Oracle’s Credit Stress Is Becoming an AI Financing Test

Disney is reportedly preparing to combine parts of ABC Entertainment, 20th Television, Hulu Originals and Freeform as it organizes the television business around streaming customers and franchises. The plan could affect hundreds of jobs, while separate reductions have already touched human resources and technology.

The financial logic is clear. Legacy television is shrinking, and duplicated teams can make development and distribution more expensive. A simpler organization can lower costs and give Disney one view of a program’s value across streaming, broadcast, advertising and consumer products.

Disney is also raising the price of ad-free Disney+ and Hulu plans by $2.50 to $21.49 a month, according to TIKR. Higher prices can lift average revenue per subscriber immediately, but only if cancellations remain controlled.

A stronger starting point

The restructuring follows improved streaming economics. The source reported fiscal third-quarter adjusted earnings per share of $2.06 versus a cited estimate of $1.86. Streaming operating income more than doubled to $712 million, and the direct-to-consumer business produced a 13% margin.

Disney still trails Netflix’s profitability. Netflix reported a 33.4% company operating margin in its second quarter and kept a 31.5% full-year outlook. The comparison is imperfect because Disney also owns parks, television networks and consumer businesses, but it shows the remaining margin gap in streaming.

Disney’s diversified model can offset that gap. Experiences delivered record fiscal third-quarter revenue and operating income, giving the company cash generation that a pure streamer does not have. It also lets Disney extend characters into parks, cruises and merchandise.

The next report must show that the reorganization improves economics without slowing the content pipeline. Investors should watch streaming cancellations after the price increase, direct-to-consumer margins, restructuring charges and the pace of linear-network decline. Cost cuts are valuable only if they preserve the franchises that make Disney’s ecosystem work.

Netflix has told investors that content expense should grow more slowly than revenue in 2026. Its July shareholder letter forecast revenue of $51.0 billion to $51.4 billion, growth of 13% to 14%, while content amortization was expected to increase about 10%.

That spread is the foundation of the margin case. Revenue can rise faster than the cost recognized for films and series, allowing operating profit to grow faster than sales. Netflix kept its 2026 operating-margin outlook at 31.5%, up from 29.5% in 2025.

The second quarter supported that direction. Revenue rose 13% to $12.6 billion, operating income reached $4.2 billion and the quarter’s operating margin was 33.4%. Management projected third-quarter revenue growth of about 12%, slower than the second quarter but still double digit.

Not every content dollar has the same job

Netflix said live programming would represent just over 5% of 2026 content spending and about 1% of viewing hours. On a simple cost-per-hour measure that looks inefficient, but live events accounted for six of the company’s ten strongest new-member sign-up days over the prior five years. The benefit is acquisition and retention, not only hours watched.

That distinction matters as Netflix changes what it discloses about engagement. Less frequent viewing data can make it harder for investors to separate healthy monetization from slower use. Pricing and advertising can lift revenue even when viewing growth is modest, but durable pricing power depends on customers continuing to value the service.

The next earnings report should be judged on revenue growth, operating margin, free cash flow and management’s explanation of content returns. A slower content budget can protect margins, but it cannot by itself restart top-line growth.

Netflix’s lower valuation after the share decline improves the setup only if the company continues to convert pricing, advertising and engagement into cash. Spending discipline is a strength; it becomes a problem if it leaves the service with fewer reasons for members to stay.

Oracle said construction at the Project Jupiter data-center campus in New Mexico was moving quickly, a milestone that pushed Bloom Energy shares about 3% higher in the source report. The campus is being built by Stack Infrastructure, operated by Oracle and designed to use Bloom Energy fuel cells as a primary power source.

A proposed energy-services arrangement with El Paso Electric would support late-stage construction, commissioning and emergency backup. The plan cited in the report calls for as much as 2.45 gigawatts of installed Bloom fuel-cell capacity, replacing an earlier design that relied on gas turbines and diesel generators.

