Photo by Mohamed Nohassi on Unsplash

It’s said that “they don’t ring a bell at the top”.

Well, there sure seem to be a LOT of bells coming from the AI revolution.

First and foremost, OpenAI founder Sam Altman, openly admitted AI is in a bubble even going so far as to compare the current market environment to that of the 1999 tech bubble.

Altman told The Verge…

“Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes.”

That is one HECK of a statement from a major AI insider. Remember, OpenAI was the company that introduced Large Language Models (LLMs) to the general public in November 2022 when it launched ChatGPT. So to see Altman admitting AI is in a bubble, that investors are overexcited, and that there are some similarities between today’s market and that of the Tech Bubble of the late-90s, is massive deal.

Altman is not mistaken here. The latest ChatGPT roll out (ChatGPT 5) was a disaster, with the software producing hallucination 33% of the time. This included basic queries that involved public information. Indeed, situation was so bad that OpenAI actually canceled the rollout and shifted back to ChatGPT 4.

This is not the only major disappointment to hit the AI world in the last month. AI models from Google, and Chinese startup DeepSeek are also experiencing MORE errors, not fewer as they become more powerful. Indeed, it appears that the more advanced the models become, the more problems they produce: Meta’s (META) latest LLM rollout (Llama 4) apparently had accuracy rates as low as 28%!

Some other major issues of note.

Reports have surfaced suggesting that LLMs obtain as much as 40% of their information from informal sources such Reddit. This raises questions as to just how accurate these models can be. After all, no matter how robust your model is, if it’s getting bad inputs, its results will be poor.

Other items to be aware of…

No other company has been as aggressive in hiring AI talent as META. CEO Mark Zuckerberg was personally involved in the process, offering deals as large as EIGHT figures (tens of millions) to poach AI talent from Apple, OpenAI and Alphabet.

In that light, it’s deeply concerning that META just announced a reshuffling of its AI division only two months after going on that hiring spree. Indeed, the company has burned through $32 billion in capital expenditures on its AI buildout during the first half of 2025. This represents 72% of the company’s cash hoard, bringing cash and cash equivalents down to just $12 billion.

And then, finally, MIT just published a report The GenAI Divide: State of AI in Business 2025 that was EXTREMELY concerning for anyone betting on AI changing everything for the better.

Some key findings…

• 95% of generative AI rollouts produce little to no revenue acceleration.
• The learning gap is so large between the technology and corporate employees that integrations are failing at a rapid clip.
• Internal buildouts at large companies only work one third of the time.
• Even external partnerships with AI-centric firms only work two thirds of the time,

I bring all of this up, because the ENTIRE bull market in stocks has been predicated on the AI revolution producing rapid results. If you need signs that stocks are in massive bubble based on unrealistic growth expectations, consider that the S&P 500 is trading at a Price to Sales multiple of 3.4. This is even greater than the P/S multiple of the 2021 pandemic bubble when the Fed was printing TRILLIONS of dollars per year and interest rates were at zero!

(Click on image to enlarge)

More By This Author: