October 2026 Summaries

No summary generated yet.

No summary generated yet.

Nebius’ quarterly updates expand cloud connectivity, data management, cost controls, AI development tools, security, automation, and managed inference capabilities. Cloud Interconnect offers private encrypted connectivity to on-premises and other cloud environments, while dynamic Preemptible VM pricing, storage transfer support for Shared Filesystem, object tagging and inventory, and enhanced cost allocation tools aim to improve efficiency and governance. Serverless AI additions, including Devlab workspaces, direct Python execution, file injection, and templates, reduce workload setup time, alongside Cloud Shell, REST APIs, Wiz integration, and AI-agent-oriented infrastructure tools such as Nebius Echo and public skills repositories. The platform also added CNCF-certified Kubernetes, a public application marketplace, tunnels, and LanceDB guidance. Token Factory introduced custom speculative-model training, a broader open-weight model catalog, a Responses API for agent workflows, expanded inference observability, scalable batch inference, and integrations with additional coding tools. Tavily enhanced search quality, expanded coverage for people, companies, and news, added integrations for agent and application workflows, and received ISO 27001:2022 certification.

Nebius AI Cloud’s Q3 updates focused on enterprise connectivity, cost management, developer workflow simplification, and platform performance. New infrastructure capabilities include Cloud Interconnect for private high-bandwidth links to on-premises, colocation, or other cloud environments, expanded Data Transfer Service support for the Nebius Shared Filesystem and S3-compatible storage, Kubernetes conformance certification, and planned security integrations such as SIEM audit-log exports and Wiz support. Cost and capacity features include demand-based Spot pricing for Preemptible Virtual Machines, cost allocation labels, monthly usage reports, project limits for reserved capacity, and new object storage tagging and inventory tools. Serverless AI additions allow direct Python execution, file injection for configuration files, reusable Devlab workspaces, and validated templates, reducing container-building and setup requirements. Other developer tools include browser-based Cloud Shell, a REST API with OpenAPI specifications, guidance for coding agents, enhancements to the Echo infrastructure agent, marketplace browsing before sign-in, and Tunnels for exposing local services during testing. The company also cited MLPerf benchmark results and a Platinum ClusterMAX 3.0 rating from SemiAnalysis as indicators of its training, inference, storage, and cluster performance.