Microsoft Reassesses AI Infrastructure as Oracle Cloud Deal Winds Down

Microsoft is winding down an agreement to rent artificial intelligence computing capacity from Oracle, highlighting changes in the software company’s approach to securing the infrastructure needed to support its expanding AI operations.

According to a Bloomberg News report cited by Reuters on 6 November, Microsoft has been switching off rented computing capacity from Oracle in recent months after deciding not to renew the contract. The decision was reportedly driven by changes in Microsoft’s computing requirements, with the hardware supplied through the arrangement using an older generation of Nvidia graphics processing units (GPUs).

The development forms part of a broader reassessment of Microsoft’s cloud infrastructure partnerships, as the company evaluates the cost, performance and suitability of external computing resources for its AI workloads.

Why Microsoft is winding down its Oracle AI agreement

Microsoft and Oracle announced a multi-year cloud infrastructure agreement in 2023. Under the arrangement, Microsoft was to use Oracle Cloud Infrastructure (OCI) AI computing resources to support the AI-powered version of its Bing search engine.

However, the computing capacity reportedly became less suitable as Microsoft’s infrastructure requirements evolved. The older-generation Nvidia chips used in Oracle’s hardware were identified in the Bloomberg report as a factor behind the decision not to renew the contract.

The development illustrates how quickly hardware requirements can change in the AI computing market. Infrastructure that meets a customer’s requirements when a contract is signed may become less suitable as newer AI models and workloads demand different computing configurations.

Oracle, meanwhile, maintained that demand for its AI infrastructure remains strong, including for earlier-generation GPUs.

Microsoft did not immediately respond to Reuters’ request for comment.

Microsoft reviews its wider AI cloud infrastructure partnerships

The Oracle development comes amid reports that Microsoft is increasing its oversight of projects involving specialist AI cloud infrastructure providers, commonly referred to as neocloud companies.

Bloomberg separately reported that Microsoft was tightening supervision of certain data centre projects involving IREN, Nebius and Nscale, amid project delays and increasing costs.

These developments suggest that Microsoft is paying closer attention to the execution and technical requirements of external infrastructure agreements rather than relying solely on securing additional computing capacity.

Microsoft’s infrastructure strategy involves several sources of capacity, including company-owned data centres, leased facilities and partnerships with external providers. The company has previously described this approach as providing flexibility in bringing computing resources online.

The reassessment is particularly relevant to neocloud providers, whose business models depend on securing large infrastructure contracts with major technology companies.

IREN delays Microsoft AI infrastructure deliveries

A related development emerged from Australian AI infrastructure provider IREN, which reported delays to some infrastructure deliveries for Microsoft.

On 5 November, IREN said changes requested by Microsoft to the network design had affected the timeline for bringing additional computing capacity online.

The delay pushed back IREN’s target of achieving more than $4 billion in annualised revenue capacity from the end of 2026 to January 2027. The company indicated that the timeline could extend to late February if equipment deliveries were delayed further.

IREN said the changes requested by Microsoft would not result in financial penalties.

The development highlights the technical and operational challenges involved in building large-scale AI data centre infrastructure. Hardware availability, network configuration, commissioning and integration all influence when contracted capacity can become operational.

For infrastructure providers, these requirements can affect project schedules, capital deployment and the timing of revenue generation.

What the Microsoft-Oracle development means for AI cloud providers

Microsoft’s decision to wind down the Oracle arrangement does not establish that the company is reducing its overall AI computing requirements. Rather, the reported circumstances point to a reassessment of the infrastructure used to meet those requirements.

For cloud providers and data centre operators, the development highlights the importance of aligning hardware generations, network architecture and computing capacity with the changing needs of large enterprise customers.

The growing use of advanced AI models is also increasing the importance of infrastructure flexibility. Providers must be able to accommodate changing GPU requirements and network configurations while managing the substantial investment needed to build and operate AI data centres.

Oracle’s response indicates that demand for earlier-generation GPUs remains present in the market. However, Microsoft’s reported decision demonstrates that demand for computing capacity does not necessarily translate into demand for every generation of infrastructure.

What to watch in Microsoft’s AI infrastructure strategy

The next developments to monitor include Microsoft’s approach to renewing external cloud capacity agreements, its investment in company-owned data centres and the progress of projects being delivered by specialist infrastructure providers.

The execution of existing agreements with companies such as IREN, Nebius and Nscale will also be important in assessing how Microsoft balances external partnerships with its own infrastructure expansion.

For the wider AI infrastructure market, the key issue is no longer simply the availability of computing capacity. Hardware suitability, network readiness, delivery schedules and the ability to meet changing customer specifications are becoming increasingly important factors in securing long-term business from major technology companies.

Follow B2B Market Signal on social media for the latest business news, industry developments, market intelligence and analysis. Connect with us on LinkedIn and follow us on Instagram at @b2bmarketsignal for regular updates