Proxmox lacks native, easy-to-configure shared storage for Fibre Channel and iSCSI, which is a hard blocker for enterprise adoption.
A software-defined storage appliance (virtual or physical) that integrates natively with Proxmox, providing enterprise shared storage with a simple setup wizard, replication, and HA — no Ceph complexity required.
Subscription per TB managed ($10-30/TB/month) or annual node license
This is a hard blocker, not a nice-to-have. Teams literally cannot run HA Proxmox clusters without shared storage. The Reddit thread with 950 upvotes specifically calls out shared FC/iSCSI as THE missing piece. Every Proxmox migration guide has a 'figure out storage' section that loses people. Ceph is the default answer and it's widely hated for small/mid deployments.
Proxmox claims 800,000+ installations. If even 5% are clusters needing shared storage (40,000), at $200-500/month average, that's a $96M-$240M TAM. The VMware exodus is adding thousands of new Proxmox shops monthly. However, this is still a niche within infrastructure — it's not a billion-dollar TAM, but it's very reachable for a bootstrapped or seed-stage company.
These teams are ALREADY paying for storage solutions — TrueNAS Enterprise, Starwind, or dedicated SAN hardware ($10K-$100K). They switched to Proxmox to SAVE money vs VMware licensing. They have budget freed up from VMware and are actively looking to spend it on making Proxmox production-ready. $10-30/TB/month is a fraction of what they were paying. Enterprise storage buyers are conditioned to pay for reliability.
This is genuinely hard. Building a reliable storage layer involves kernel-level work (DRBD, LVM, ZFS), iSCSI target daemons, Fibre Channel HBA integration, replication engines, failure detection, and fencing. You can build an MVP using existing open-source components (DRBD + LINSTOR, or ZFS + targetcli) with a Proxmox-native UI wrapper, but getting it production-reliable for enterprise data is a 3-6 month effort minimum for an experienced storage engineer. A solo dev without deep Linux storage experience cannot build this in 4-8 weeks. This is infrastructure software where bugs cause data loss.
Nobody owns 'Proxmox-native shared storage' today. Ceph is integrated but hated for small clusters. Every other option (TrueNAS, Starwind, LINSTOR, Open-E) treats Proxmox as a secondary citizen. There is a massive gap for a solution that speaks Proxmox natively — auto-discovers nodes, integrates with the Proxmox API, shows up in the Proxmox UI, handles VM-aware snapshots and replication. First mover advantage is real here.
Storage is inherently recurring — data grows, clusters expand, support is ongoing. Per-TB pricing scales naturally with customer growth. Subscription model is standard in enterprise storage (LINBIT, Open-E, Starwind all use it). Customers cannot easily switch storage once VMs are running on it — very high switching costs create strong retention. Annual node licensing also works well.
- +Massive unmet need validated by large-scale community signal (950 upvotes, Proxmox's own $50M revenue growth)
- +VMware-Broadcom chaos creating a generational migration wave toward Proxmox
- +No one owns the 'Proxmox-native storage' position — first-mover opportunity
- +High willingness to pay from teams already budgeted for enterprise storage
- +Extreme customer lock-in once deployed — storage is the stickiest infrastructure layer
- +Per-TB subscription model scales naturally with customer growth
- !Proxmox itself could build native shared storage (they have the resources now at $50M revenue and this is their most-requested feature)
- !Storage software bugs cause DATA LOSS — the reputational and legal liability is existential
- !Requires deep Linux storage engineering talent, which is rare and expensive
- !LINSTOR/DRBD community plugin could mature and become 'good enough' for most users
- !Enterprise sales cycles are long (3-6 months) even when pain is acute
- !Ceph is improving its small-cluster story with each release — the gap may narrow
Open-source distributed storage system with native Proxmox integration. Provides block, object, and file storage across commodity hardware.
Enterprise-grade NAS/SAN platform supporting iSCSI, NFS, SMB, and Fibre Channel. Available as TrueNAS CORE
Software-defined storage that creates shared HA storage for hypervisors. Supports iSCSI and NFS with synchronous replication between nodes.
Open-source SDS using DRBD for synchronous block replication. LINSTOR is the management layer that orchestrates DRBD volumes across cluster nodes.
ZFS-based software-defined storage platform providing iSCSI, Fibre Channel, NFS, and SMB with HA clustering and replication.
A virtual appliance (OVA/qcow2) that deploys as a VM on Proxmox, auto-discovers other Proxmox nodes via API, presents iSCSI targets using ZFS + targetcli under the hood, includes a simple web wizard for creating shared storage pools, and auto-configures Proxmox to use them. Ship with 2-node synchronous replication using DRBD. Skip Fibre Channel for MVP — iSCSI covers 80% of SMB use cases. Include a Proxmox UI plugin that shows storage health in the native dashboard.
Free tier: single-node, up to 2TB, community support → Paid tier ($10-30/TB/month): multi-node replication, HA failover, snapshots, email support → Enterprise tier ($50+/TB/month or annual node license): FC support, priority support, SLA, compliance certifications, professional services for migration from VMware/Ceph → Scale: managed storage-as-a-service for MSPs managing multiple Proxmox clusters
4-6 months to first paying customer. Month 1-2: build MVP appliance with iSCSI + basic replication. Month 3: private beta with 5-10 Proxmox community power users from Reddit/forums. Month 4-5: iterate based on feedback, harden reliability. Month 6: launch paid tier. The long pole is trust — storage customers need to believe you won't lose their data, which takes testimonials and time. Early revenue likely comes from annual prepaid licenses from beta users who helped shape the product.
- “If Proxmox had a good file system so that you can easily do shared Fibre Channel or iSCSI storage”