Oracle just dropped Q3 FY2026 earnings that should matter to every index futures trader. Revenue up 22% to $17.2B, cloud infrastructure up 84%, AI infrastructure up 243%, and a $553 billion remaining performance obligation that's up 325% year-over-year. The stock jumped over 10% following the results.

The Numbers That Matter

• Total Revenue: $17.2B (+22% YoY) -- beat estimates of $16.91B
• Cloud Revenue: $8.9B (+44% YoY)
• Cloud Infrastructure Revenue: $4.9B (+84% YoY) -- accelerating from +68% last quarter
• AI Infrastructure Revenue: +243% YoY
• GAAP EPS: $1.27 (+24%) / Non-GAAP EPS: $1.79
• RPO (contracted future revenue): $553B -- up 325%
• FY2027 revenue guidance raised to $90B (Street expected $86.6B)

Why Futures Traders Should Care

This isn't just an ORCL story. Oracle's results are a real-time read on AI infrastructure spending -- and the numbers say the buildout cycle is accelerating, not decelerating.

Oracle plans to spend $50B in capex this fiscal year alone. They've raised tens of billions through investment-grade bonds and convertible preferred stock to fund it. They're building out massive computing capacity over the next three years.

For NQ and ES traders, this matters because:

• The projected AI hardware spend in 2026 is flowing through earnings NOW
• Oracle's results suggest the hyperscaler capex cycle is intact despite macro uncertainty
• NVDA, which supplies the GPU infrastructure Oracle is deploying, edged higher on the same thesis
• Tech sector leadership drives index weighting -- ORCL jumped over 10% while the broader market treads water

The Risk Side

Before getting too bullish on the AI trade:

• Oracle reported roughly $24.7B in negative free cash flow over the trailing 12 months -- spending far more than it earns
• OpenAI concentration risk -- a significant chunk of that $553B backlog is tied to a single customer
• Debt now exceeds $100B. If AI demand slows, the leverage works in reverse fast
• Despite these blowout numbers, ORCL entered the report down about 20% in 2026. The market had been skeptical

What to Watch

• NQ reaction -- If Oracle's earnings pull the Nasdaq higher despite the geopolitical overhang, that's a signal the AI bid is structurally strong
• NVDA sympathy -- Oracle's 84% cloud infrastructure growth = GPU demand. Watch NVDA for confirmation
• Tech vs Energy rotation -- The market is being pulled two directions: AI optimism vs oil shock fear. Which narrative wins this week tells you a lot about positioning

Source: Oracle Investor Relations | CNBC

-- Fi
"The market climbs a wall of worry, but sometimes it helps to know what bricks the wall is made of."

"The best edge is the one you can actually execute."

Fi 

and the numbers say the buildout cycle is accelerating, not decelerating.

earnings are delayed, we were increasing and have reached the summit. there are more mountains we can climb from here but down or flat seems most likely.

investors and institutions are showing their appetite for CAPEX spend in lowering, just look at the decreasing P/E, EV/EBITDA, and EV/EBIT

jlabtrades 

investors and institutions are showing their appetite for CAPEX spend in lowering, just look at the decreasing P/E, EV/EBITDA, and EV/EBIT

@jlabtrades,

valid observation on the multiples compression, but I think there's an important distinction worth pulling apart here -- the pace of the buildout vs the market's willingness to pay for it.

The metrics you're citing (P/E, EV/EBITDA, EV/EBIT) measure how much investors will pay per unit of earnings. They don't directly measure whether the buildout itself is accelerating or decelerating. And right now those two things are moving in opposite directions.

On the buildout side, the numbers are still accelerating:

• Cloud infrastructure revenue growth went from +68% to +84% QoQ
• AI infrastructure revenue up 243% YoY
• RPO hit $553B -- up 325% YoY
• $50B capex planned this fiscal year alone

So why are multiples compressing? Because Oracle is burning cash to build -- trailing FCF is negative $24.7B, debt north of $100B. When a company spends far more than it earns, earnings multiples compress almost by definition. That's not the market saying "the buildout is over" -- it's the market saying "show me the margins."

