Oracle is reportedly evaluating sweeping cost-cutting measures that could include tens of thousands of layoffs and the potential sale of key assets as the company grapples with the enormous financial demands of building infrastructure for artificial intelligence.

According to a research report by investment bank TD Cowen, Oracle could eliminate between 20,000 and 30,000 positions as it attempts to strengthen its finances while continuing to expand its cloud and AI operations. If implemented, the workforce reduction would rank among the largest in the company’s history.

Analysts estimate that cutting that many jobs could help Oracle free up between $8 billion and $10 billion in additional cash flow. The savings would provide some financial relief as the company pours billions of dollars into building massive data centers capable of supporting the computing needs of advanced AI systems.

At the same time, Oracle is also examining the possibility of selling its healthcare technology division, Cernerwhich it acquired in 2022 for $28.3 billion. Divesting the unit could generate significant capital and help the company focus more heavily on its core cloud and AI infrastructure strategy.

Expanding AI Infrastructure Requires Enormous Investment

Oracle’s financial pressures are largely tied to its aggressive push to become a major player in the rapidly growing AI infrastructure market.

To support AI workloads for enterprises and technology partners, companies must build enormous data centers equipped with specialized hardware, high-performance networking systems, and advanced cooling infrastructure. These facilities require substantial upfront investment before they begin generating revenue.

TD Cowen estimates that Oracle’s planned data center expansion could require around $156 billion in total capital spending. The sheer size of this investment has raised concerns among investors and lenders about how the company plans to finance such a massive buildout.

The report notes that both equity investors and debt markets have started questioning whether Oracle can sustain this level of spending while maintaining healthy financial performance.

U.S. Banks Pull Back From Data Center Financing

Financing has become more difficult in recent months as several major U.S. banks have reportedly grown cautious about lending money for Oracle- data center projects.

This pullback has increased pressure on the company’s financing strategy and contributed to rising borrowing costs.

According to TD Cowen, lenders have nearly doubled the interest rate premiums they charge Oracle for project financing to data center construction since September. As a result, Oracle is now facing borrowing costs similar to those typically associated with companies considered below investment grade.

Higher financing costs have already slowed some of the company’s infrastructure plans. Several proposed data center leases that Oracle was negotiating with private operators have reportedly stalled because developers struggled to secure funding for the facilities.

Without access to financing, those developers cannot build the large-scale data centers Oracle needs. This situation creates a bottleneck that could limit the company’s ability to expand its computing capacity.

Billions Already Raised Through Debt Markets

Despite the growing financing challenges, Oracle has already secured significant funding to support some of its projects.

In a span of roughly two months, the company raised about $58 billion in financing tied to new data center developments. That funding includes:

• $38 billion allocated for facilities in Texas and Wisconsin


• $20 billion for a major development project in New Mexico

Although these sums are substantial, analysts say they represent only a portion of the funding Oracle will ultimately require to complete its planned infrastructure rollout.

This means the company may still need to identify additional financing sources or adjust its expansion strategy.

Asian Banks Step In as Alternative Lenders

While some American banks are becoming more cautious, financial institutions from Asia have shown greater willingness to provide loans connected to AI infrastructure projects.

These lenders are reportedly prepared to finance data center development, although often at higher interest rates than what companies previously paid.

For Oracle, this could create an alternative pathway for funding certain projects, particularly those outside the United States. However, analysts say this approach does not fully address the company’s immediate needs for data center capacity within the U.S. marketwhere many of its customers operate.

TD Cowen warned that continued financing difficulties could ultimately affect Oracle’s long-term growth if the company cannot build the facilities required to support its clients.

New Strategies Shift Some Costs to Customers

To ease financial pressure, Oracle is exploring ways to reduce the amount of capital it must spend directly on infrastructure.

One of the strategies involves requiring substantial upfront payments from customers. According to TD Cowen, the company has begun asking new clients to provide deposits of around 40 percent when reserving data center capacity.

These deposits effectively allow customers to help fund the construction of the infrastructure they will eventually use.

Another option being explored is a “bring your own chip” (BYOC) model. Under this arrangement, clients would supply their own specialized AI hardware rather than relying entirely on Oracle’s equipment.

Such a structure would shift a portion of the hardware costs away from Oracle while still allowing customers to run AI workloads on the company’s cloud platform.

Potential Layoffs Would Follow Earlier Job Cuts

If Oracle proceeds with the workforce reduction under consideration, it would follow earlier layoffs carried out in recent years.

In late 2025, the company cut approximately 10,000 jobs as part of a restructuring effort estimated to cost around $1.6 billion.

The company has also repeatedly reduced staff within its Cerner division since acquiring the healthcare software firm.

Some of those layoffs occurred after challenges emerged with a contract involving the U.S. Department of Veterans Affairswhich had selected Cerner’s technology for a large electronic health record modernization program.

Selling the Cerner business could therefore serve multiple purposes—generating capital while allowing Oracle to streamline its operations and focus on its cloud and AI ambitions.

Changes in Customer Partnerships

The company’s infrastructure constraints are also beginning to influence its relationships with major technology customers.

According to the TD Cowen report, OpenAI has shifted some of its near-term computing demand to other cloud providers, including Microsoft and Amazon.

This shift represents a notable change from earlier arrangements in which Oracle had leased approximately 5.2 gigawatts of U.S. data center capacity across states such as Texas, Wisconsin, Michigan, and New Mexico to support OpenAI workloads.

The development highlights how delays in building new infrastructure can quickly affect where AI companies choose to deploy their computing resources.