Oracle is experiencing one of the strongest periods of revenue growth in its history. Oracle Cloud Infrastructure revenue increased 121% in the latest quarter, total cloud revenue increased 62%, consolidated revenue increased 30%, and remaining performance obligations reached an extraordinary $664 billion. On the surface, these figures make Oracle look like one of the strongest beneficiaries of the AI infrastructure boom.

The financial statements tell a considerably more complicated story. Oracle spent $28.5 billion on capital expenditures in a single quarter, generated negative free cash flow despite $23.1 billion of reported operating cash flow, carried roughly $125 billion of borrowings, issued 141 million new shares for $19.9 billion, and received $11.4 billion of customer prepayments that materially inflated reported operating cash flow. Meanwhile, property, plant and equipment has increased to $127.8 billion as Oracle rapidly transforms from an asset-light enterprise-software company into a capital-intensive AI infrastructure provider.

The question is therefore not whether Oracle can grow. It clearly can. The question is whether the returns Oracle eventually generates on this unprecedented infrastructure investment will adequately compensate shareholders for the debt, dilution, negative free cash flow, customer concentration and execution risk required to produce that growth.

Oracle's fundamental business is stronger than its current financial risk profile might suggest. Its database franchise remains deeply embedded across governments and large enterprises, Fusion and NetSuite provide recurring application revenue, and OCI has emerged as a legitimate hyperscale infrastructure competitor.

The problem is the price Oracle is paying to capture the AI opportunity.

Oracle has transitioned from monetizing high-margin intellectual property toward constructing enormous amounts of physical infrastructure. Servers, GPUs, networking equipment, data centers, power contracts and leases require enormous upfront capital before revenue can be recognized.

This changes the investment framework.

The relevant question is no longer simply how fast OCI grows. Investors need to determine the incremental return Oracle earns on every dollar invested in AI infrastructure.

My concern is that the financial risk is increasing faster than Oracle has demonstrated the long-term economics of these investments.

Fiscal Q1 2027 was operationally exceptional.

Revenue increased 30% to $19.35 billion, cloud revenue increased 62% to $11.61 billion and OCI infrastructure revenue increased 121% to $7.39 billion. GAAP operating income increased 57% to $6.73 billion, while GAAP EPS increased to $1.56.

The operating results themselves are not the bearish argument. The disconnect between accounting earnings, reported operating cash flow and actual capital requirements is.

Oracle's income statement currently looks excellent.

Revenue grew significantly faster than operating expenses, creating strong operating leverage.

But interest expense deserves particular attention. It increased 55% to $1.43 billion in one quarter as Oracle carried the financing burden associated with $43 billion of senior notes issued during FY2026.

Annualizing Q1 interest expense produces approximately $5.7 billion—not a trivial expense even for a company of Oracle's size.

Oracle is rapidly becoming a cloud infrastructure company.

Cloud generated $11.6 billion of Oracle's $19.3 billion quarterly revenue, meaning approximately 60% of total revenue now comes from cloud operations. OCI alone represented roughly 38%.

The growth is impressive, but the mix shift matters financially because infrastructure has very different economics from traditional software licensing.

Software requires relatively little incremental physical capital. AI infrastructure requires servers, accelerators, networking equipment, buildings, power and cooling systems.

Oracle's revenue growth is therefore becoming considerably more capital intensive.

OCI revenue reached $7.4 billion, increasing 121% year over year. Oracle also added approximately 850 MW of data-center capacity during Q1 and delivered more than 300,000 GPUs since Q4.

Demand is clearly not Oracle's current problem.

Supply is.

But this creates an important distinction between revenue growth and economic value creation. Oracle can grow OCI rapidly by deploying enormous amounts of capital into GPUs and data centers. That does not automatically mean those investments will produce attractive long-term returns.

The metrics I would prioritize are increasingly:

Oracle currently provides much more evidence for the first metric than the remaining seven.

