Oracle's $18B→$73B cloud forecast sends shares up 27%; RPO soars 359% and analysts lift targets to $360

Big moves around Oracle (NYSE: ORCL) this week. The stock ripped higher after management laid out aggressive cloud-infrastructure revenue targets and analysts responded by upping price targets across the board.

Management is projecting cloud infrastructure revenue to jump to roughly $18 billion this fiscal year - about a 77% lift - then to $32 billion in fiscal 2027 and a staggering $73 billion in fiscal 2028. Even more eyeball-grabbing: remaining performance obligations (RPO) climbed 359% year-over-year, a big back-book metric that tells you how much contracted business is sitting in the pipeline. That backlog helped investors shrug off softer headline quarterly earnings; ORCL shares spiked as much as 27% in after-hours trade and are up about 45% so far in 2025.

Analysts moved fast. Jefferies bumped its target to $360 from $270 and stuck with a Buy, saying the RPO figures "stole the show" and bolstered the case for revenue acceleration.
BMO Capital Markets raised its target to $345 from $275 and kept an Outperform call, but it warned there's risk in turning RPO into realized revenue.
Mizuho raised its target to $350 from $300, highlighting demand from AI-heavy customers like OpenAI, Meta, xAI and Nvidia as a growth driver.
D.A. Davidson lifted its target to $300 from $220 but stayed Neutral, flagging worries about long-term margins - noting Oracle may be running virtual machines and GPU rentals at single-digit operating margins, or even at a loss in some instances.

The next calendar marker traders are watching is Oracle's AI World conference on October 13-16 - a likely stage for product detail, customer announcements and timing on when backlog might flow into revenue.

So what's the playbook for market participants? No directives here - just patterns to watch. Price-target upgrades often fuel momentum, but the key tension is converting booking growth into sustained revenue and healthy margins. A huge RPO number is a promising read on demand; it isn't the same as cash from customers. And the margin profile on GPU/compute rentals matters if Oracle is to translate scale into profit.

Short, blunt takeaway: the numbers give a strong narrative for Oracle's role in AI infrastructure, and analysts have rewarded that view with higher targets - yet there are legitimate questions about conversion timing and profitability. Will the backlog convert fast enough and at attractive margins to justify these loftier valuations?