Tom Lee thinks AI won't reach its true bubble peak until 2033

Fundstrat's Tom Lee thinks investors are way too early with their warnings of an AI bubble. According to him, the current market does not look like 1999, just before the bursting of the dotcom bubble, but rather like 1997.

That would mean that the biggest increases are still years ahead of us.

Lee compares the current investment wave with the roll-out of the internet, mobile networks and smartphones. Converted to today's dollars, that infrastructure cost about $35 trillion, according to him.

For AI, he estimates the total investment at about $30 trillion. His point is that such technological revolutions can go on much longer than investors expect.

For example, Cisco did not reach its final peak until March 2000, while the share had already experienced several strong corrections before that. Lee therefore thinks that the current cycle may not reach its real peak until around 2033.

Another argument from Lee is psychological. According to him, the biggest bubbles occur when hardly anyone thinks there is a bubble. Now investors, economists and strategists are constantly warning against overvaluation.

That makes him skeptical about the idea that the summit is already close.

That's where Ray Dalio comes in. The founder of Bridgewater calls the current situation a classic bubble. His biggest concern is guilt.

More and more AI investments are no longer paid for out of free cash flows, but with bonds and loans. With high interest rates, this becomes increasingly expensive. Dalio therefore fears that investors will eventually start taking profits as soon as doubts arise about the payback period of all those investments.

This means that the difference between the two camps comes down to one question. Can the huge investments eventually be paid for out of rising cash flows?

If that succeeds, Lee can be right and the real peak is still years ahead. If companies become more and more dependent on debt while interest rates remain high, Dalio is more likely to be right.

So the investments are not the problem. The question is above all who pays them and how much return they ultimately yield.