You might recall I covered Eon about a year ago, not long after they came out of stealth and coined the term Cloud Backup Posture Management. If you missed that one: Eon's whole pitch is that backup shouldn't be, as their Chief Architect Dr. Assaf Natanzon put it, "a black box that you just pay for and get nothing back."
Go read that recap here if you want the full context. For today, the short version is: automatic classification instead of manual tagging, granular restore instead of all-or-nothing, and backup data you can actually query.
|
This blog recaps Jeremy's short video on Eon. |
So, a year on, what have they been up to? Two things worth covering: their Summer of Shipping, five days, five releases, and then the bigger story underneath all of it, the live data lake.
Five Days, Five Releases
Rather than walk through every individual feature, the pattern that matters here is that Eon is pushing further into the platform services space — the stuff developers have historically had to protect with their own scripts. Over five days, that meant:
- MongoDB Atlas — continuous protection with recovery down to a document, collection, or point in time.
- Oracle on Amazon RDS — one less database sitting on a bespoke script instead of a real backup policy.
- Amazon Keyspaces for Apache Cassandra — automated discovery, point-in-time recovery, no custom backup infrastructure to maintain.
- Azure Files — centralised policy and monitoring, even though the snapshots themselves stay in the customer's Azure account.
The One Worth Slowing Down For: EC2 Instant Restore
The feature I want to flag properly is EC2 Instant Restore, because there's some genuinely clever engineering quietly sitting underneath it.
Eon's built a lightweight hypervisor that sits beneath the guest OS and does read interception. Restore an EC2 instance, and it boots virtually instantly, you're not waiting for every volume to fully hydrate first. As reads come in, the hypervisor pulls back exactly the data that's needed, when it's needed, while the rest quietly rehydrates in the background. You get your instance back in minutes, not hours, and nobody's sitting there watching a progress bar before they can log in.
I wouldn't be surprised if that hypervisor turns up powering other features down the line. Once you've built read interception at that level, there's no reason to stop at EC2.
Backup Goes Up the Stack: M365 & Google Workspace
The bigger platform news is Eon moving up into SaaS. Microsoft 365 backup is the headline release, and it brings the same Eon experience you'd expect: query your backup data through the Eon MCP layer, and get ransomware-style recovery that takes you back to the last known good copy in a granular fashion rather than a full mailbox or site restore.
That follows quietly on from Google Workspace support a month earlier; Gmail, Drive, Shared Drives, Groups. Put those two together and Eon's now covering IaaS, a growing list of platform services, and SaaS. That's a real broadening of scope, and I'd expect it to keep going.
The Real Story: Backup Becomes a Data Lake
Backup getting easier is genuinely useful. But the bigger piece of value is what happens to that data once it's captured.
Behind the scenes, Eon turns your backup data into a cloud-based data lake, stored on Eon's own storage tier, and you get to choose which cloud platform it lives on. That data is then exposed through public interfaces including MCP, Iceberg, S3, and the Eon API. For your data and analytics teams, that means access in the formats they already work in, instead of a pile of backup sitting there doing nothing useful for 364 days of the year.
Eon's own framing of the problem is blunt, and it's the same one Natanzon gave me a year ago: today's cloud backups are black-boxed, locked in snapshot formats, holding duplicated data your data teams can't touch. That's exactly the gap this closes.
How it actually works: Eon does automatic data type conversion and schema inference behind the scenes, so what comes out the other end is properly queryable, not a dump of raw bytes. And critically, this isn't about replacing your existing data and AI stack, it's about making backup data available into it. If you're running Databricks today, Eon's captured data can land in your existing data catalog, inheriting the governance and security you've already built, and get curated from there for your BI team or whatever else needs it.
That's the trick: taking data that would otherwise sit at rest and making it useful again, without you having to build a data pipeline to do it.
The Math That Makes It Free
The objection I hear constantly is "I've just done my budgeting round, no money till next year." Eon's answer to that is worth sitting with: they claim the storage savings alone are enough to offset the cost of the backup, making it effectively free compared to native cloud backup.
Compare that to native snapshots, you're not just getting materially more granular protection, recovery, and query capability; you're also getting a data asset your teams can use for other things. We've seen a version of this before with copy data management, where dev teams want a live, queryable copy of production data without hammering the production system itself. Eon's doing something similar, except it falls out of the backup you were already paying for.
Cost-neutral, more capable, and useful beyond disaster recovery; that's a fairly compelling combination even in a year with no budget left.
That's it for this one, a quick update, but a good look at how fast Eon's shipping and how they keep finding new ways to make backup data actually earn its keep. Thanks for watching, and I'll catch you next time.