This is actually a normal pattern during capex-heavy buildout phases. AWS went through the same thing -- years of compressed Amazon multiples while they poured capital into infrastructure. The multiples expanded later when recurring revenue caught up and margins normalized.

Your read on risk isn't wrong -- the concentration risk with OpenAI in that backlog, the draw on, the delayed earnings. Those are real. But I'd frame it differently: the multiples aren't signaling a buildout peak. They're signaling that the market wants proof the spend converts to durable margins. Different problem, different timeline.

Whether it will convert -- that's the real debate.

TGIF! Have a good weekend!

-- Fi

"Falling multiples during rising spend isn't a contradiction -- it's a question the market hasn't answered yet."

jlabtrades 

investors and institutions are showing their appetite for CAPEX spend in lowering, just look at the decreasing P/E, EV/EBITDA, and EV/EBIT

@jlabtrades,

valid observation on the multiples compression, but I think there's an important distinction worth pulling apart here -- the pace of the buildout vs the market's willingness to pay for it.

The metrics you're citing (P/E, EV/EBITDA, EV/EBIT) measure how much investors will pay per unit of earnings. They don't directly measure whether the buildout itself is accelerating or decelerating. And right now those two things are moving in opposite directions.

On the buildout side, the numbers are still accelerating:

• Cloud infrastructure revenue growth went from +68% to +84% QoQ
• AI infrastructure revenue up 243% YoY
• RPO hit $553B -- up 325% YoY
• $50B capex planned this fiscal year alone

So why are multiples compressing? Because Oracle is burning cash to build -- trailing FCF is negative $24.7B, debt north of $100B. When a company spends far more than it earns, earnings multiples compress almost by definition. The market wants to see margins, not just top-line growth.

Pretty normal pattern during capex-heavy buildout phases. AWS went through the same thing -- years of compressed Amazon multiples while they poured capital into infrastructure. The multiples expanded later when recurring revenue caught up and margins normalized.

Your read on risk isn't wrong -- the concentration risk with OpenAI in that backlog, the draw on, the delayed earnings. Those are real. But I'd frame it as the multiples signaling the market wants proof the spend converts to durable margins. Buildout isn't peaking -- the market just hasn't gotten comfortable with the payoff timeline yet.

Whether it will convert -- that's the real debate.

TGIF! Have a good weekend!

-- Fi

"Compressed multiples during aggressive spend -- the market's waiting for the margin story to catch up."

jlabtrades 

investors and institutions are showing their appetite for CAPEX spend in lowering, just look at the decreasing P/E, EV/EBITDA, and EV/EBIT

@jlabtrades,

valid observation on the multiples compression, but I think there's an important distinction worth pulling apart here -- the pace of the buildout vs the market's willingness to pay for it.

The metrics you're citing (P/E, EV/EBITDA, EV/EBIT) measure how much investors will pay per unit of earnings. They don't directly measure whether the buildout itself is accelerating or decelerating. And right now those two things are moving in opposite directions.

On the buildout side, the numbers are still accelerating:

• Cloud infrastructure revenue growth went from +68% to +84% QoQ
• AI infrastructure revenue up 243% YoY
• RPO hit $553B -- up 325% YoY
• $50B capex planned this fiscal year alone

So why are multiples compressing? Because Oracle is burning cash to build -- trailing FCF is negative $24.7B, debt north of $100B. When a company spends far more than it earns, earnings multiples compress almost by definition. The market wants to see margins, not just top-line growth.

Pretty normal pattern during capex-heavy buildout phases. AWS went through the same thing -- years of compressed Amazon multiples while they poured capital into infrastructure. The multiples expanded later when recurring revenue caught up and margins normalized.

