1. Functional Features of the Service
1.1. Nebius provides Customer with scalable virtual computing capacity to place, store, and process Customer Content, to create and run Virtual Machines, to manage them, and to use other functional features of the Service. Disks and Shared Filesystems that may be attached to Virtual Machines are provided under the respective Terms of Use available at https://docs.nebius.com/legal/specific-terms. 1.2. As part of the Service, the following resources are available: (a) Virtual Machines based on Linux, including Virtual Machines with GPUs; (b) Preemptible VMs; (c) Images and Snapshots. 1.3. Customer may: (a) create Virtual Machines in Regions selected by Customer; (b) create Virtual Machines with the selected technical parameters including platform and preset; (c) access Virtual Machines via SSH using a key pair; (d) monitor the state of Virtual Machines; (e) use other functional features of the Service, which Nebius makes available to Customers. Functional features of the Service are described in Documentation on Site. 1.4. Service Level for the Service, including the Service Level for Rack-Scale GPU System, is determined in SLA. Preemptible VMs are not subject to SLA. 1.5. Rack-Scale GPU System. (a) A Rack-Scale GPU System is allocated to Customer as a single resource on one Rack dedicated to Customer for the duration of the allocation. The Rack includes the Maintenance Buffer, which forms part of Customer’s allocation. (b) Nebius may, in the event of a failure of a Rack-Level Component or of multiple Nodes, offer Customer the option to relocate the Rack-Scale GPU System, including all its Virtual Machines, to a replacement Rack through a coordinated stop and restart of the Virtual Machines. Nebius will notify Customer of such option via the Management Console or other means described in the Documentation. (c) Nebius may perform maintenance of a Rack, including replacement of Nodes and Rack-Level Components, in accordance with SLA. Customer shall cooperate with Nebius in scheduling and performing such maintenance. 1.6. Preemptible VMs are provided on a best-effort basis. Nebius may stop a Preemptible VM at any time in accordance with the Documentation. 1.7. Where a Virtual Machine cannot continue to run on its current physical host, Nebius may make another host available and notify Customer by the means described in the Documentation. Restart of the Virtual Machine on such host is performed by Nebius in accordance with the recovery settings of the Virtual Machine or by Customer.2. Pricing
2.1. The use of the Service is chargeable. 2.2. Service Fee is determined as per Service Rates and billing units specified on Site. Nebius may change Service Rates as stipulated by Agreement. 2.3. Service Fee for a Virtual Machine usage depends on the allocated computing resources, operating system, and usage time. Service Fee is calculated for the time of using the Virtual Machine, from the moment it is started and to a complete stop. The time when the Virtual Machine is stopped is not charged for computing resources. Resources attached to or reserved for a Virtual Machine other than computing resources, such as Disks, Shared Filesystems, Images, Snapshots and reserved public IP addresses, are charged for the time of their existence irrespective of whether the Virtual Machine is running. 2.4. Service Fee for a Rack-Scale GPU System is determined in the applicable Order Form or Service Rates and applies to the whole Rack allocated to Customer, including the Maintenance Buffer, irrespective of the number of running Virtual Machines. 2.5. As part of the Service, the volume of Outgoing Traffic passing through Customer’s Virtual Machine, the cost of Images, Snapshots, user licenses for subscription to remote desktop services, and other additional benefits ordered and used by Customer and related to the use of the Service are charged separately. 2.6. Nebius may from time to time provide Customer with additional functional features of the Service for an additional fee.3. Limitations
3.1. Customer shall monitor and ensure the security and relevance of software used in Virtual Machines, timely update software versions or make configuration changes as per the Documentation or requirements published by developers of respective software. 3.2. Customer shall not install and shall take measures to prevent the installation of: (a) malware on Virtual Machines and Disks; (b) software specifically developed to gain unauthorized access to information; or (c) software, databases, or other Content in violation of the intellectual property rights of third parties. 3.3. If any gross violations of information protection requirements threatening the operation of Platform or third-party resources are revealed on the part of Customer, Nebius may interrupt the operation of Virtual Machines that violate security requirements.4. Definitions
Disk has the meaning set forth in the applicable Terms of Use of the respective service. Host means the physical server of the Platform on which a Virtual Machine runs. Image means a full copy of the structure of file system and data on a Disk used to create Virtual Machines. Instance means a Virtual Machine as represented in the Nebius Compute API. The terms Instance and Virtual Machine are used interchangeably in the Linked Documents and the word Instance in the names of resources of the Nebius Compute API, such as NVL Instance Group, refers to that representation. Maintenance Buffer has the meaning set forth in the Service Level for Compute Cloud. Node means a Host forming single physical compute tray within a Rack. Each Node hosts one Virtual Machine. Preemptible VM means a Virtual Machine that Nebius may stop at any time in accordance with Documentation. Outgoing Traffic means an amount of information outgoing to external information systems and passing through the virtual computing resources or other cloud resources of Platform provided to Customer. Information on consumed outgoing traffic is provided to Customer as Statistics aggregated from all Services. Rack means a single rack-scale physical unit of a product offered as Rack-Scale GPU System, comprising the number of Nodes specified in Documentation and in the Service Level for Compute Cloud, interconnected by a unified NVLink domain and dedicated to one Customer at a time. Rack-Level Component means any element of the shared physical infrastructure within a Rack that is managed by Nebius and supports multiple Nodes simultaneously, including rack power distribution, rack cooling systems, NVLink switching hardware, and rack backplane connectivity. Rack-Scale GPU System means a group of Virtual Machines with GPUs sharing a unified high-speed GPU interconnect, allocated to Customer as a single resource on one Rack and represented in the Nebius Compute API as an NVL Instance Group. Each Rack-Scale GPU System maps to exactly one Rack at a time. Shared Filesystem has the meaning set forth in the applicable Terms of Use of the respective service. Snapshot means a copy of Disk file system at a given moment. Virtual Machine means virtual computing resource of Platform, emulating hardware and used by Customer to remotely process Content.Web address: https://docs.nebius.com/legal/specific-terms/compute Publication date: October 6, 2026 Effective date: October 6, 2026 Previous version of the document: https://docs.nebius.com/legal/archive/specific-terms/compute-20241125 Previous version of the document: https://docs.nebius.com/legal/archive/specific-terms/compute-20240925