Oracle broadens its accelerator roadmap
Oracle announced Google TPUs on OCI for 2027. Future availability remains subject to rollout.
Read the original source →RESEARCH & INTELLIGENCE
Evidence before assumptions.
A focused reading room for cloud computing decisions. Official announcements are useful evidence of a provider’s direction; they are not independent performance guarantees.
Editorial review: 9 November 2026. This is a curated snapshot, not a live news feed. Publication dates are shown where available. Roadmaps may change. Recheck sources before a commercial decision.
Three questions to take into a provider discussion
What is available now?
Separate an announced architecture from a bookable allocation in your required region and start window.
What is measured?
Ask for evidence using your software, dataset, precision and target scale. Peak specifications and delivered throughput describe different things.
What is included?
Make storage, interconnect, support, egress and contract commitments explicit alongside compute rental.
Official source ledger
AI cloud extends across the workload lifecycle
Nebius’s Q2 shareholder letter describes its integrated approach to training and inference.
Read the original source →GPU clusters need a connected fabric
Vultr documentation distinguishes compute, bare metal and GPU fabric clusters.
Read the original source →Vultr describes its enterprise AI infrastructure
Vultr discusses self-service CPU and GPU clusters and identity features. Some capabilities are described as forthcoming, so verify availability directly.
Read the original source →Oracle announces Zettascale10
Oracle’s announcement connects large AI clusters with networking, reliability and facility design. Announced maxima are not workload benchmarks or evidence of capacity available to every customer.
Read the original source →OCI explains its HPC platform
Oracle’s HPC page brings together bare metal, networking and storage for scientific and engineering applications. Provider claims require application-specific validation.
Read the original source →Nebius describes its cloud approach
Nebius presents a platform spanning AI development and deployment. Its descriptions are vendor information, rather than independent comparative testing.
Read the original source →Vultr adds CPU cluster support
Vultr’s product update describes extending its cluster offering to CPU resources. This illustrates why the CPU layer belongs in AI infrastructure planning.
Read the original source →