I haven't bought too many AI stocks in my portfolio yet, except for a few data center REITs and one major chipmaker. But recently, I pulled the trigger and added shares of Oracle (ORCL +1.77%) to my portfolio after shares declined by more than 60% from their 52-week high.
The short version is that the market has legitimate concerns about Oracle's massive backlog and the debt it is taking on to fulfill its contracted orders. But if the company's strategy works out, the stock could be an incredible bargain at the current share price. Here's a rundown of why Oracle's stock has been beaten down, why I bought, and why I may add even more to my position.
Image source: Getty Images.
Oracle: The good and the bad
The biggest bull case for Oracle is also the same reason many investors are skeptical. The company has a $638 billion remaining performance obligation (RPO), which represents the contracted future revenue its customers have committed to. This is 363% higher than it was a year ago, and as you can probably guess, the surge in AI infrastructure spending is the main reason.
For its 2027 fiscal year, which started June 1, Oracle is guiding for $90 billion in revenue. For reference, in its 2026 fiscal year, the company generated about $67 billion. And keep in mind that the $638 billion figure represents only contracts that have already been signed -- as customer needs grow, this figure could increase.
However, many investors are justifiably skeptical about the massive backlog. And it's fair to say that if we knew Oracle would actually get all $638 billion of that contracted revenue, the stock wouldn't be as beaten down as it is.
For one thing, about half of the backlog comes from a single customer -- AI platform giant OpenAI. That company recently delayed its IPO, which added to major concerns about its ability to meet its obligations. In addition, Oracle is not only spending all of its income to set itself up to fulfill its backlog but is also taking on debt. The company added $43 billion in debt to its balance sheet in the 2026 fiscal year, and expects to raise an additional $40 billion between debt and equity in the current fiscal year.
In short, Oracle's spending plan is frightening investors, and it's not hard to see why. The company is spending billions of real dollars (much of which it doesn't have) in pursuit of promised revenue.
A substantial discount for a wonderful business
After the recent decline, Oracle trades for less than 16 times forward earnings. The stock has a 1.6% dividend yield and has produced a net margin of nearly 27% over the past four quarters. This is a highly profitable business that is sacrificing short-term cash flow to take full advantage of the tailwinds from the AI infrastructure build-out.
Of course, there are some legitimate risks. It's entirely possible that OpenAI won't be able to raise the capital needed to cover all of the spending it has committed to (Oracle isn't the only one with a large, multi-year agreement with the AI giant). We could see a general downturn in AI capex instead of the multi-year spending surge investors are expecting.
I don't view those as particularly likely scenarios. OpenAI hasn't had much of an issue raising money at lofty valuations whenever it has tried to. And all recent signs point toward more infrastructure capex than originally expected in 2027 and beyond, not less.
The key question is whether a significant portion of the backlog will convert into actual revenue before Oracle's spending plan becomes unmanageable. If the answer is yes, Oracle at 16 times forward earnings could be a steal.
The bottom line is that Oracle has been a wonderful business for decades, and for most of its 35-year publicly traded history, it has been a mistake to bet against it. With the stock trading at its lowest price since 2023, before the AI investment boom even kicked into high gear, I decided to open a position in my portfolio. And if it stays this cheap, I plan on adding more shares very soon.
