Microsoft Oracle AI cloud: According to a Reuters account of a Bloomberg report dated 6 November 2026, Microsoft is winding down an arrangement to rent AI computing capacity from Oracle. The company has been turning off rented capacity in recent months after declining to renew the contract. Public discussion also tracked a post by @RobynHD quoting a short alert from @DeItaone.
Bloomberg cites people familiar with the matter: the rented capacity became less useful because it no longer matched Microsoft’s specifications. The Oracle hardware used an older generation of Nvidia chips. Financial terms of the original deal were not disclosed in 2023 and are still not quantified in the current coverage — so claims of a specific “billion-dollar” Oracle contract size remain unverified.
What Bloomberg and Reuters report
The story places the move inside a broader reassessment of billions of dollars in cloud partnerships. An Oracle spokesperson told Reuters that demand for Oracle Cloud Infrastructure (OCI) AI capacity remains strong, including for earlier-generation GPUs. Microsoft did not immediately respond to Reuters.
Background: On 7 November 2023, Oracle announced a multi-year agreement under which Microsoft uses OCI AI infrastructure alongside Azure AI for inferencing models that power Bing conversational search (Oracle/PR Newswire). No deal value was published then either.
Context: Nscale option and IREN delay
On 5 November 2026, Reuters reported — citing the Financial Times — that Microsoft will not exercise an option to more than double its existing Texas agreement with AI cloud provider Nscale. The FT said a potential extension would have been worth “tens of billions of dollars” for Nscale. A person familiar with the matter told Reuters the decision was made months before Nscale’s September IPO filing and that the existing Texas contract was unaffected.
Also on 5 November, IREN said some Microsoft infrastructure deliveries were delayed after the customer requested network-design changes. Its target for bringing more than $4 billion of annualized run-rate revenue capacity online slipped from year-end to January, possibly late February. IREN said it faces no financial penalties for those customer-driven delays. Bloomberg separately reported that Microsoft is tightening oversight of some neocloud projects involving IREN, Nebius and Nscale.
Reading without hype
What holds: Microsoft is not renewing and is winding down rented Oracle AI capacity; the primary newswire line is Bloomberg via Reuters and Bloomberg Law. What does not hold as a hard figure: the euro or dollar size of this specific Oracle contract. RobynHD’s X thread still framed the Oracle item as not independently confirmed at posting time — that caveat is overtaken by Bloomberg/Reuters, but it does not invent a contract value.
Whether this signals a broad pullback from AI infrastructure spending or mainly a shift (more owned capacity, selective third-party deals) is not settled by the sources above. On neocloud topics, Microsoft has described a mix of owned sites, leased facilities and partnerships — without specifically commenting on the Oracle wind-down in that quote.
What to watch next
Investors and infra teams: Oracle and Microsoft commentary on OCI utilization and capex mix; Nscale backlog versus unexercised options; IREN delivery dates and SEC language on “Microsoft-requested change”. For reporting: prefer Bloomberg/Reuters and company filings over X alerts as evidence.
Related reading on code404zero: Grid Protectors — OpenAI and the power grid, Meta acquires Conductor, and GPT-6 Cyber — related infrastructure and AI-cloud context without overlapping focus keywords.
Bottom line: On 6 November 2026, Bloomberg reports Microsoft is winding down its Oracle AI cloud rental deal because older Nvidia hardware no longer fits. Nscale and IREN supply the week’s context — declined expansion and a design-driven delay — without proving a sector-wide “AI crash”.
