Oracle Stock Falls, Down 57% From Its Peak: AI Bargain or Debt-Fueled Trap?

Oracle's stock has fallen around 57% from its peak of $345.72, hit in September 2025. In the same period, the company reported its strongest revenue growth in 15 years, signed AI infrastructure contracts worth hundreds of billions of dollars, and guided for 34% revenue growth in FY27. 

Oracle has a $638 billion backlog, growing cloud infrastructure revenue. Yet the stock keeps falling. Understanding why requires moving past the capex headline and into the structure beneath it.

Let's break down what is actually happening at Oracle, why the market is more worried than the revenue numbers suggest, what could prove the bears catastrophically wrong, and what any investor should understand before forming a view on Oracle (ORCL) stock.

Why is Oracle Stock Falling?

Oracle's business numbers for FY26 are genuinely strong. The selloff in Oracle stock is not about revenue growth but who is paying for it.

These are not soft numbers dressed up with adjustments. Revenue growth above 20% at Oracle's scale, sustained over multiple quarters, is a real operating outcome. The problem is the balance sheet statement two pages later. 

To fund the gap, Oracle raised $43 billion in debt and $5 billion in equity during FY26 alone. Total debt now exceeds $100 billion. And for FY27, the company expects net capex of approximately $70 billion, with plans to raise another $40 billion in debt and equity, including a $20 billion at-the-market equity issuance.

A $20 billion ATM program means Oracle plans to sell new shares into the open market. At a stock price of ~$147, that is significant dilution. Existing shareholders see their percentage ownership shrink with each share issued, and the fact that Oracle chose equity over pure debt is itself a signal, the company's CFO said so on the earnings call when she noted that the ATM program was structured to protect Oracle's BBB credit rating.

That last detail is where the story turns structural rather than cyclical.

Why Oracle’s AI Bet Carries a Different Kind of Risk

There is a useful way to think about what Oracle is attempting. Call it the BBB Hyperscaler Problem.

• Meta funded its 2025 infrastructure buildout from $115.8 billion in operating cash flow and remained free cash flow positive at $43.6 billion for the year. 
• Alphabet (Google) deployed well over $270 billion in capital spending over two years while retaining the cash generation of Search and YouTube. 

These companies are building at hyperscaler scale because they have hyperscaler-scale cash generation to match.

Oracle is building at hyperscaler scale too. FY27's $70 billion net capex estimate puts Oracle in the same numerical league as Meta. The difference is the funding source. Oracle's $32 billion in operating cash flow covers only part of the FY27 commitment. The rest comes from bond markets and equity dilution, from a balance sheet already rated BBB, with Moody's and S&P both carrying negative outlooks.

This is not a fatal flaw. It is a risk parameter. But the market priced Oracle in September 2025 as if it were the first developer. The stock correction since then is, in part, a reassessment of which category Oracle actually belongs in.

Oracle’s $638B Backlog: Demand, Delay and Customer Concentration

Oracle's Remaining Performance Obligations (RPO) reached $638 billion at the end of FY26, up 363% year over year. This is the single most cited number in every bullish Oracle case, and it deserves careful unpacking rather than simple citation.

RPO is real. It represents contracted future revenue, meaning the work Oracle has signed agreements for. It is not speculative. The $85 billion jump in a single quarter reflects contracts actually signed, not pipeline estimates. For context, this figure now exceeds Microsoft's disclosed backlog.

However, three caveats matter:

• The conversion timeline is long: Oracle’s backlog is huge, but the cash conversion is slow. Only about 12% of its $638 billion RPO is expected to become revenue in the next 12 months, leaving $560+ billion spread over several years. Oracle is spending on infrastructure today, funded through debt and equity raises, while the cash returns come much later.
• The concentration is significant: Analysts estimate that over half of Oracle’s RPO may come from one customer: OpenAI, mainly through the Stargate AI infrastructure project. The $300 billion OpenAI relationship has been widely reported. Oracle knows this risk and is trying to diversify through multicloud partnerships with Amazon Web Services (AWS) and Google Cloud, a Meta deal, and other enterprise customers.
• OpenAI's own financial position is a factor: OpenAI's reported revenue was around $12 billion in mid-2025 and growing, but even at $20 billion annually, Oracle’s implied pricing looks hard to support without continued external funding. As DA Davidson’s Gil Luria put it, Oracle is building for a customer whose ability to pay depends on that customer’s own fundraising success.

