Oracle Earnings Today: Q2 2026 Results Put AI Cloud Boom — and Debt Fears

Oracle’s fiscal Q2 2026 earnings, released after the close on December 10, 2025, delivered exactly what everyone wanted to see on the surface: big AI-driven cloud growth, a massive backlog, and a huge earnings beat. Underneath, though, the numbers also intensify questions about debt, cash burn, and reliance on OpenAI — and that’s why the stock is trading lower after hours.

Oracle (ticker: ORCL) reported:

Revenue: $16.1 billion, up 14% year over year
Non-GAAP EPS: $2.26, up 54%
Total cloud revenue (IaaS + SaaS): $8.0 billion, up 34%
Cloud infrastructure (IaaS): $4.1 billion, up 68%
Remaining Performance Obligations (RPO): $523 billion, up 438% year over year and 15% sequentially ^[1]

Yet ORCL shares are down more than 5% in after-hours trading around $210.85, as investors label the report “mixed” and focus on quality of earnings and balance-sheet risk rather than just the headline beats. ^[2]

Below is a breakdown of what Oracle reported today, how it compares to expectations, and what it really means for the AI and cloud trade.

1. Oracle Q2 FY2026 by the numbers

Headline figures (GAAP and non-GAAP)

According to Oracle’s official release, for the quarter ended November 30, 2025: ^[3]

Total revenue: $16.1 billion, up 14% in USD (13% in constant currency)
Cloud revenue (IaaS + SaaS): $8.0 billion, up 34%
- Cloud Infrastructure (OCI/IaaS): $4.1 billion, up 68%
- Cloud Applications (SaaS): $3.9 billion, up 11%
- Fusion Cloud ERP: $1.1 billion, up 18%
- NetSuite Cloud ERP: $1.0 billion, up 13%
Software revenue (non‑cloud): down 3% to $5.9 billion
GAAP operating income: $4.7 billion
Non-GAAP operating income: $6.7 billion, up 10%
GAAP EPS: $2.10, up 91%
Non-GAAP EPS: $2.26, up 54%

Cash and backlog:

Remaining Performance Obligations (RPO): $523 billion, up 438% year over year and up $68 billion sequentially, driven by new commitments from customers including Meta and NVIDIA ^[4]
Short-term deferred revenue: $9.9 billion
Trailing 12‑month operating cash flow: $22.3 billion, up 10%

Ampere sale turbocharges EPS

Both GAAP and non‑GAAP EPS were boosted by a one‑off gain: Oracle booked a $2.7 billion pre‑tax gain from selling its stake in chip firm Ampere. Management explicitly notes this Ampere sale helped lift earnings per share in the quarter. ^[5]

That detail matters: the earnings beat looks spectacular, but a big slice of it comes from a non‑recurring gain, not just core operations.

2. How the results compared with expectations

Heading into today’s release, Wall Street expectations for Q2 FY2026 were clustered around:

Revenue: roughly $16.15–$16.21 billion, up about 15% year over year ^[6]
Adjusted EPS: about $1.63–$1.65, up low double digits from $1.47 a year ago ^[7]

Options markets were pricing in a ~10% move in either direction, highlighting how “event‑risk‑y” this print was expected to be. ^[8]

On that scorecard:

Revenue: At $16.1 billion, Oracle landed very close to consensus, arguably just a hair light versus the most optimistic forecasts but well within the expected range. ^[9]
EPS: Headline non‑GAAP EPS of $2.26 looks like a monster beat versus the ~$1.64 consensus — but again, that includes the Ampere gain. Without that one‑off lift, the underlying EPS outperformance is much smaller, even if still positive. ^[10]

So you can think of this quarter as:

“Revenue roughly in line, EPS optically huge, but boosted by a big asset sale rather than pure operating leverage.”

That nuance goes a long way in explaining the stock’s muted — and now negative — post‑earnings reaction.

3. Cloud and AI: Oracle’s growth engine is still roaring

If you only looked at the cloud numbers, you’d think this was a dream quarter.

