Oracle plunges as AI spending blows past forecasts, $50B capex plan rattles Wall Street

Oracle tumbled after capex blew past estimates and full-year spend rose, while revenue missed. Keep an eye on capex cadence, backlog conversion, OCI margins, and credit spreads.

Categorized in: AI News Finance
Published on: Dec 12, 2025
Oracle plunges as AI spending blows past forecasts, $50B capex plan rattles Wall Street

Oracle stock sinks as AI build costs surge past estimates - what finance teams should track next

Oracle shares fell about 11% after the company posted a heavy AI spend and a slight revenue miss. The headline: capex shocked to the upside, guidance for more spending accelerated, and investors repriced the equity and credit risk in one session.

The print at a glance

  • Capex: $12B in fiscal Q2 vs. ~$4B last year and ~@$8B expected. Full-year capex guide raised to $50B from $35B.
  • Revenue: $16.06B, +14% y/y but below $16.21B consensus.
  • EPS (adj.): $2.26, beating $1.64 consensus and up from $1.47 a year ago.
  • OCI: $4.1B revenue, +68% y/y, roughly in line with estimates.
  • RPO: $523B, up ~440% y/y and 15% q/q; driven by commitments from Meta, Nvidia, and others.
  • Longer-term guide: FY27 revenue outlook raised by $4B to $89B.
  • Stock move: biggest daily loss since January; shares are ~40% off September peak while the "Magnificent Seven" are up ~10% over that span (per Bloomberg).

Why the market sold first

Capex at ~$12B in the quarter is roughly 75% of quarterly revenue - a pace that pulls forward cash needs and compresses near-term free cash generation. Management also lifted full-year capex plans to $50B, signaling the spend is not a one-off.

Revenue fell a touch short, which weakened the "growth funds the build" narrative, even with a solid EPS beat. Add in concern over debt-funded data center projects and a high reliance on OpenAI for targets, and risk appetite faded fast.

Growth is there - cash is the constraint

OCI is scaling with 68% y/y growth and a swelling backlog. The $523B RPO builds a case for future revenue, but conversion timing, margins, and customer concentration matter more than the headline number.

Oracle cited new multiyear commitments from major AI customers. That supports the capacity build, but the market wants evidence that these obligations translate into high-quality, high-margin revenue on a sane capex cadence.

Credit lens

The cost to insure Oracle's debt rose to the highest level since January 2009, according to Intercontinental Exchange, pointing to a higher implied risk premium on funding. That can flow through to interest expense and lowers flexibility if the spending cycle lengthens or demand slows.

Oracle investor relations offers the primary source materials and updates. For reference on CDS dynamics and pricing methodology, see Intercontinental Exchange credit data.

Key debates for your model

  • Capex cadence and utilization: How quickly does $50B of planned spend translate into usable capacity and revenue? What's the slope of capex-to-revenue over the next four quarters?
  • RPO burn: What portion of the $523B converts within 12-24 months, at what gross margin, and with what energy and depreciation drag?
  • Customer concentration: How sensitive are bookings and backlog to OpenAI's trajectory and funding environment? What is the exposure to a single program slip?
  • OCI margins: Track unit economics as data centers fill. Early ramps are margin-dilutive; inflection requires high utilization and disciplined pricing.
  • Balance sheet: Watch debt mix, maturities, and interest expense. The CDS move implies higher funding costs; update WACC and coverage ratios accordingly.

What would change sentiment

  • Evidence of backlog converting to revenue on time with improving OCI gross margin.
  • Clear path from $50B capex to capacity, utilization, and unit economics - plus a taper or glide path once the current build cycle stabilizes.
  • Less reliance on any single AI partner via diversified commitments and disclosed contract economics.
  • Stable or tightening credit spreads as the market gains confidence in cash generation.

Positioning considerations

  • Model higher near-term capex and interest expense; stress test slower backlog conversion.
  • Sensitivity test scenarios where OpenAI demand underdelivers or ramps later than expected.
  • For credit, track spreads and rating commentary; assess liquidity against data center commitments.

Bottom line

Oracle is building for AI demand at full speed, and the bill is arriving ahead of the cash. Growth is real; the question is pacing - how quickly spending turns into high-quality revenue and free cash flow without leaning harder on debt. Until there's proof on conversion and margins, the market will price the risk.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)
Advertisement
Stream Watch Guide