I’m not the only one who has been talking about the impending pop of the AI bubble. GenAI (or LLMs) are way too expensive to train and run, but so far, there is no killer application that can cover the cost, let alone generate a profit for the investors.
We are starting to see the signs of the bubble bursting, but rather than a loud pop, I think it will go out quietly with a hiss.
I was reading this article on the anatomy of an AI reckoning.
By definition, a bubble results in investments that fail to achieve promised results. This means that economic growth during the bubble phase depends on continually building infrastructure, not using infrastructure.
For the last 3 years, the “economic growth” has been from the sale of the NVIDIA chips and memory used for AI infrastructure. These chips are to be installed in data centers to deliver even “more AI”, although no one knows what shape or form that AI will take other than LLMs. The other growth was from building data centers to house those chips.
OpenAI has been making deals left, right and center to build data centers, despite having no experience or expertise in building them. OpenAI reportedly signed a $300B data center investment with Oracle in September of 2025, then Oracle took on debts to finance building the data centers, and now that deal is dead as of March 2026. Oracle stock has been crashing since then.
Several other deals for data centers have also not materialized. Ed Zitron covers the data center “deals” and how it is all a con quite extensively in his free newsletter.
From announcing a ton of new models, products, deals, and financing rounds, OpenAI has now shifted to shutting down products such as Sora and cancelling deals, like the one with Oracle, Disney, and NVIDIA.
So I would say we are the beginning of the end of the AI bubble. We’ve reached peak financing/deals, and are now beginning to deflate.
But will it be a loud pop?
I think it will be a quiet hiss, with companies quietly scaling back their investments in AI. Note that for the companies that can afford it, they will still continue using AI in their products (I am looking at you, big tech), but their generosity might go down, and they might start charging more or limit tokens or both. The stock market is currently taking a beating from Trump and Netanyahu’s illegal war in Iran, so the regulators might want to stop panic selling so that the markets remain stable. They may continue lending to banks to maintain liquidity, so that the AI bubble doesn’t pop too rapidly. (Speaking of selling stock, it does not bode well for the Coreweave price that the CEO sold millions worth of stock, even if it was prearranged. CoreWeave is a company that builds data centers to lease to their customers, usually big tech.)
To quote the WEF article:
Bursting a bubble is primarily a financial market and media event, and the initial economic fallout is limited. The immediate macroeconomic concern is to avoid disorderly financial markets. However, the world’s central banks are adept at providing liquidity in a crisis. The financial market focus of AI has been primarily on the US, and so the US Federal Reserve would be looked to as the main liquidity provider. This would not necessarily require interest rate cuts, although political pressure to do so would likely be intense.
The same 7-10 companies have been involved in the circular financing of the AI bubble, and the Magnificent 7 stocks (NVIDIA, Microsoft, Apple, Alphabet, Meta, Amazon, and Tesla) accounting for any growth in the US stock market recently.
Hopefully, it will be a hiss with reduced effect on the overall US economy (and by extension the world’s), but hey, I am no economist. I may be wrong, and it might give a satisfying pop but leave lasting damage.
However it ends, we don’t know, but the end has definitely begun. So I thought it would be a fun project to monitor the Mag7 stocks as they deflate (there will be some ups, but the overall trend will be downward). I have also included Oracle and Adobe. (Bear with me as I am still trying to find suitable, interactive stock market widgets. I will keep adjusting this page, so please come back again for the updated dashboard. If you have any recommendations, please let me know in the comments!)