Intelligence Foundry now supports Z.ai's GLM-5.3 and GLM-5.3-Flash models, both with a 1M-token context window. See Available models.
]]>You can now trigger, track, and download bug reports programmatically, in addition to generating them from the Console. Use the new crusoe diagnostics vm commands to collect diagnostics across many instances or from a script. See Diagnostics and the crusoe diagnostics CLI reference.
When you sign up with an email address, you'll now receive a six-digit verification code. Enter the code in the console to verify your email address and finish creating your account. See Create an account.
]]>Serverless Inference now enforces per-project TPM and RPM limits per model through Quotas. See Serverless rate limits.
]]>New and existing users will receive credits they can use in the Intelligence Foundry to try out Serverless Inference and Serverless Fine-Tuning—without adding a payment method. Credits burn down with usage and update in near real-time.
]]>Load balancer metrics are now available in the Console. Navigate to Networking > Load Balancers, select a load balancer, then select the Metrics tab. See Load Balancer Metrics.
]]>Intelligence Foundry now supports NVIDIA's Nemotron 3.5 Lightning 30B A3B model. See available models for Serverless Inference, Serverless Fine-Tuning, and Self-Serve Deployments.
]]>You can now modify your notification preferences on the Notification Settings page to focus on the alerts that matter most to you. See Notifications.
]]>You can now use Customer-managed encryption keys to encrypt and decrypt your data with a KMS key you own in your AWS account. See Customer-Managed Keys.
]]>You can now generate bug reports for AMD GPU instances directly from the Console, for both standalone VMs and Crusoe Managed Kubernetes nodes. See VM Telemetry and CMK Telemetry.
]]>VM disk I/O and NVMe drive health metrics are now available via API, Telemetry Conduit, and Crusoe MCP. See VM Telemetry.
]]>Object Storage connection latency, throughput, and retransmit metrics observed from the VM are now available via API, Telemetry Conduit, and Crusoe MCP. See VM Telemetry.
]]>You can now spin up dedicated, self-serve inference deployments on Crusoe's managed infrastructure. To get started, see Self-Serve Deployments.
]]>Use Serverless Fine-Tuning to fine-tune an open model using a LoRA-based supervised training workflow through the Intelligence Foundry. To get started, see Serverless Fine-Tuning.
]]>New users can now sign up with Google or Github, and existing users can link their accounts to their Google/Github to sign-in more easily in the future. Note: This doesn't apply to organization-wide enforcement of single sign-on, which is currently only supported for Okta.
]]>AMD MI355x VMs are now available on Crusoe Cloud. Full instance specs are available on the VM Overview page. Please contact sales for availability and interest.
]]>Intelligence Foundry now supports Kimi K2.6. See Available models.
]]>Intelligence Foundry now supports GLM 5.2. See Available models.
]]>Intelligence Foundry now supports Nemotron 3 Ultra 550B. See Available models.
]]>New AMD (Turin) CPU instance types c2a and s2a are now available in eu-norway1-a. Full instance specs are available on the VM Overview page.
b200-180gb-sxm-ib-8x instances are now also available in eu-norway1-a and us-west1-a, in addition to eu-iceland1-a.
Intelligence Foundry now supports GLM 5.1. See Available models.
]]>Intelligence Foundry now supports DeepSeek V4 Pro, DeepSeek V4 Flash, and NVIDIA Nemotron 3 Nano Omni. See Available models.
]]>Object Storage is now generally available on Crusoe Cloud. See Object Storage.
]]>B300 VMs are now available for customer POCs. Full instance specs are available on the VM Overview page. Please contact sales for availability and interest.
]]>A100's are no longer available in us-west1a.
]]>Intelligence Foundry now supports Google's Gemma 4 31B Instruct model, the latest update to the Gemma family focused on maximizing intelligence-per-parameter. See Available models.
]]>Load Balancer Metrics are now available in Limited Availability. Monitor throughput, traffic patterns, and connection behavior for your load balancers via PromQL API. See Load Balancer Metrics.
]]>Command Center now supports AMD GPU Metrics and managed Slurm Metrics.
]]>Intelligence Foundry now supports NVIDIA's Nemotron 3 VoiceChat model, a full-duplex speech-to-speech model. See Available models.
]]>Command Center is now available for Virtual Machines users. See Command Center.
]]>Object Storage is released in Limited Availability on Crusoe Cloud. See Object Storage.
]]>Intelligence Foundry now supports NVIDIA's Nemotron 3 Super, Nemotron 3 Nano, and Nemotron Cosmos models. See Available models.
]]>You can now scope user access to individual projects, rather than granting organization-wide permissions. Roles are assigned in the Console under Manage Organization > User Access > Team. See Managing Users.
]]>The Crusoe Cloud Model Context Protocol (MCP) Server launched in public preview. This enables AI models and agents such as Claude and Cursor to query and analyze your Crusoe AI infrastructure. See MCP Server.