This is the structural version of counterparty risk. It is not that OpenAI will default tomorrow. It is that the entire financial model requires OpenAI's business to grow faster than Oracle's debt service costs accumulate.

Oracle Financials: Record Growth Meets a Cash Flow Shock

Sources: Oracle earnings release (SEC Form 8-K); Oracle earnings call transcript. 

The takeaway is that FY26 was a year of genuine operational strength (revenue, EPS, operating cash flow all hit records) and genuine financial stress (FCF turned sharply negative, debt surged). Both things are simultaneously true.

Oracle’s Growth Engine OCI Comes With a Margin Trade-Off

Oracle's cloud business now splits cleanly into two segments with very different dynamics.

• Cloud Infrastructure (IaaS / OCI): FY26 revenue of $18.1 billion, up 77%. Q4 alone saw 93% year-over-year growth. This is Oracle's fastest-growing segment and the core of its AI story. OCI (Oracle Cloud Infrastructure) is winning share because, unlike AWS or Azure, it has been willing to build bespoke, multi-gigawatt data centers to hyperscaler-scale customer specifications. This wins large contracts but compresses margins because it requires enormous upfront capital.
• Cloud Applications (SaaS): FY26 revenue of $15.9 billion, up 11%. This is Oracle's Fusion Cloud ERP, NetSuite, and healthcare software business. It is growing steadily, carries structurally better margins than infrastructure, and generates recurring subscription revenue with high retention. It is the part of Oracle that the pre-AI Oracle story was built on.
• Legacy Software Licensing: FY26 revenue of $24.5 billion, down 1%. This is Oracle's on-premise database and software licensing business. It is slowly declining as customers migrate to cloud, but it is cash-generative and supports overall margin structure.

The critical issue is that OCI's rapid scaling, while driving the top-line growth, is also compressing margins. Oracle's CFO explicitly said gross margins would "step down" in FY27 as new data centers ramp their revenue contribution. In other words, the infrastructure is being built now, costs are being incurred now, but the full contractual revenue won't arrive until later.

This margin dip is temporary by design but the timing of the recovery is precisely what investors cannot price with confidence.

Oracle vs Cloud Giants: Faster Growth, Weaker Balance Sheet

Note: AWS, Google Cloud, and Azure figures are approximate estimates from public company disclosures. 

Oracle's growth rate is genuinely superior. The structural question is whether its capital position can sustain that growth rate long enough for the cash flow to follow.

Oracle Stock Price Predictions: Why Wall Street Is Split

Sources: Benzinga, Tickflow, Yahoo Finance and Reuters

What explains the spread? The $160 to $400 price target range is one of the widest for any large-cap stock. The spread reflects one fundamental disagreement: does Oracle's RPO represent real, monetizable, near-term contracted revenue (bullish view), or does it represent a long-dated, concentrated, credit-dependent obligation that the balance sheet cannot comfortably support (bearish view)? 

Both positions are internally consistent. The outcome depends on variables like OpenAI's revenue growth, interest rate trajectory, Oracle's ability to execute $70B of capex flawlessly, that no analyst can resolve today with certainty.

Oracle’s Cheap PE May Be Misleading

At roughly $147 per share, Oracle's forward price-to-earnings (PE) ratio calculated against its own FY27 non-GAAP EPS guidance of $8.05, comes out to approximately 18.3 times. That number will surprise most people who have been following the stock's decline. 

A company guiding for 34% revenue growth and 18% EPS growth (in constant currency, excluding one-time FY26 gains) trading at 18.3 times forward earnings sounds cheap. It is cheaper than Microsoft, which trades at approximately 19 to 20 times forward earnings while growing revenue at roughly half Oracle's guided rate.

Sources: Oracle earnings release (SEC Form 8-K), GuruFocus, S&P Global Market Intelligence, MacroTrends, Microsoft, Salesforce investor relations.

On PE alone, Oracle looks like the better value. Growth twice as fast, priced lower. So why is the stock down 57% from its peak?

Oracle’s $40B Earnings-to-Cash Gap Investors Cannot Ignore

Oracle’s income statement and cash flow statement are telling different stories. There is a $40.7B gap between accounting profit and actual cash flow, mainly because GAAP earnings deduct depreciation over time, not the full cash spent on capex.

So, Oracle’s $55.7B capex does not fully hit earnings immediately, even though the cash has already gone out. That is why PE can be misleading here. PE uses earnings, not cash. When earnings and cash flow are $40B apart, PE does not show what the business is actually generating.