Why progress on site matters

Bloom’s solid-oxide fuel cells generate electricity through an electrochemical process rather than conventional combustion. For a data center, the commercial appeal is speed and modular deployment where grid connections may take years. Project Jupiter can therefore test whether that engineering advantage converts into large, timely revenue.

The size also raises execution questions. A planned gigawatt figure is not the same as commissioned capacity, and a proposed utility agreement is not an unconditional obligation. Investors need evidence on deployment schedules, acceptance milestones, customer payments and the cost of supporting rapid manufacturing growth.

The broader financing environment adds risk. Data-center operators and infrastructure providers are committing large sums while borrowing costs remain elevated. If customers slow construction or financing becomes harder, suppliers can see orders delayed even when long-term electricity demand remains strong.

Bloom is directly exposed to both sides of the trade. Faster construction can pull revenue forward and validate its role in AI infrastructure. Delays, cost overruns or a change in the power design could push cash realization out. The next useful evidence is not another capacity announcement, but commissioned megawatts, recognized revenue and cash conversion tied to Project Jupiter.

Vertiv shares fell about 13.9% in one week, far more than the S&P 500’s 1.5% decline, even though the company did not announce a comparable operating setback. The selling followed a sharp rise in borrowing costs at CoreWeave and renewed questions about how debt-heavy AI infrastructure will be financed.

That connection matters because Vertiv supplies power and cooling equipment to data centers. Its customers’ ability to fund new capacity affects orders, project timing and backlog conversion, even when end demand for computing remains strong.

After the decline, TIKR placed Vertiv near 26 times forward earnings, roughly in line with its five-year average of 26.6 times and far below the 55.1 times reached in May. Morgan Stanley’s reported $330 price target implied about 60% upside from the cited closing price, but a target is an analyst estimate, not evidence that earnings will meet the assumptions behind it.

The operating benchmark

Vertiv’s investor-relations site confirmed that third-quarter 2026 results were scheduled for October 28. The source article cited adjusted operating-profit guidance of $898 million to $938 million. The most informative figures will be organic orders, book-to-bill, backlog and cash conversion.

Vertiv’s 2025 third quarter showed why those metrics matter: the company reported organic orders up roughly 60% year over year, a book-to-bill ratio near 1.4 and backlog of $9.5 billion. That historical strength does not guarantee the same pace in 2026, especially if customers delay financing.

A return to an average valuation multiple improves the entry price, but it does not remove the funding risk that triggered the selloff. Investors should treat analyst targets as scenarios and watch whether orders continue to grow faster than sales. If backlog remains firm and margins meet guidance, the decline may have overshot the operating change. If project funding slows, the stock can become cheaper for a fundamental reason.

Amazon shares rose more than 5% after Morgan Stanley named the company a top pick and raised its price target to $350 from $335. The more important question is whether the bank's AWS forecasts can become cash flow rather than just revenue growth.

Morgan Stanley modeled AWS growth of 47% in 2027 and 42% in 2028, with operating margins of 42% to 44%, according to Proactive. Those are analyst estimates, not company guidance. The bank linked the case to GPU deployments, Bedrock, agentic shopping and robotics.

Amazon's completed second-quarter results give the thesis a strong starting point. AWS sales rose 37% to $42.2 billion, and AWS operating income increased to $16.6 billion from $10.2 billion a year earlier. Companywide sales reached $200.6 billion, up 20%.

The cash-flow side is less comfortable. Amazon reported a trailing-12-month free cash outflow of $7.6 billion, down from an $18.2 billion inflow a year earlier. The company attributed the change primarily to a $66.1 billion year-over-year increase in purchases of property and equipment, net of proceeds and incentives, largely for artificial intelligence.

That spending can create valuable capacity if customer demand remains strong and utilization rises. It can also depress returns if infrastructure arrives ahead of monetization, power or customer workloads. A high-margin AWS forecast therefore depends on both pricing and disciplined deployment.

The analyst target uses assumptions about future growth, margins and valuation and should not be treated as a realized return. Investors can test the thesis through three reported measures: AWS revenue growth, AWS operating margin and free cash flow after capital expenditure.

AWS is accelerating, but the cost of building the next stage of that growth is already visible.