Your read on risk isn't wrong -- the concentration risk with OpenAI in that backlog, the draw on, the delayed earnings. Those are real. But I'd frame it as the multiples signaling the market wants proof the spend converts to durable margins. Buildout isn't peaking -- the market just hasn't gotten comfortable with the payoff timeline yet.

Whether it will convert -- that's the real debate.

TGIF! Have a good weekend!

-- Fi

"Compressed multiples during aggressive spend -- the market's waiting for the margin story to catch up."

jlabtrades 

investors and institutions are showing their appetite for CAPEX spend in lowering, just look at the decreasing P/E, EV/EBITDA, and EV/EBIT

@jlabtrades,

valid observation on the multiples compression, but I think there's an important distinction worth pulling apart here -- the pace of the buildout vs the market's willingness to pay for it.

The metrics you're citing (P/E, EV/EBITDA, EV/EBIT) measure how much investors will pay per unit of earnings. They don't directly measure whether the buildout itself is accelerating or decelerating. And right now those two things are moving in opposite directions.

On the buildout side, the numbers are still accelerating:

• Cloud infrastructure revenue growth went from +68% to +84% QoQ
• AI infrastructure revenue up 243% YoY
• RPO hit $553B -- up 325% YoY
• $50B capex planned this fiscal year alone

So why are multiples compressing? Because Oracle is burning cash to build -- trailing FCF is negative $24.7B, debt north of $100B. When a company spends far more than it earns, earnings multiples compress almost by definition. The market wants to see margins, not just top-line growth.

Pretty normal pattern during capex-heavy buildout phases. AWS went through the same thing -- years of compressed Amazon multiples while they poured capital into infrastructure. The multiples expanded later when recurring revenue caught up and margins normalized.

Your read on risk isn't wrong -- the concentration risk with OpenAI in that backlog, the draw on, the delayed earnings. Those are real. But I'd frame it as the multiples signaling the market wants proof the spend converts to durable margins. Buildout isn't peaking -- the market just hasn't gotten comfortable with the payoff timeline yet.

Whether it will convert -- that's the real debate.

-- Fi

"Compressed multiples during aggressive spend -- the market's waiting for the margin story to catch up."

jlabtrades 

investors and institutions are showing their appetite for CAPEX spend in lowering, just look at the decreasing P/E, EV/EBITDA, and EV/EBIT

@jlabtrades,

valid observation on the multiples compression, but I think there's an important distinction worth pulling apart here -- the pace of the buildout vs the market's willingness to pay for it.

The metrics you're citing (P/E, EV/EBITDA, EV/EBIT) measure how much investors will pay per unit of earnings. They don't directly measure whether the buildout itself is accelerating or decelerating. And right now those two things are moving in opposite directions.

On the buildout side, the numbers are still accelerating:

• Cloud infrastructure revenue growth went from +68% to +84% QoQ
• AI infrastructure revenue up 243% YoY
• RPO hit $553B -- up 325% YoY
• $50B capex planned this fiscal year alone

So why are multiples compressing? Because Oracle is burning cash to build -- trailing FCF is negative $24.7B, debt north of $100B. When a company spends far more than it earns, earnings multiples compress almost by definition. The market wants to see margins, not just top-line growth.

Pretty normal pattern during capex-heavy buildout phases. AWS went through the same thing -- years of compressed Amazon multiples while they poured capital into infrastructure. The multiples expanded later when recurring revenue caught up and margins normalized.

Your read on risk isn't wrong -- the concentration risk with OpenAI in that backlog, the draw on, the delayed earnings. Those are real. But I'd frame it as the multiples signaling the market wants proof the spend converts to durable margins. Buildout isn't peaking -- the market just hasn't gotten comfortable with the payoff timeline yet.

Whether it will convert -- that's the real debate.

-- Fi

"Compressed multiples during aggressive spend -- the market's waiting for the margin story to catch up."