he numbers are out for Oracle’s Q2 fiscal 2026, and the market reaction is ugly. Shares are down more than 10% in after-hours trading and still struggling to find a base as $200 gave way.
Here is the data:
-
EPS: $2.26 vs $1.64 expected
-
Revenue: $16.06bn vs $16.19bn expected
-
Cloud Infrastructure Revenue: +68% y/y
-
Software License Revenue: -21% y/y
This is a classic case of “the bottom line doesn’t matter if the growth story has holes.” The big turnabout in sentiment is the huge AI spending commitments and the wisdom of them.
That thinking briefly led founder Larry Ellison to be the world’s richest man in September but that’s clearly marked the top and shares are down 43% since, including the after-hours drop.
On the surface, a 62-cent beat on EPS looks fantastic. In reality, it was due to a one-time $2.7 billion gain on the sale of chip designer Ampere to SoftBank for $6.5 in March. Given the performance of chip makers, that sale may have been unwise and it cuts off a potential major win for Oracle.
Investors have grown concerned about the amount of debt the company is taking on to fund a $300 billion commitment for OpenAI and other generative AI investments. Oracle is forecasting $50 billion in full-year capital expenditures, from $35 billion in September. That’s sure to rise further next year.
The legacy business also isn’t exactly healthy with software license revenue down 21%. That is a massive drag. It tells you that while the cloud business is booming (up 68%), it’s cannibalizing the old cash-cow business quickly.
Total software revenue was actually down 3%. That is not what you want to see when you are trying to pitch yourself as a high-growth AI play.
The Price Action
ORCL was already heavy coming into this print, down about 25% from the highs over the last month. The bulls were praying for a blowout to reverse the trend