Remaining performance obligations increased by $209 billion year over year to $664 billion, while Oracle signed more than $30 billion of additional AI cloud contracts during Q1.

This is an extraordinary amount of contracted revenue.

But RPO should not be confused with current revenue, profit or free cash flow.

A large AI infrastructure contract can create substantial future revenue while simultaneously requiring Oracle to commit enormous amounts of capital today.

That distinction is critical.

The economic value of $664 billion of backlog depends on the margins and capital intensity associated with fulfilling it.

Oracle's relationship with OpenAI provides both enormous upside and significant concentration risk.

Approximately $300 billion of Oracle's $664 billion RPO has been linked to OpenAI, meaning one customer represents an exceptionally large portion of Oracle's future contracted business.

OpenAI's rapidly increasing revenue is encouraging. Recent reporting indicates annual recurring revenue is approaching $70 billion.

But Oracle is effectively making enormous long-duration infrastructure commitments based partly on the future financial strength and compute requirements of an AI company whose economics are still evolving rapidly.

The OpenAI relationship therefore increases Oracle's growth visibility while simultaneously increasing counterparty and concentration risk.

This is the most important section of the Oracle thesis.

Oracle spent $55.7 billion on capex during FY2026, up dramatically from prior periods. Management expects FY2027 gross capital expenditures could reach approximately $95 billion. Reuters reported that roughly $70 billion would be directly funded by Oracle, while another $20–25 billion could be funded by customers.

Q1 already demonstrates the scale.

Q1 capex increased 235% year over year, far faster than consolidated revenue.

That relationship is currently my biggest concern.

Oracle's balance sheet is expanding almost as rapidly as OCI.

The most striking number is PP&E.

Oracle added almost $28 billion of net PP&E in just three months.

Construction in progress alone reached $48.5 billion, while computer, networking and machinery assets reached $77.8 billion.

This is physical evidence of how dramatically Oracle's business model is changing.

Oracle had approximately $125.3 billion of current and long-term borrowings at quarter-end against $37.1 billion of cash and securities.

That implies roughly:

Net debt ≈ $88 billion.

Oracle's debt load is substantially different from the balance sheets of the largest hyperscalers against which it competes.

Reuters reported in August that Oracle was approaching the boundary of junk-grade credit as investors questioned the debt-funded AI expansion.

Oracle remains investment grade, but credit risk matters because the company intends to continue investing heavily while borrowing costs remain elevated.

Higher Treasury yields only increase the importance of this issue.

Interest expense increased from $923 million to $1.43 billion, a 55% increase.

That represents approximately 21% of quarterly operating income.

Oracle can currently absorb this expense because operating income is expanding rapidly.

But debt servicing becomes much more problematic if:

Oracle's financing strategy therefore reduces its margin for execution error.

Oracle didn't fund its expansion solely with debt.

During Q1, the company issued 141 million new common shares through its ATM program, raising approximately $19.9 billion.

Outstanding shares increased:

A roughly 5% increase in outstanding shares in one quarter is significant.

Even if Oracle's aggregate earnings increase substantially, existing shareholders own a smaller percentage of the company.

This is one of the most important differences between Oracle and hyperscalers funding AI expansion primarily from enormous internal cash generation.

The cash-flow statement provides the strongest evidence supporting a cautious stance.

At first glance, operating cash flow increasing 184% looks extraordinary.

It is.

But the composition matters enormously.

Oracle received $11.4 billion of customer prepayments containing a significant financing component during Q1. Those payments contributed directly to the $23.1 billion operating-cash-flow figure.

Without simply adjusting them dollar-for-dollar as though they have no economic value, it is still useful to recognize that approximately half of quarterly OCF came from this unusual working-capital source.

Reported OCF: $23.1B

Customer financing prepayments: $11.4B

Capex: $28.5B

Reported FCF: -$5.4B

The prepayments are beneficial because customers are helping Oracle finance construction.

