Oracle's AI Push Meets Wall Street's Cold Shoulder

Oracle launches a US clinical AI agent and Java 27, as cloud infrastructure revenue jumps 121% and analysts trim targets on spending concerns.

Oracle is racing to prove that artificial intelligence can do more than burn through capital — it can change how hospitals run and how developers write code. Within a single week, the software giant rolled out an AI agent for clinical staff in the United States and shipped Java 27, a release built around post-quantum cryptography and AI-ready features. The twin launches signal a deliberate shift: away from the pure data-center land grab that has dominated the sector's narrative, and toward software that lives inside customers' daily workflows.

That pivot arrives at an awkward moment. For months, the industry debate has fixated on the staggering sums poured into server farms, yet the real test of AI's value plays out in the operating rooms, wards, and code repositories where automation either saves time or doesn't. Oracle's bet is that wiring intelligent automation into its vast installed base of databases and programming tools creates a durable edge. Healthcare, in particular, ranks among the most lucrative — and most demanding — markets for such technology. If hospitals can measurably cut their documentation burden, Oracle unlocks a reliable stream of recurring software revenue.

Blockbuster Numbers, Cautious Analysts

The fundamentals in the infrastructure business certainly look explosive. Cloud Infrastructure revenue jumped 121% year over year to $7.4 billion, effectively doubling from the prior-year period, while total quarterly revenue climbed 30% to $19.3 billion. Adjusted earnings per share of $1.92 sailed past estimates. Most striking of all, remaining performance obligations swelled by $209 billion year over year to reach $664 billion.

Management leaned into the momentum, guiding for fiscal 2027 adjusted earnings of $8.10 per share on at least $90 billion in revenue. Yet instead of cheering, several research houses trimmed their targets. BMO Capital Markets cut its price objective to $195 from $220 on September 11, keeping an "Outperform" rating. Scotiabank followed on September 9, lowering its target to $215 from $241 while maintaining its sector outperform call.

The hesitation reflects a sober calculation: building out cloud and AI-grade data-center capacity devours cash, and the market increasingly wants proof that those outlays translate into lasting margin expansion rather than a drag on future profitability.

Cutting Deep to Fund the Future

To protect margins, Oracle's leadership has turned to the knife. A fresh round of layoffs began roughly a week ago, and the stock has gained 2.7% since. Media reports suggest as many as 10,000 positions could be eliminated, with the company now pegging the total cost of its restructuring program at around $2.8 billion.

Painful as those cuts are for the people affected, they are hard to avoid for a company repositioning itself at the front line of cloud and AI. Carrying expensive excess capacity in legacy business lines is a luxury Oracle can no longer afford, and the moves show a willingness to redirect capital decisively toward growth segments.

A Stock Still Searching for Traction

Investors have yet to buy the story. Oracle shares closed Friday at EUR 128.70, down 23% since the start of the year. The stock also sits 10.0% below its 200-day moving average of EUR 142.97, a technical signal of persistent consolidation pressure.

The coming quarters will test whether Oracle can bridge the gap between heavy upfront spending on modern infrastructure and profitable software delivery in the day-to-day business. Clinical assistants and next-generation programming interfaces are compelling calling cards — but the market now wants to see the cash register ring before it rewards the vision.