In simple terms, two companies can show the same profit, but if one is generating cash and the other is burning cash to fund expansion, they are not equally strong. That is Oracle’s issue today: earnings are real, but cash conversion is weak.

Why Oracle’s Valuation Story Really Depends on FY29

The market is not really debating whether Oracle can reach $8.05 non-GAAP EPS in FY27. With current revenue guidance, that looks achievable. The bigger question is when its data centers start generating full contracted revenue, capex pressure eases, and FCF turns meaningfully positive.

• Bulls call this a potential “cash flow waterfall” from around FY29. If Oracle executes well, future FCF could support a higher valuation.
• Bears argue FY29 is still three years away. Until then, Oracle must carry rising interest costs on $100B+ debt, while the $20B ATM equity program could dilute shareholders.

Takeaway: Oracle is not expensive on PE. It is priced on the expectation that cash flow will eventually catch up with earnings. Whether that happens, and how long it takes, is the entire investment thesis.

Oracle Bull Case: Backlog, OCI Momentum and AI Demand

• Demand is locked in: Oracle’s RPO jumped $85B in one quarter to $638B by FY26-end, contracted future revenue, not a speculative pipeline.
• OCI growth is far ahead of peers: Cloud infrastructure revenue grew 93% YoY in Q4 FY26 on an $18B annual IaaS base, compared with AWS growth of around 19%.
• Oracle is winning a specific AI niche: It is building large, bespoke, multi-gigawatt AI training clusters where customer-specific infrastructure gives it an edge.
• Multicloud adds stickiness: Deals with AWS and Google Cloud make Oracle databases easier to use across clouds, reducing the need for customers to migrate away.
• FY27–FY29 upside is still real: Oracle guided for $90B FY27 revenue, up 34%, while bulls see a possible FY29 “cash flow waterfall” if contracts convert into cash.

The bull case is simple: Oracle already has the demand. The question is whether it can convert that demand into cash before debt, capex, and dilution pressure catch up.

Oracle Bear Case: Debt, Dilution and OpenAI Dependency

• Backlog concentration: Oracle’s $638B RPO looks strong, but analysts estimate that more than half may be linked to OpenAI.
• OpenAI’s funding risk: OpenAI spends far more than it earns and depends on external funding. If its growth or fundraising slows, Oracle’s backlog assumptions could weaken.
• Financing pressure is rising: Some banks avoided Stargate-linked data center financings where Oracle was the anchor tenant, citing concentration risk and Oracle’s weaker credit profile.
• FY27 capex execution risk: Oracle plans around $70B of FY27 capex. If financing tightens for its infrastructure partners, data center execution could become harder.
• Dilution could hurt shareholders: Oracle’s $20B ATM equity program could issue over 136M new shares at $147/share, equal to around 4.7% dilution.
• Debt risk: Oracle needs the equity program to protect its BBB credit rating. A downgrade could increase borrowing costs across its $100B+ debt stack.
• Margins may stay under pressure: Management guided for gross margin to “step down” in FY27 because data center costs start before revenue fully ramps.
• Margin recovery: If gross margins do not recover, the bull case remains based more on execution faith than proven financial results.

The bear case is simple: Oracle has demand, but it is paying heavily upfront, relying on concentrated customers, rising debt, possible dilution, and future margin recovery to make the math work. 

Bottom Line: Oracle Stock - AI Bargain or Balance Sheet Trap?

A 57% fall from peak in a company posting 93% cloud infrastructure growth looks like a textbook buying opportunity. At around 18x forward earnings, Oracle trades below its five-year average of 34x and roughly in line with Microsoft, despite growing faster. That headline valuation will attract value investors.

But PE misses the core problem: Oracle has a $40B gap between earnings and cash, a $100B+ debt stack, and a $20B equity dilution program still ahead. These risks have not gone away because the stock has fallen. They are still unfolding.

There is credible downside risk from here. Q1 FY27 earnings in August/September will be the first real test of whether gross margins are stabilizing or weakening further. A deeper margin step-down, active ATM equity issuance, or any negative update on OpenAI’s funding trajectory could quickly reprice the RPO-led bull case.

Oracle is not a once-in-a-decade bargain. It is a leveraged AI infrastructure bet at a lower valuation. The upside is real, but investors may need to sit through 18–24 months of negative FCF, margin pressure, and dilution headlines before the FY29 cash flow story plays out.

The risk is treating low PE as a safety net. It is not. Until gross margins stabilize and FCF starts recovering, Oracle remains a high-conviction bet, not a clear buy signal.