Version 9.2 is out: here is what you need to know
The new official release came out in mid-May 2026 (yes, a stable release!) and is based on Debian 13.5 “Trixie”. It brings a set of improvements well suited to today's virtualisation needs.
In a nutshell, what does version 9.2 change?
It comes with several significant software improvements that make it a major release compared with 9.0 or 9.1: a more recent base system, stronger security and components better optimised for modern production environments.
The essentials at a glance 🔑 (comparison table)
| Feature | Previous version | Version 9.2 |
|---|---|---|
| Debian base | Debian 13 Trixie | Debian 13.5 Trixie |
| Linux kernel | Kernel 6.x LTS/Testing | Kernel 7.0 (stable) |
| QEMU | QEMU 10.x | QEMU 11.0 |
| LXC | LXC 6.x | LXC 7.0 |
| ZFS | ZFS 2.3 | ZFS 2.4 |
| SDN | IPsec/OpenVPN | Native WireGuard + BGP |
A quick comparison of the major changes in Proxmox VE 9.2
🍊 Debian 13.5 + Kernel 7.0 by default
This is the most important change under the bonnet. The 9.x series has been based on Trixie (Debian 13) since 9.0; 9.2 now builds on Debian 13.5 and moves to kernel 7.0 by default.
What it brings:
- ✅ A secure system without complex configuration
- ✅ Better hardware compatibility: recent GPUs and NVMe drives supported natively
- ✅ VMs start around 15% faster on average
Please note:
Some very old drivers or hardware that is several years old may need to be checked beforehand.
⚖️ Dynamic Load Balancer: automatic workload management
A key feature for multi-node clusters: Proxmox balances the load across its servers by itself.
How does it work?
- Set up a cluster with 3 or more nodes
- If server no. 1 starts to saturate (high CPU/RAM usage), Proxmox automatically moves some VMs to the other nodes
- The monitoring cycle runs roughly every 30 minutes
Why it is useful in practice:
- ✅ No need to be the human firefighter (🚒) whenever a server is overloaded
- ✅ Fair distribution of resources across all nodes
- ✅ The best possible virtualisation density
Important: this only works with at least 3 servers in the cluster. On single-node or two-node setups, the classic approach is perfectly sufficient!
🌐 SDN with native WireGuard and BGP
Proxmox networking takes a leap towards modern technologies: WireGuard is now built in, alongside EVPN for BGP.
The practical benefits:
- 💨 WireGuard: much better performance than IPsec (often twice as fast over the same physical link)
- 🔐 Modern, auditable cryptography (ChaCha20-Poly1305)
- 🛠️ Built-in BGP to filter routes between multiple datacentres
Typical use cases:
Secure connection to a remote cloud, interconnection of several corporate sites, a virtual VLAN shared between hosts without complex network configuration.
All of this can now be configured from the Proxmox VE web interface! No more copying and pasting configuration files manually over SSH.
🔄 QEMU 11.0 + LXC 7.0 + ZFS 2.4: an optimised stack
As with every major release, the low-level components have been updated to get the most out of the whole ecosystem:
| Component | Previous version | Version 9.2 | Visible gain |
|---|---|---|---|
| QEMU | ~10.x | 11.0 | Faster snapshots 🚀 |
| LXC | ~6.x | 7.0 | Containers ready in a fraction of a second ⏱️ |
| ZFS | 2.3 | 2.4 | LZ4-ZF compression with a lighter CPU footprint 💪 |
In practice: your snapshots are created faster and compression makes better use of NVMe storage (~60K random IOPS vs ~45K with the previous generation, according to comparative tests).
🛡️ Simpler HA: the Disarm/Arm maintenance workflow
High availability management has been reworked to avoid surprises when nodes are taken offline manually.
Previously, disabling HA on a server could trigger unpredictable failovers. The new “Disarm” mode allows targeted maintenance by preparing the transfer of active services before a power-down (or a forced reboot if necessary).
Why it is better:
- ✅ Simpler planned maintenance
- ✅ Manual control over failover before hardware work
- ✅ Fewer unexpected service outages
⚠️ Before upgrading in production: points to watch
| Aspect | Practical advice |
|---|---|
| Backup | Full backups of your VMs and configurations are a must! It is always a good habit. 🔒 |
| Older hardware | If your hardware is 5 to 7 years old, check driver compatibility (GPU/CPU drivers). |
| Small clusters | The DLB and SDN features only work on clusters of 3 or more nodes. |
| Ceph users | Ceph users are advised to move to Tentacle 20.2 for official support. |
🤔 Quick FAQ: the questions we are often asked
Q: I am on Proxmox 8.x. Can I upgrade straight to 9.2? 💡
A: Technically yes, using the pveupgrade command, but the community recommends going through 9.0 or 9.1 first to avoid conflicts between packages and the kernel.
Q: Will it have a negative impact on my performance? 📊
A: No! Benchmarks show improvements across the board: network throughput, storage IOPS and VM start-up. The new ZFS, for example, compresses better with a lower CPU load.
Q: Is it essential on a single machine? 🖥️
A: There is no rush if your current setup meets your needs. New features such as DLB or SDN are of no use if you only run a single standalone server.
🎯 Our recommendation (honest and clear)
Already running Proxmox 9.0 or 9.1 in production?
→ Upgrading is recommended but not urgent if everything is working well. The current release is stable and optimised for today's workloads.
Moving from 8.x to the new 9.x?
→ Now is the ideal time, especially if you run multi-node clusters or use advanced networking features.
Existing multi-node cluster 🏢
→ Upgrade without delay to benefit from the new Dynamic Load Balancing and the WireGuard/BGP networking improvements.
New to virtualisation? 🆕
→ The move to Debian 13.5 + kernel 7.0 provides a solid, secure foundation for the long term. As a beginner you will not gain a great deal, but upgrading is still a good idea.
✅ Checklist before you start
For those who want to deploy this release on their existing Proxmox infrastructure:
- Full backup of all critical VMs
- Driver compatibility test on any older hardware
- Gradual upgrade (one node first, then the rest of the cluster)
- Validation/rollback script ready if needed, just in case
- Maintenance window scheduled by your administrators
💬 In conclusion: 3 key takeaways
| Summary | |
|---|---|
| Type of release | Not a revolutionary overhaul, but a stable release with useful software improvements for production. |
| For clusters | DLB + WireGuard/BGP are well worth it if your infrastructure has several nodes. 🏢 |
| Standalone VMs | You can wait for the next version if your current setup fully meets your needs. 🖥️ |
All in all, Proxmox VE 9.2 is a natural step forward for modern system administration tools, with tangible but incremental improvements (no sudden revolution). Migrating is not urgent, but it is strongly recommended for growing production environments.
🚀 Ready to move to Proxmox VE 9.2?
Download the official ISO: download.proxmox.com
Read the official release notes: Proxmox documentation
