Should I replace our ZFS file servers to improve Veeam backups?

For ordinary file shares, a simple replica or rsync-style copy can be useful as an additional recovery layer, but it should not be treated as the only backup. Replication can copy accidental deletions, corruption, or ransomware just as efficiently as healthy files. Keep versioned snapshots and an isolated or immutable off-site copy for disaster recovery.

Windows Server with DFS Namespaces is a reasonable option if your organization already has the Windows skills, licensing, and operational processes. Veeam deployments commonly protect Windows file servers without trouble, especially when the data is divided into reasonably sized volumes or shares.

That said, DFS is mainly a namespace and referral technology; it is not a backup or storage-replication solution by itself. Before migrating, test real ACLs, long paths, large engineering files, and both individual-folder and large-scale restores. The migration is only worthwhile if it improves administration, access, or recovery—not just because the current backup job was designed as one huge chain.

Another option is to reconsider the backup product rather than replacing the file servers. Products designed around large NAS datasets or object storage may handle this layout more naturally, and object storage can avoid the practical limitations of a single massive VBK file. Since your Veeam environment is already established for the HCI infrastructure, though, it may be simpler to keep Veeam there and use a separate, purpose-built method for the file data.

Whichever route you choose, define the required restore time for one project, the acceptable amount of lost changes, and the retention needed after a site failure. Then perform a real restore test to a clean location. Storage capacity alone is not enough to determine the best design.