]]>Crusoe Managed Slurm leverages Slurm's powerful job scheduling capabilities while benefiting from Crusoe Managed Kubernetes (CMK) container orchestration. See Managed Slurm.
]]>Added guidance on applying network optimizations for shared disk migration.
]]>With our new usage-based reservation system, you no longer need to manually associate individual VMs with your RIPAs. Management is now focused on monitoring your aggregate usage against your contracted capacity. See the updated docs at Reservations Overview and Managing Reservations.
]]>Managed metrics for CMK node pool, virtual machines, InfiniBand, and shared disks are now generally available.
]]>The Crusoe Container Registry is now generally available. Read more here.
]]>Updated best practices for the Crusoe Container Registry (CCR) to recommend intra-region image pulls to reduce latency.
]]>Managed inference is now available through the Intelligence Foundry on Crusoe Cloud. Read more here.
]]>Updated compute instances to clarify between core count and vCPU count. Read more here.
]]>Support for NFS Shared disks and migration from virtiofs to NFS.
]]>CMK nodepool metrics and InfiniBand metrics launched in public preview. Please see nodepool metrics details here and InfiniBand metrics details here. Now generally available.
]]>VM metrics launched in public preview. See details here. Now generally available.
]]>Custom images are now generally available. Please see details here.
]]>GB200 VMs are now available. Full instance specs are available on the VM Overview page. Please contact sales for availability and interest.
]]>The Crusoe Container Registry launched in public preview, allowing you to store your container images in a given location for faster access as you scale. Read more here. Now generally available.
]]>Single Sign-On enforces secure OIDC-based sign-in to the Crusoe Cloud Console from your identity provider. Set up here.
]]>Added DeepSeek-V3, DeepSeek-R1, and GPT-OSS 120B models to our inference API. Read more here.
]]>Added documentation for NAT Gateways, which can be enabled per subnet. Read more here.
]]>Crusoe Managed Kubernetes (CMK) now supports Kubernetes v1.32. Read more here.
]]>Added documentation for Shared Disk Metrics. Read more here.
]]>L4 passthrough load balancers launched in public preview. Read more here.
]]>B200 VMs are now available. Full instance specs are available on the VM Overview page. Please contact sales for availability and interest.
]]>Updated NCCL configuration guidance. Read more here.
]]>Removed Ubuntu 20.04 references following Ubuntu 20.04 end-of-life on May 31, 2025. This is in line with the End of Life (EOL) of Ubuntu 20.04. Read more here.
]]>You can now view and export your Crusoe Cloud on-demand / spot spending via a new dashboard. See more at Viewing Billing.
]]>Crusoe Cloud will be ending support for curated images based on Ubuntu 20.04 effective May 31st, 2025. This is in line with the End of Life (EOL) of Ubuntu 20.04. Read more here.
]]>Crusoe Cloud now supports customizing your OS images and using it to create new VMs. Read more here.
]]>Crusoe Managed Kubernetes (CMK) now supports Kubernetes v1.31. Read more here.
]]>Added h200 instance support.
Added visibility for resource quotas at the organization and project levels. See more at Viewing Quotas.
]]>You can now add MFA methods to your account for extra security when signing into the Console. Additionally, organization admins can require MFA for their entire organization. Read more on how to set up MFA here.
]]>Launched an alpha for private inference API, initially with OpenAI-API compatible endpoints.
]]>Audit logs of control plane activity are now available in the UI, under the "Usage" page, for admin users. Read more here.
We support listing IB networks and updating existing IB partitions using Terraform.
]]>Audit logs give you a 90-day history of activity in your organization, spanning actions on cloud resources as well as administrative actions. Read more here.
]]>Updated our firewall rules documentation to restrict traffic on ports 25, 465, or 2525, as well as provide recommendations for outbound communication.
]]>Shared Disks offer a high level of scalability and performance and are attached to many VMs with shared access. We offer shared disks from 1 TiB to 1000 TiB. See more at Storage Overview.
]]>New, larger s1a instance types (s1a.120x, s1.160x) are now available on Crusoe Cloud, with up to 8x12.8TBs of NVMe storage per VM.
]]>Launched an improved reservation system, where customers can track reservations and specify logic for whether VMs are placed on a reservation or on-demand.
]]>Added a curated image, ubuntu22.04-nividia-slurm, for customers spinning up VMs with Slurm.
We've added a new bulk VM creation method that allows you to create multiple VMs atomically. Learn more here.
]]>You can now save the configuration of an instance to an instance template, allowing you to easily create multiple VMs that share the same configuration. Documentation is available here.
]]>We now support non-default VPC networks and subnets, with customer defined IP ranges in RFC 1918 space. This allows you to better segment resources within your network, and create additional independent networks (e.g. management and data plane). Learn more here.
]]>We now support creating firewall rules with direction: egress. This allows you to control what traffic is allowed to leave a VM, subnet, or network. Learn more here.
