One of the most surprising storylines in the market is the rapid rise of Oracle as an AI infrastructure giant. The company that was once recognized for its regular software business and slight capital requirements has now changed its whole business around the AI trend.
Now, it creates massive and expensive GPU-driven data centers, along with making multi-million dollar agreements such as the $300 billion deal with OpenAI.
Its buildup became more than half a trillion dollars, and for a brief period Oracle seemed to be one of the primary victors of the AI race. However, as soon as the initial buzz wore off, investors had a huge pullback, which dropped the share price below the prior OpenAI deal level.
This shocking turn of events creates a very uncomfortable thought, where one wonders whether Oracle’s AI infrastructure strategy is really so profitable as to be considered at the start of the year or not.
Weak Economics Behind AI Infrastructure
Doubt about profitability is the main reason for the selloff. AI infrastructure might look like a golden business, where they develop data centers, lease them, and repeat, but the economic foundations are proving to be less sparkly.
Oracle itself estimated the adjusted gross margins to be 30% to 40%, however the reports are indicating that actual margins have been much lower. As per the Information, the gross margin from the AI infrastructure for Oracle was as low as 14% in the quarter that ended in August which is a stark difference from the targeted range.
The disparity mentioned is quite alarming, since Oracle is very much adopting a capital-intensive approach with a substantial increase in its capital expenditure. The entire company has boosted its capex forecast for the year by $15 billion, based on the idea that by doing so, the company would be able to access the revenue that was covered by the recently signed contracts.
However, the numbers do not seem very attractive, as the company spends $15 billion, it will be able to earn only $4 billion in the next fiscal year. This would then be converted to approximately $1.4 billion in gross profit at average margins.
The cost of operating will take an even bigger bite out of the profit, which is likely to be less than 10%. This is the kind of situation with capital and low returns that will not be able to support the claim that AI infrastructure will be a long-term profit source.
Downside Risk
Oracle’s debt and dependence have raised the stakes on its downside risk. The company is now heavily relying on debt for its aggressive expansion, which makes its risk profile more complicated. Currently, Oracle has a total debt of $108 billion, most of it from when AI was not a big deal, but all of it is still affecting market perception negatively.
Unlike its competitors, Microsoft, Google, and Amazon, all of which are loaded with cash, Oracle does not have the same financial support to onboard mistakes or survive the downs in the industry.
Therefore, the company has become extremely susceptible to scenarios such as the AI sector overbuilding capacity or the situation where demand does not grow as fast as it was thought. The backlog is very huge, but it relies a lot on OpenAI’s financial stability in the long run.
If OpenAI has a hard time getting funding or slows down its demand for infrastructure, Oracle’s revenue ramp-up could very become a thing of the past. The narrative has changed from being about opportunity to being about vulnerability.
Should You Buy the Dip?
Investors might consider Oracle’s decline as an opportunity to buy, given that shares are down about 40% from their peak in the last 12 months. However, it is advisable to act cautiously. The market is not questioning the importance of Oracle in AI, instead it is questioning the present day profits of the company.
Oracle is perceived as a great contributor in AI infrastructure, however the returns are still vague, the requirements of the capital are increasing, and the burden of debt is raising the risk of any future errors. There is no doubt that Oracle has played a significant part in the AI world, but the market is showing that such importance is not enough.
Investors need to be assured that the heavy investments in capital will give rise to durable, high-margin returns and the figures so far are not convincing. Until Oracle proves that it can convert its AI aspirations into sustainable profitability instead of piling up debt and having thin margins, the skepticism surrounding its AI-led future will continue to overshadow the promise.