Posted on 07/18/2025 11:02:45 AM PDT by lasereye
A top Wall Street economist is sounding the alarm on sky-high valuations in AI stocks — and drawing comparisons to the tech bubble of the late 1990s.
"Yes, AI will do incredible things for all of us," Torsten Sløk, chief economist at Apollo Global Management, said on Yahoo Finance's Opening Bid. "But does that mean I should be buying tech companies at any valuation?"
According to Sløk, the answer is increasingly no. In a research note to clients this week, he pointed to internal data showing the price-to-earnings ratios (P/E) of the 10 largest companies in the S&P 500 — many of them AI stock picks like Meta and Nvidia — have eclipsed P/E levels seen at the height of the dot-com bubble in 1999.
That signals a dangerous concentration of investor exposure in just a handful of tech giants, Sløk argued.
"Almost 40% of the S&P 500 is made up by the 10 largest companies," he said. "So if I take $100 as an investor and buy the S&P 500, I think I have exposure to 500 different stocks, but I'm really just betting on the Nvidia and the AI story continuing."
In his note, Sløk noted that the current valuations in megacap tech stocks, and the index as a whole, may not be sustainable. His concerns echo a growing unease on Wall Street over how much of the recent stock market rally is driven by AI euphoria and momentum trades.
BTIG analysts flagged similar warning signs in a note this week, describing market sentiment as "frothy" and raising the possibility of a near-term pullback in high-flying AI names.
Their focus was on the BUZZ NextGen AI Sentiment Index, a benchmark of AI-related stocks popular with retail investors. The index is up 45% over the past 16 weeks.
(Excerpt) Read more at finance.yahoo.com ...
But some big companies also got way overpriced back then, which is the focus of this analyst. For example Microsoft was at $62 in December 1999. After the bubble popped it didn't get back to that until April 2017. It's around $511 today, which works out to an annual return of around 8.5% if you bought it at the bubble peak.
It may be a good idea to take some profits on any AI related stocks.
Theoretically, AI is the wisdom of crowds.
In practice, it is GIGO.
The slopes of the NVDA and PLTR charts are reminiscent of late 1999.
“In practice, it is GIGO.”
Like some people!
Is there a market bubble in AI company valuations?
Yup.
I’m using it in my CPA practice recently and I find it extremely helpful tool and its uses going to help me increase my fees and decrease my time a win-win
Sell Microsoft?
Definitely true of public domain use of AI. I can see practical use of AI with certain amounts of datasets constrained to only data that's been verified. Or perhaps use AI for early layers of forensic research, even looking at large public data sets of unverified data, but only with the understanding that eventually the AI research results themselves have to be investigated to be certain.
But AI as we often see it used is definitely just GIGO on steroids.
You should use real intelligence to construct a coherent sentence.
I don’t think the issue is the efficacy of AI, the issue is too many AI engines, and start ups centered around them, which has led to too much investment concentrated in the category.
Use one company’s AI to vet the work product of another company’s AI—while you take a nap!
Limited, narrowly-focused applications, such as automated navigation, either in the air or on the ground, can work well because those are primarily mechanistic tasks. But computers are not good at those tasks because they’ve somehow obtained the ability to think. They’re good at them, even better than humans at times, because they can use a multitude of sensors to give them better situational awareness, and they can calculate at extremely high speeds. These are tasks that are perfectly suited for what is just a very advanced calculator.
But the pipe dream of “general AI” will never work, for among many reasons the fact that these software algorithms make constant errors (because they can’t actually think!), and that the larger they grow the more errors they make and the more difficult (and eventually impossible) it becomes for humans to detect errors and correct them. A sufficiently comprehensive “AI” would be impossible to quality check, because no human or even team of humans would be able to know how and why it made every one of its trillions of decisions per second.
Unless we’re just going to create an idiocracy that unquestioningly trusts anything some “AI” algorithm spits out (a very real risk given the current foolish feeding frenzy), I see no long-term place for “general AI.” The very concept is fatally flawed, no matter how advanced the calculation engines that it runs on become.
You mean NVIDIA at 55 and TESLA at a PE of 177 might be a little over priced? Tell me it isn’t so. /s
These are so overvalued and stay that way because they are held by so many who would now rather not but are in too deep.
Ferrari has a PE of about 51 and is more profitable than Tesla.
Tesla is simply a car maker and ranks globally in the bottom of the pack by number of units but second to Ferrari in profitability per unit. Watch what happens there without the subsidy.
If I did not already own S&P 500 ETF I would opt for the S&P 493.
Yeah, and when you sell it continues to go up! It's all a gamble, either way. We own some AI related stocks. One that my wife bought for $10 went up 2500 percent (she bought a lot of shares). If she had taken any profits earlier she would have kicked herself. We're sitting on most unless there is a substantial reversal.
“Tesla is simply a car maker”
You should do your homework before disseminating false information.
Remember, AI is nothing more than what some programmer designed it to be, including that programmer’s ignorance, biases, depth or lack of knowledge, experiences, and data sources.
Ask it a question like “who is the best presidential candidate” and see what happens.
So, AI stock is good for laundering money.
One more thing, restoring and recommissioning old nuclear plants to feed the server or data farms? No risk in that is there?
Anything is possible if you can get your hands on enough of OPM. It becomes all about he multiples and keeping the show attractive.
On the other side of the equation though, I asked google what dividend stocks to invest in for long term income. Three out of the four named are the same ones I spent days doing research and screening to find and ultimately bought. The citations given were complete, current and mostly well reasoned. Same goes for medical diagnosis. Not a complete answer but a decent pointer followed by confirmation.
We have been here before at least once in my lifetime. I bought Broadcom waaaaay back but I also bought a few others in that day that have since vanished. I also bought 20% zero coupon bonds back then with the longest maturities I could possibly get. Just wish I had bought more of them. However, like one of my landmen said, “25% of a good deal is probably enough.” He made Billy Bob look like the yokel that he really is.
Wrong, and full of FUD. Tesla is much more than a car maker. Heavily invested in AI, robotics, energy creation, etc. Just their energy division may surpass their car profits (they're becoming a power-utility company). Then there is their contributions to research with Musk's other companies, like SpaceX with Starlink and metals research, and AI with NeuraLink. I suggest you do some research into Tesla and their operations. Musk suggested that they won't be focused on selling cars in the future, it will be a side operation for producing RoboTaxis and commercial vehicles.
What? Solar, home power batteries, charge points? All made more attractive if not feasible by incentives. Let’s see how it works out now that those incentives are disappearing.
There is not enough blue sky in Tesla for me to ever pay to access a PE of 177.
I will eventually put in solar but the payout on batteries just does not float even at higher electricity prices, which I expect to arrive as the net zero falls apart and is replaced in a panic by expedient sources. Excess capacity and reverse metering don’t work either. You can’t pay a low enough cost to make excess capacity payout in a reasonable time. The RoR of that investment is a loser at the rejected cost of power.
Your thoughts are your own and you are welcome to them. Maybe one of us will be right.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.