AI infrastructure is the star

Oracle’s cloud infrastructure (OCI) business — the compute, storage and networking that power AI training and inference — grew 68% year over year to $4.1 billion, far outpacing the broader company. ^[11]

That’s on top of Q1 FY2026, when OCI revenue surged 55% to $3.3 billion and total cloud revenue hit $7.2 billion, up 28%. ^[12]

Management has previously laid out an aggressive roadmap:

Targeting $18 billion in OCI revenue this fiscal year, up ~77%
Then ramping to $32 billion, $73 billion, $114 billion and $144 billion over the following four fiscal years ^[13]

Crucially, Oracle argues most of this multi‑year forecast is already “booked” in its RPO. That’s how you end up with a backlog that jumped 359% year over year to $455 billion in Q1, and now sits at $523 billion after Q2. ^[14]

AI mega‑deals and multicloud push

A big part of that backlog comes from a handful of mega AI deals:

A reported $300 billion, five‑year contract with OpenAI, one of the largest cloud deals on record ^[15]
Participation in the $500 billion “Stargate” AI super‑cluster project with SoftBank and OpenAI ^[16]
New commitments from Meta, NVIDIA and other hyperscale customers, highlighted in today’s RPO commentary ^[17]

At the infrastructure level, Oracle says it now has over 211 live and planned cloud regions, more than any of its big‑tech cloud competitors, and is over halfway through building 72 multicloud datacenters embedded inside AWS, Google Cloud and Microsoft Azure. Its multicloud database business is described as its fastest‑growing, up over 800% year on year in Q2. ^[18]

On the AI side, Oracle also stresses that all of the top five AI models are available on its cloud, and that it’s embedding AI into:

Its cloud infrastructure stack
The Autonomous Database and analytics layer
Front‑end applications (Fusion, NetSuite, Oracle Health, and more) ^[19]

In short: the AI/cloud growth story is absolutely still intact. But that’s only half the story investors care about right now.

4. The catch: debt, cash burn and OpenAI concentration

The other half of the story — and the one driving a lot of the market’s anxiety — is how Oracle is funding this AI land grab, and who it’s banking on to pay for it.

Debt‑funded AI buildout

Over the last year, Oracle has become a poster child for debt‑fueled AI capex:

It has raised about $18 billion in new bonds, pushing its debt‑to‑equity ratio above 400% and leaving it with a BBB‑level credit profile, well below peers like Meta and Alphabet. ^[20]
Analysts estimate outstanding debt around $105 billion, and IG notes Oracle is exploring another ~$38 billion in debt financing to keep building AI data centers. ^[21]
Capital expenditures for FY2026 are projected around $35 billion, up roughly 65% from prior levels. Free cash flow has turned negative since Q4 FY2025 as spending outruns cash generation. ^[22]

Debt markets have noticed. Oracle’s credit default swap spreads have climbed to their highest levels since 2009, signaling increased concern about its leverage and cash burn. ^[23]

Heavy reliance on OpenAI and AI “bubble” fears

At the same time, Oracle’s AI backlog is highly concentrated:

Reuters reports that a $300 billion OpenAI contract explains a huge portion of that 359% backlog spike, and that OpenAI itself remains unprofitable and is expected to spend more than $1 trillion on compute by 2030. ^[24]
MarketWatch has dubbed Oracle “the canary in the coal mine” for Big Tech’s debt‑fueled AI spending spree — arguing that if Oracle struggles to make the math work, it may signal trouble for the broader AI infrastructure boom. ^[25]

Pre‑earnings pieces from Saxo, IG and others were already sketching out the main investor worries:

Can Oracle turn RPO into actual cash fast enough to keep up with debt service and capex? ^[26]
How diversified is that backlog beyond OpenAI and a few mega‑customers? ^[27]
What happens if AI spending cools, or OpenAI stumbles?

Today’s Q2 numbers grew the backlog further and showcased robust cloud growth — but they didn’t fully calm those structural concerns.

5. Why the stock is down on “good” numbers

So why is ORCL trading down more than 5% after hours when the company just delivered:

Double‑digit revenue growth
A huge EPS beat
Accelerating cloud infrastructure growth
And another surge in backlog?

A few reasons:

1. Quality of the EPS beat

Investors are sensitive to the quality of earnings, not just the size of the beat. In this quarter:

The Ampere sale adds a large, non‑recurring boost to EPS.
Underlying EPS growth — the kind that comes from core operations — looks much more modest than the headline 54% non‑GAAP jump. ^[28]

Given how much AI optimism is already baked into the stock, traders wanted clean operational upside, not one‑time help.