But they also mean investors should not annualize $23.1 billion of Q1 operating cash flow into a $90+ billion normalized annual figure.

Oracle's working capital improved to $8.1 billion from $4.8 billion at fiscal year-end.

But the source of that improvement matters.

Oracle explicitly stated that working capital benefited from:

Those benefits were partially offset by:

This is not the same quality of balance-sheet improvement that would result purely from internally generated FCF.

The bearish valuation case should not obscure Oracle's exceptional competitive assets.

Oracle Database remains embedded throughout banking, telecommunications, government and large enterprises. Replacing these systems is expensive, technically difficult and operationally risky.

Oracle has also made its database increasingly available through AWS, Azure and Google Cloud rather than forcing customers exclusively into OCI.

That is strategically intelligent.

Oracle can monetize its database moat even when another hyperscaler controls the underlying infrastructure relationship.

This remains one of the strongest arguments against an aggressively bearish long-term view of the company.

Cloud applications grew approximately 10% to $4.2 billion during Q1.

Fusion, NetSuite and Oracle Health are considerably less exciting than OCI's 121% growth rate, but their economics are attractive because they provide recurring revenue without requiring Oracle to construct tens of billions of dollars of GPU infrastructure every quarter.

In many ways, this is the part of Oracle I prefer economically.

Applications produce slower growth but require substantially less incremental capital.

The irony of Oracle's current transformation is that the market is focused primarily on the fastest-growing portion of the company even though it may ultimately generate lower returns on incremental capital than Oracle's traditional software franchises.

Oracle is attempting to compete for AI infrastructure workloads against companies with some of the strongest balance sheets in corporate history.

Its primary competitors include Amazon AWS, Microsoft Azure and Google Cloud.

Oracle has several legitimate advantages:

But AWS, Microsoft and Alphabet can fund enormous AI investments largely from internal cash generation.

Oracle increasingly depends on debt, customer financing and equity issuance.

That is a significant strategic disadvantage during a prolonged infrastructure arms race.

Oracle raised $43 billion of debt and $5 billion of equity during FY2026 and previously said it expected approximately another $40 billion of combined debt and equity financing during FY2027.

The $20 billion ATM program has now been completed.

Management is effectively making one of the largest capital-allocation bets in technology history.

If AI infrastructure produces exceptional long-term returns, this aggressive financing could prove prescient.

If infrastructure returns resemble lower-margin commodity cloud economics, shareholders may discover that enormous revenue growth did not translate into comparable per-share value creation.

Oracle expects Q2 revenue to grow 30–34% and cloud revenue to grow approximately 65–71% in USD. Non-GAAP EPS is expected between $1.85 and $1.93.

These are excellent growth rates.

That is precisely why the thesis should not be framed as "Oracle's business is deteriorating."

It isn't.

The bearish thesis is that exceptional growth may still produce disappointing shareholder returns if the capital required to generate that growth becomes too expensive.

Revenue growth and value creation are not synonymous.

ORCL closed September 29 at $137.79. It had already fallen substantially from its 52-week high of $322.54.

Despite that collapse, valuation remains more complicated than the forward earnings multiple suggests.

A 17–18× forward P/E does not look expensive for a company growing revenue 30%.

But Oracle is not currently converting those earnings into free cash flow.

That makes earnings-based valuation potentially misleading.

The bull case assumes today's capex eventually produces enough revenue that infrastructure spending normalizes as a percentage of sales.

That may happen.

But until Oracle demonstrates it, I would apply a lower multiple than I would to a comparably growing asset-light software company.

A useful EPS sensitivity looks like this:

My base valuation assumes Oracle can produce approximately $8–9 of sustainable forward EPS, but that the combination of negative FCF, leverage, dilution and infrastructure risk justifies only roughly a 14–15× multiple until cash economics improve.

That produces fair value around $115–130.

A $120 midpoint represents my 12-month base target.

There are several developments that could cause me to become more constructive:

Recent developments are instead highlighting execution risk.

