment in Big Tech stocks, because they are the rock in the sock that’s driven all the markets over the last year and a half.
What could go wrong with Big Tech? One thing is the whole market. It’s sentiment driven. If sentiment changes—for whatever reason, Trump, tariffs, a bond market sell-off—you’ll see the gold price going up, which says people are beginning to worry about something.
If there was a major correction, it would damage sentiment-driven stocks. That includes the big investment banks in the US and Big Tech in the US. Some correction would kill Big Tech.
The other obvious thing is if there was any sort of clear communication from the hyperscalers—the people building the big clouds, the language learning models, the people buying all of Nvidia’s chips. If there was any mention from them about backing off on capital expenditure on AI, that would tip the sector over as well.
Any sort of demand plateau for chips would do it. It takes quite a long time to build a data centre, build a language learning model. If Nvidia suddenly said demand for chips is being bottlenecked by the speed at which data centres are built, that could kill it too. Any sort of peak in demand would kill the story.
If Nvidia did something specific to drop its own share price, that could trigger it as well. There’s all sorts of flux around their Chinese chip sales. They didn’t include any chip sales in their guidance in their last quarterly results. If they were prevented from selling into China, or we saw Chinese chips competing, that would matter. OpenAI and Meta are developing their own chips—if they start to compete with Nvidia, that’s a problem.
It’s a $4.4 trillion stock, 8% of the S&P 500. Any company-specific hiccup could kill the story.
Competition is another risk. That DeepSeek moment earlier this year made it clear AI is a commodity. It could come cheaper. Maybe the rolled-gold language models in the US are too expensive. If it can be done cheaper, the whole thing becomes a commodity. Prices drop, margins drop, and willingness to spend disappears. Another DeepSeek moment—cheap AI—would kill the baby as well.
