Oracle AI backlog has climbed to $664 billion, giving investors fresh evidence that its expensive cloud expansion has customers waiting. Shares rose after quarterly results showed accelerating infrastructure revenue alongside stronger-than-expected earnings and a smaller cash drain.
The company added more than $30 billion in new AI cloud contracts during its fiscal first quarter. Those agreements helped push remaining performance obligations higher by $26 billion compared with the previous quarter.
Oracle reported quarterly revenue of $19.3 billion, representing 30% growth compared with the same period last year. Cloud revenue climbed 62% to $11.6 billion as artificial intelligence workloads continued driving infrastructure demand.
Cloud infrastructure revenue grew even faster, jumping 121% to $7.4 billion during the quarter. That performance helped offset continuing concerns about how much Oracle must spend before its AI infrastructure becomes consistently profitable.
The Oracle AI backlog has become especially important because investors have questioned whether current spending can produce sustainable returns. Oracle shares had fallen more than 21% this year before the latest results offered investors some relief.
Roughly half of Oracle’s $664 billion backlog should convert into revenue during the next 36 months. That visibility gives management a clearer picture of future cloud sales while infrastructure continues expanding around committed demand.
Just as importantly, Oracle says much of its newest contracted business will not require additional capital beyond existing plans. Some customers are supplying their own GPUs, while others are providing prepayments that help finance infrastructure construction.
That structure matters because Oracle’s biggest challenge has increasingly shifted from finding customers toward financing enough capacity. AI data centers require enormous upfront spending on chips, power systems, networking, cooling equipment, and construction.
Oracle delivered another 850 megawatts of data-center capacity during the quarter as its infrastructure business continued expanding rapidly. Since the previous quarter ended, the company also delivered more than 300,000 GPUs to cloud customers.
Oracle says that capacity was almost three times what it delivered during the previous fiscal quarter. Such rapid deployment shows how aggressively management is trying to convert contracted artificial intelligence demand into usable computing infrastructure.
However, the Oracle AI backlog does not eliminate the financial risks surrounding that expansion strategy. The company still generated negative free cash flow of $5.4 billion despite performing better than analysts expected.
Analysts had expected Oracle to burn approximately $9.56 billion in free cash flow during the quarter. The smaller loss therefore offered investors some evidence that spending pressure may be becoming more manageable.
Still, Oracle remains committed to spending heavily because AI customers continue requesting more capacity than its cloud network can provide. The company has said demand for training and inference services continues growing faster than available infrastructure.
That shortage creates both an opportunity and a problem because Oracle can sign contracts faster than infrastructure becomes available. Every major agreement eventually requires enough power, chips, buildings, and networking capacity to deliver the promised computing resources.
Oracle expects to raise approximately $40 billion through debt and equity during the current fiscal year. The company already completed a $20 billion stock sale during the first quarter under that broader financing plan.
Those financing requirements explain why Oracle’s debt position remains central to how investors judge its artificial intelligence strategy. Revenue can grow rapidly while returns remain disappointing when infrastructure expansion absorbs enormous amounts of cash simultaneously.
Morningstar analyst Luke Yang expects the pressure to continue because cloud revenue still needs considerably greater scale. He believes Oracle may need years before cloud income consistently supports expansion while also producing positive free cash flow.
That tension has defined much of the debate around Oracle since its AI infrastructure ambitions accelerated. Investors can clearly see exceptional demand, yet they must also decide how much financial pressure remains acceptable while capacity catches up.
The Oracle AI backlog strengthens the bullish argument because contracted revenue continues rising while customers increasingly help fund infrastructure. That arrangement lowers some risk compared with Oracle independently financing every GPU and data-center project before receiving meaningful revenue.
Oracle disclosed during its previous quarter that prepaid and customer-supplied hardware connected with large AI contracts totaled $75 billion. Those agreements substantially reduced how much outside capital Oracle needed for portions of its planned data-center construction.
The latest quarter appears to extend that model as the company takes on additional large artificial intelligence contracts. Oracle says the newest agreements will not force it to increase previously announced capital-raising plans.
This financing model could become particularly important as AI infrastructure costs continue rising across the technology industry. Cloud providers are spending heavily while chip prices, electricity requirements, and construction costs create pressure across project budgets.
Oracle is also competing against considerably larger cloud businesses operated by Amazon, Microsoft, and Google. Those companies possess enormous balance sheets and established platforms capable of funding long-term infrastructure expansion from broader technology businesses.
Yet Oracle has carved out momentum by providing large GPU clusters for companies demanding enormous amounts of computing capacity. Its infrastructure business increasingly serves artificial intelligence companies whose workloads require dedicated systems operating at extraordinary scale.
That shift is visible inside Oracle’s revenue mix because traditional software is no longer driving its fastest expansion. Software revenue declined 3% during the quarter while cloud infrastructure increased by triple digits.
The contrast shows how quickly Oracle is repositioning itself around the infrastructure supporting generative artificial intelligence. Its traditional database and enterprise software businesses remain important, but the strongest growth now comes from cloud computing.
Oracle expects that momentum to continue during its second fiscal quarter, with total revenue increasing between 30% and 34%. Cloud revenue is expected to grow between 65% and 71% in reported dollar terms.
Those forecasts place even greater importance on whether Oracle can keep bringing contracted capacity online without worsening its balance sheet. Demand alone will matter less if infrastructure expenses consume too much cash before customers generate enough recurring revenue.
The Oracle AI backlog provides management with something many infrastructure builders lack: unusually large visibility into future sales. Hundreds of billions in committed business can justify aggressive expansion when customers are contractually waiting for capacity.
However, backlog should not be confused with immediate cash generation because much of that revenue arrives over several years. Oracle must still construct enough infrastructure, manage financing costs, and maintain attractive economics while those contracts gradually convert.
Investors appeared more willing to accept that trade-off after the latest quarter showed stronger growth and smaller-than-expected cash burn. Oracle shares gained roughly 3% in early trading as the new numbers improved confidence around its AI strategy.
For Oracle, the next stage involves proving that its massive backlog can become profitable cloud revenue without continually expanding debt. Customer prepayments and supplied hardware may help, but infrastructure remains enormously expensive at this level of demand.
The Oracle AI backlog has now made one part of the story significantly clearer: customers are arriving in extraordinary numbers. What remains unresolved is whether Oracle can build quickly enough while turning that demand into durable cash generation.