Oracle issued a force-majeure notice connected to the massive Project Jupiter data-center development in New Mexico after delays in securing power. Reuters reported that the situation has affected financing discussions across the broader AI infrastructure market.

Today, a separate report raised concerns about potential power-connection delays at Oracle's Wisconsin Project Lighthouse campus, although Oracle had not responded to that report at the time it was published.

These developments do not prove Oracle's AI strategy will fail.

They demonstrate that building hundreds of billions of dollars of physical infrastructure introduces risks that did not exist to the same degree in Oracle's traditional software model.

The bear case assumes Oracle continues spending at an extraordinary rate while infrastructure delays and financing costs increase. Revenue can still grow strongly under this scenario, but continued negative FCF and additional balance-sheet pressure force investors to value Oracle more like a leveraged infrastructure provider than a premium software company.

The base case assumes OCI remains a major success but the financial benefits take longer to materialize than revenue growth suggests. Oracle produces approximately $8–9 of normalized earnings while investors apply a roughly 14–15× multiple because FCF remains weak and leverage elevated.

The bull case requires the $664 billion backlog to convert into high-margin revenue quickly enough that capex growth moderates, free cash flow turns strongly positive and Oracle no longer needs external capital to fund expansion. Under that scenario, the current valuation could prove inexpensive.

Oracle's business performance is considerably better than a SELL rating would normally imply. Revenue is growing 30%, cloud revenue is growing 62%, OCI is growing 121%, operating income is growing 57%, and RPO has reached $664 billion.

The SELL thesis instead comes directly from the three financial statements.

The income statement shows exceptional growth but also a 55% increase in interest expense to $1.43 billion as Oracle finances its infrastructure expansion.

The balance sheet shows approximately $125 billion of borrowings and nearly $88 billion of net debt, while PP&E has increased to $127.8 billion. Construction in progress alone stands at $48.5 billion. Oracle is rapidly replacing an asset-light software model with an infrastructure-heavy balance sheet.

The cash-flow statement is the largest concern. Oracle reported $23.1 billion of Q1 operating cash flow, but $11.4 billion came from customer prepayments with a significant financing component. Meanwhile, capex reached $28.5 billion, producing negative $5.4 billion of free cash flow.

And Oracle has already turned to shareholders for capital. The company issued 141 million shares for $19.9 billion, increasing outstanding common shares from 2.88 billion to 3.024 billion in a single quarter.

The $664 billion RPO therefore cannot be viewed in isolation as $664 billion of future value. Oracle must spend enormous amounts of capital to fulfill those contracts, and approximately $300 billion of RPO has been linked to OpenAI alone.

Recent project difficulties add another layer of risk. Oracle's force-majeure notice connected to the Jupiter project has reverberated through AI infrastructure financing markets, while power availability is becoming a bottleneck across multiple proposed data-center developments. The current environment of elevated Treasury yields also makes Oracle's financing-heavy model less forgiving than it would have been during the zero-rate era.

At roughly 17.8× forward earnings, Oracle does not look expensive using P/E alone. But that multiple ignores the defining issue: earnings are currently not converting into free cash flow after the capital required to produce OCI's growth.

My $120 target assumes Oracle's operating business remains strong and OCI continues expanding rapidly, but the market assigns only approximately 14–15× normalized earnings until management demonstrates three things: capex intensity is falling, free cash flow is recovering, and AI infrastructure can increasingly be financed internally rather than through debt, customer financing and shareholder dilution.

The upside risk to this thesis is substantial. If Oracle converts RPO rapidly, achieves attractive infrastructure margins and moves back toward strongly positive FCF without additional material equity issuance, $120 would prove too conservative and the shares could re-rate toward $180 or higher.

Until that cash-flow inflection becomes visible, I believe the financial risk outweighs the headline growth.

Newsletter Verdict: SELL — $120 12-month price target.

The rating and price target represent an equity-research framework based on the assumptions above and are not individualized investment advice.