The 'Instances' page on the console is now paginated with a configurable number of items per page. Additionally, you can now search for specific instances via a search bar and filter by Status, Instance Types and Locations. Additionally, our v1alpha5 API now supports pagination on GET .../instances.
We've launched a new usage dashboard allowing you to view and export usage details for instances and disks provisioned within your organization. Learn more here. New APIs are also available to export usage information as well.
]]>You can now update an existing firewall rule via the UI, API and CLI.
]]>We've released our v1alpha5 API which introduces multiple projects per organization. Projects are a way to organize resources within an organization. Resources are not shared between projects. Each project will have its own VPC network and quotas. New projects can be created and listed with the projects API. Project ID will also be a required argument for resource-related APIs, such as VM management. By default, the CLI and UI will use the default project automatically created for every organization, unless another is specified.
Crusoe Cloud has launched general purpose (c1a) and storage specific (s1a) CPU only instances. Learn more at CPU instances.
You can now enable serial console access to provide out of band access to VMs for debugging purposes. Learn more at serial console access.
]]>Curated Images have been updated to support Ubuntu 22.04 as a base OS. Learn more at Images Overview.
]]>You can now update the IB network and/or partition of a VM currently connected to a different network and/or partition. Learn more at Updating the IB partition on a VM.
]]>Curated Images are now available, which allow developers to create a VM with a different base operating system or different preinstalled software. Learn more at Images Overview.
]]>All VMs within a VPC now have a default DNS address, which is reachable from all other VMs on the same VPC Network. Learn more at Internal DNS.
]]>The default user for all images is now ubuntu instead of root. Please access VMs using ssh ubuntu@<IP>; root access should be considered deprecated. See default user for more details.
Developers can now attach or detach persistent disks without having to stop and restart the VM.
]]>H100 HGX servers are now available. Full instance specs are available at VM Hardware. Please contact sales to enable this feature for your organization.
]]>Resource names (including VM names) are limited to a-z, 0-9, and non-leading - and _ charaters. This is to support future DNS addressing by resource name.
A100 HGX servers are now available. Full instance specs are available at VM Hardware. Please contact sales to enable this feature for your organization.
]]>InfiniBand networks and user-configurable partitions are now available. Visit Managing InfiniBand Networks to learn more.
]]>Organization Admins can now invite other users to join their organization as an Admin, Editor, or Reader.
]]>Organization Admins can change the role of other organization members to the Admin, Editor, or Reader role, which grants different permissions to Crusoe Cloud resources.
]]>We have released a Capacity API that lets users make requests to determine the current on-demand capacity available on the platform. Please contact support to enable this feature for your organization.
]]>We've released our v1alpha5 API which introduces block storage and networking resources. This includes the ability to create and attach block storage volumes to your VMs, and the ability to create and manage firewall rules for your virtual private cloud (VPC). Examples of new APIs include createDisk and createVPCFirewallRule. You may attach disks to stopped VMs via the attachDisks API.
Our UI and CLI now use our v1alpha5 APIs, with the ability to create and manage block storage and networking resources.
We've released our v1alpha2 API which introduces asynchronous operations for creating and updating resources. Asynchronous APIs enable better error handling and a lower risk of timeouts. Some examples include createInstance which returns an Operation and getComputeVMsInstancesOperations which returns all a list of Operation resources associated with your VMs.
Our UI and CLI now use our v1alpha2 APIs, providing better error handling, multiple parallel operations, and more robust operations.
status.crusoecloud.com is now available for developers to check for known issues or scheduled maintenance and subscribe for updates to Crusoe Cloud services. Status notifications are also visible as indicators on the user profile icon withing the console.
]]>VMs are now created with their own IP address, rather than a shared IP. Currently only port 22 (ssh) is open, but in the future we may allow you to allow ingress or egress traffic through other ports. This IP address is not static, and may change if you start and stop your machine.
Users who have forgotten their password can now reset it from the home page of the console by clicking the Forgot password? link.
VMs are now able to automatically run scripts on startup and shutdown. Learn more in Lifecycle Scripts.
]]>The Crusoe Cloud CLI is now available for download via brew and apt package managers. Get started by installing and configuring the CLI.
The NVIDIA container toolkit is now installed by default on all newly created VMs.
]]>Fixed a bug where the NVIDIA drivers could get out of sync with the operating system version, resulting in the GPUs no longer being found. The drivers are now updated appropriately and should always be on the correct version.
]]>Users can store SSH keys to be re-used on subsequent VMs without needing to add a key to each VM on creation.
]]>Users can switch to dark mode in the settings drop down or by pressing cmd + j.
Additional error and progress messages have been added to notify users of the VM creation progress.
]]>Fixed a bug where the UI would timeout when attempting to launch VMs with four or more GPUs. Large VMs can now consistently be created without timing out.
]]>