Release Notes: 2025.11.4
This release includes significant enhancements to spark provisioning, security updates across the platform, and general UI improvements.
This release includes significant enhancements to spark provisioning, security updates across the platform, and general UI improvements.
This update brings a fully redesigned website, Veritate AI, free threat intel feeds, Spark snapshots/restores, and billing.
This update brings a free public AI chat, more capable AI assistants, a much richer object storage experience, and clearer itemized billing.
New usage-based billing means you pay only for what you run, plus Object Storage, per-network bandwidth plans, threat intel feeds, and a refreshed dashboard.
Version 2025.5 is here. Region-aware dashboard, AI chat with history and markdown, browser RDP/VNC consoles, real in-VM disk usage, tighter security.
Security hardening for 2FA and API keys, SSH gateway lockout on suspicious activity, deployment firewall visibility, and notification history.
Carpathian Cloud Version 2026.4.0 is available to all users. Contains many bug fixes, better security, and more granular networking and firewall control.
This update includes significant Spark server provisioning enhancements, general security updates, and a redesigned UI for better navigation and usability. We've brought provisioning times down from 3-5 minutes to about 30 seconds, launched our USC2 data center with GPU infrastructure, and added granular permission controls. We also cover the Nov 25 outage and how we've improved failover to minimize downtime. Plus, get a preview of upcoming features like team management, WordPress hosting, object storage, and managed databases.
We renamed the "Web Servers" page to "Networking" and improved the copy. This page is the primary area to configure web traffic and internet routing to make your VMs accessible from the internet.
Authors have been added to the publications page. This is in preparation for guest posting articles and forum creation (though users can still post and discuss the platform on our General Subreddit and Carpathian Cloud Subreddit).
We updated the layout and added more helpful links so users can more easily access resources.
Testers told us provisioning was taking too long, so we completely overhauled the flow and brought the time down to about 30 seconds. Based on a user's Spark allocations, CynosEngine will pre-provision an instance so it's ready to go as soon as you provision. There are still a few pre-flight checks it needs to do, but overall, there's a significant improvement.
If a standby instance still isn't available, it will fall back to the normal provision flow and take 3-5 minutes. Upgrades may also increase the time even if a standby unit is available. We'll continue testing and looking for ways to speed up the provisioning flow.
USC2 is now online! We're still working on getting the final integrations set up, but this will soon allow users to access our GPU infrastructure as well as more options for CPU types and larger core count upgrades.
With USC2 coming online, we'll be working to implement load balancing on the backend and automatic traffic handoff.
This update contains general security fixes across the backend. We added additional permissions for more granular control over what members are and aren't allowed to do.
You'll be able to add team members to help manage your infrastructure. Release is expected early 2026, though we'd love feedback if this is a feature users would appreciate at this scale.
Launch WordPress sites at the click of a button. All the configuration is automatic. We're building a special portal to manage the admin side of your WordPress site.
We plan to release this alongside WordPress hosting since it will make uploading and displaying content easier. No release date yet.
We're still actively developing database functionality. This requires some rearchitecting of the backend to allow virtual machines to be on the same subnet. We want to make sure this supports user workflows, so as we receive feedback, we'll implement accordingly.
We have the capability for users to create reseller accounts where they can manage client projects and cloud environments. We're using it for our own clients and software projects. We'd love feedback if this is something you'd like, otherwise we'll keep it internal.
We experienced an outage on Nov 25, 2025 at 2:30 CST due to a misconfiguration on the hypervisor self-healing agents. These agents are supposed to run security checks, rotate authentication, and send health reports to the database, but a security key was typo'd and caused them to go offline.
We got everything back online and rebuilt the failover so that if anything goes offline, the region will hand off to a running region. Users shouldn't experience more than 2-5 minutes of downtime (at the most). This is a major improvement from general cloud platforms since our goal is that if ANY node goes down, we always have a center that can run autonomously. We'd have to see every data center go offline to bring the infrastructure down.
Questions about these changes? Email us at info@carpathian.ai or check out our documentation.