2. Debt and cash flow worries haven’t gone away

Pre‑earnings coverage from Reuters, MarketWatch and others framed this report as a referendum on debt‑funded AI spending. ^[29]

Even with strong cloud growth, the market still has to grapple with:

A heavily leveraged balance sheet
Negative or thin free cash flow amid massive capex
The need for additional financing in coming years to deliver on multi‑hundred‑billion AI commitments

Today’s press release doesn’t fundamentally change that picture — it just confirms Oracle is pushing even harder on the AI gas pedal.

3. “Show me” mode for AI — especially after the Q1 spike & crash

Back in September, Oracle’s Q1 FY2026 results triggered a nearly 40% single‑day spike, as the huge RPO jump and AI deals pushed the stock to an all‑time high around $345. ^[30]

Since then:

The shares have fallen roughly 33–37% from that peak as investors questioned how much of the backlog is tied to OpenAI and how sustainable those commitments really are. ^[31]
Multiple analyses have called out ORCL as a bellwether for an AI “bubble”, with concerns that AI spending might be racing ahead of proven monetization. ^[32]

After that ride, the market is firmly in “show me” mode:

Show that backlog converts to revenue and cash at a healthy pace
Show that AI infrastructure margins can support debt service and still deliver attractive returns
Show that growth isn’t overly dependent on a single high‑risk partner

On that score, today’s numbers are encouraging but not decisive — hence the cautious reaction.

6. Key themes to watch on the earnings call

For investors and analysts listening to Oracle’s Q2 call, a few questions will likely dominate:

1. Backlog quality and conversion

How much of the $523 billion RPO is tied to OpenAI versus other customers?
What portion is contractually binding, and over what timeframes?
How quickly does Oracle expect to convert that backlog into recognized revenue and cash? ^[33]

2. Capex and financing plan

Updated guidance for FY2026 capex (currently projected around $35 billion). ^[34]
How much additional debt Oracle expects to raise, on what terms, and with what impact on its credit rating. ^[35]
Management’s roadmap for returning to positive free cash flow while still meeting AI demand.

3. AI customer mix and diversification

Evidence of new AI customers beyond OpenAI, xAI and Meta, and how diversified future AI workloads really are. ^[36]
Any commentary on contract structures — for example, minimum commitments vs. usage‑based pricing, and protections if customers pull back.

4. Margins in cloud infrastructure

How OCI gross and operating margins are trending as Oracle builds out gigawatt‑scale data centers and locks in long‑term AI contracts. ^[37]
Whether management still expects AI infrastructure contracts to land in the 30–40% margin range over time, as some analyses have suggested. ^[38]

7. What today’s earnings mean for AI and cloud investors

Oracle has become one of the purest public proxies for the AI infrastructure buildout:

Its AI backlog and cloud growth show how explosive demand for compute and data can be when enterprises go all‑in on gen‑AI. ^[39]
Its balance sheet and cash‑flow strain highlight how capital‑intensive that opportunity is — and how much financing risk is creeping into the AI story. ^[40]

For the broader AI and cloud complex — from NVIDIA (GPUs) to Meta, Microsoft, and Alphabet (hyperscale cloud) — Oracle’s report serves as:

A confirmation that AI workloads are scaling fast and flowing into the cloud
A warning that funding those workloads via heavy leverage can quickly become a market flashpoint

If Oracle ultimately proves that its backlog is high‑quality, diversified and cash‑generative, it strengthens the bullish case for AI infrastructure across the board. If it stumbles, it may fuel calls that the sector’s capex race has run ahead of fundamentals.

8. Bottom line

Today’s Oracle earnings are a Rorschach test for the AI trade.

Bulls will point to 68% OCI growth, 34% total cloud growth, and a $523 billion backlog as evidence that Oracle’s AI and multicloud strategy is working — and that its massive investments will pay off over time. ^[41]
Bears will point to debt‑funded capex, negative free cash flow, heavy OpenAI exposure and one‑off EPS boosts as signs that the risk side of the AI boom is rising just as fast as the opportunity. ^[42]

For now, the market’s initial verdict — a 5%+ drop in after‑hours trading — says investors still need more convincing that Oracle’s AI super‑cycle will translate into durable, high‑quality earnings and cash, not just spectacular headlines.

This article is for information and analysis only and does not constitute financial or investment advice. Anyone considering ORCL or related AI/cloud names should do their own research and, if needed, consult a qualified financial adviser.