September 10, 2025 was one of the best days of the year for the market.
At the center of the party was Oracle, the legacy technology company that knocked on the gates of the AI industry with predictions of a blockbuster cloud infrastructure business.
Investors were so enthusiastic about Oracle's AI guidance that the stock soared 43% that day. At one point, the company was worth more than JPMorgan. Larry Ellison has overtaken Elon Musk as the world's richest person — at least for a few hours. The S&P 500 ended the day at an all-time high. The atmosphere was perfect.
Looking back now, this whole ordeal feels like this year's overreaction.
Oracle's stock price has fallen 46% since those blissful highs, and its annual positive return is now just 7%, roughly half the return of the S&P 500.
In the process, the stock became emblematic of a new market trend that has been percolating for weeks. Companies in the AI industry are priced perfectly, and those deemed to fall short of high expectations will be penalized.
Click here to sign up for Business Insider's market newsletter First Trade
Just ask Meta, Microsoft, or any other AI company that underperformed this past earnings season about their forecast lags.
Oracle was particularly hard hit by two disappointing developments in quick succession. Last Thursday, the company posted an 11% loss after spending far more than analysts expected on AI data centers last quarter. And on Wednesday, the company fell another 5% after a major investor pulled out of a $10 billion data center deal.
The market's reaction perfectly sums up the message investors have been sending to technology companies in recent weeks: Stop spending so much and start paying. The much-anticipated data center latency potential does not fit that requirement.
CoreWeave is another data center-focused company that has felt the same ire from investors. The former market darling soared more than 400% in the weeks following its IPO in March, but has fallen more than 60% since then. Who is the culprit? Data center delays.
All of the above have combined to undermine the invincibility of AI trading. Gone are the days when anyone could benefit indiscriminately by doing something related to AI. Investors are becoming more discerning and holding companies to a higher standard of progress.
The AI bubble may not have completely burst yet, but there is clear deflation in areas of the market that are struggling to meet expectations.
