Oracle’s AI Moonshot Meets a Reality Check as Shares Slide

CNBC and Investor’s Business Daily contributed to this report.

Oracle’s stock rally finally hit a speed bump. After weeks of AI-fueled euphoria, investors tapped the brakes Friday, sending the shares down about 7% — Oracle’s worst day since January — after the company amped up already-lofty targets at its AI World shindig in Las Vegas.

On Thursday, Oracle laid out a supersized roadmap for the end of the decade:

• AI revenue: from $3B in FY2026 to $20B in FY2030.
• Cloud infrastructure (OCI): from $18B in FY2026 to a staggering $166B in FY2030.
• Total revenue & earnings: $225B in FY2030 and $21 in adjusted EPS — implying ~31% annual sales growth and ~28% EPS growth.

Wall Street loved it — briefly. The stock popped 3.1% on Thursday and has soared more than 160% over two years on the idea that Oracle is becoming a core platform for AI model training and inference. But by Friday, skepticism crept in.

“It feels like the stock may take a bit of a breather here,”

— Rishi Jaluria, RBC Capital Markets, telling CNBC that investors need time to vet how achievable those long-term numbers really are. He’s at Hold.

Expectations were sky high. Oracle’s long-term growth math is thrilling; execution is another story. Building an AI-scale cloud requires massive data center expansion, specialized chips, power, and supply-chain coordination — all capital-intensive.

• Capex fog: The company didn’t offer much detail on forward capex needs. Jefferies’ Brent Thill warned spending must ramp with OCI growth, raising financing questions. Consensus sees > $26B in negative free cash flow over the next three fiscal years. Oracle already raised $18B in bonds last month; more funding likely.
• Concentration risk: Bulls love the marquee logos, but skeptics flag heavy reliance on mega-customer pipelines. UBS’s Karl Keirstead (Buy; PT to $380) still sees upside, but his “bear case” highlights potential overexposure to OpenAI and the practical snags that come with hyper-aggressive buildouts.
• Lofty starting point: After a monster run — including the best post-earnings day since 1992 last month — Oracle had little room for disappointment.

Plenty of analysts still see a long runway:

• Guggenheim’s John DiFucci reiterated Oracle as a “best idea,” lifting his PT to $400, and highlighted 30–40% gross margins for AI infrastructure over the life of contracts—better than many expected.
• Barclays’ Raimo Lenschow (Overweight) also nudged his PT to $400, noting the FY2030 guide is ahead of prior Street models.
• Evercore ISI’s Kirk Materne (Outperform; PT $385) said Oracle is “firing on all cylinders,” but acknowledged the stock had already rallied ~30% since the last earnings print.
• Stifel’s Brad Reback (Buy; PT $350) expects consolidation as investors digest the capex-heavy pivot and watch for operating leverage to hold.

And Oracle is working to defuse the “it’s just OpenAI” narrative. Co-CEO Clay Magouyrk said this quarter’s fresh $65B in OCI commitments came via seven contracts across four customers — none of them OpenAI. The company also confirmed a cloud deal with Meta and keeps pointing to that mega RPO backlog of $455B, up 359% year over year.

Because lofty targets raise the bar — and the bill. Turning $18B of OCI into $166B in five years implies relentless delivery, a smooth buildout, and no major supply chain or power hiccups. Investors want a cleaner line of sight on spend, timing, and payback before chasing new highs.

By the close, Oracle shares were off roughly 7% around $291 (they fell more than 8% intraday), a reset that looks more like digestion than a trend reversal.

Oracle just told the world it plans to be an AI cloud superpower by 2030. The Street didn’t balk at the ambition — it balked at the execution gap. If management can demystify capex, show diversified demand beyond the headliners, and keep margins near the 30–40% guide on AI infra, the bull story stays intact. Until then, even great narratives need to catch their breath.