I wrote a post in 2011 pointing hosting and cloud providers at Parallels Automation as the fastest path to a Microsoft Hyper-V cloud service — automated provisioning, subscription billing, control panels, the whole turnkey stack. Every link in that post is dead now, and so is the product line it described. But the company behind it didn't disappear. It split.
The Split Nobody Explained at the Time
In December 2015, Parallels broke into two separate companies. The automation and billing business — the actual subject of my 2011 post — had already been renamed Odin that March, and Ingram Micro bought it for $163.9 million the same December. Odin became the CloudBlue brand in 2018. If you want that side of the story, including CloudBlue's own recent change of hands, the CloudBlue side of that split has its own recent story worth a separate read.
The other half — the OS-level virtualization and container technology, tracing back to SWsoft in 1997 and the OpenVZ open-source project in 2005 — kept the Virtuozzo name and went independent. It's still independent today, headquartered in Schaffhausen, Switzerland, and it's the company I want to talk about here, because 2026 gave it two reasons to matter again.
Why Virtuozzo Is Relevant Again in 2026
The first reason is Broadcom. VMware's acquisition pushed a lot of service providers into an unplanned platform search, and Virtuozzo has been positioning directly at that gap — a free migration path for partners leaving VMware's Cloud Provider Program, with a named case study (Voyager, a New Zealand sovereign-cloud provider) as proof the migration path works in practice, not just on a slide.
The second reason is structural. In February 2026, Virtuozzo unified its virtualization, storage, networking, and PaaS products into a single offering it calls the Virtuozzo Infrastructure System — hyperconverged, multi-tenant, IaaS and PaaS in one stack, built on an OpenStack foundation but packaged so a service provider isn't assembling it themselves. That's a materially different product than what shipped a decade ago, when Virtuozzo was closer to a standalone container hypervisor than a full cloud platform. The platform breadth traces to acquisitions, too — OnApp (cloud management for service providers) and Jelastic (multi-cloud PaaS) both joined in 2021, and the current stack is the result of folding those in rather than building everything natively.
The AI Pivot, and Where to Be Skeptical of It
In May 2026, Virtuozzo leaned hard into AI and GPU infrastructure messaging — positioning itself as "built with AI, by AI, for AI" and claiming 60-80% lower total cost of ownership than competitors for AI workloads. That's Virtuozzo's own claim about its own product, not a figure I've seen independently verified anywhere, and I'd treat it exactly that way until a third party tests it. Vendor TCO math on AI infrastructure is a genre unto itself right now — almost everyone selling GPU capacity claims a dramatic cost advantage, and the assumptions buried in those numbers rarely survive a real workload comparison.
What I can say with more confidence: the AI positioning sits on top of infrastructure Virtuozzo was already building for a different reason — multi-tenant hyperconverged capacity that MSPs can carve up and resell. Whether that translates into a credible AI/GPU offering for a mid-market service provider, versus repackaged messaging on the same virtualization stack, is the open question I'd want answered before recommending it for that use case specifically.
The Throughline Back to 2011
Here's the part that actually connects to what my old post was about. That post wasn't really about Hyper-V — it was about the plumbing a service provider needs to sell cloud as a product: provisioning, metering, billing, a control panel a customer can log into. Virtuozzo's current Application Platform still ships that plumbing. It includes a built-in control panel for host tuning, tariff management, and quota enforcement, an internal metering and charging subsystem (Virtuozzo calls it JBilling), and pre-built integrations with billing platforms MSPs already run — HostBill, OBAS, Stripe, WHMCS — plus support for custom integrations. Virtuozzo's own documentation cites more than 100 service providers using this billing layer; that figure comes from Virtuozzo's docs, not an independent count, so I'd treat it as a directional claim rather than an audited number.
That's the legitimate lineage here, and it's more useful than the corporate-history version. The 2011 post was about the tooling a service provider needed to turn infrastructure into a billable product. Fifteen years and one corporate split later, that's still what this company sells — it's just running on containers and hyperconverged infrastructure instead of a Hyper-V integration guide, and it's now pitched at providers absorbing VMware refugees and AI capacity demand instead of providers standing up their first Hyper-V cloud.
Who's Running It Now
Kurt Daniel became CEO in February 2026. Founder Serg Bell stayed on as Chairman, styled internally as "Chief Constructor" — worth noting because founder-retained-as-technical-lead is a different governance signal than founder-exits-entirely, and it usually means product direction stays close to the original technical thesis even as go-to-market leadership changes.
What I'd Verify Before Recommending It
A few things I'd want more than a single source on before putting this in front of a client:
- Partner count and geography. The figure I have — roughly 550 MSP and channel partners across 80+ countries, with about 90% of revenue coming from MSPs hosting SMB tenants — comes from one trade-press source (Blocks and Files, May 2026), not from Virtuozzo's own investor-grade disclosure. Treat it as directional.
- Employee count. Same source puts headcount at roughly 250-300. No independent confirmation.
- The TCO claim. Covered above — vendor math, not verified math.
- The "100+ service providers" billing-platform figure. Vendor documentation, not an audited or independently reported number.
None of that means the platform isn't worth evaluating — it means it's worth evaluating with your own reference calls, not with the vendor's own numbers as the whole case. That's the same standard I'd apply to any provider pitching a VMware-exit story or an AI-infrastructure story right now, because both categories are full of vendors making claims the market hasn't had time to test yet.
Comments
Post a Comment