Sheetz, the Pennsylvania-based convenience store chain with more than 800 locations across the eastern US, is packing up 11,000 virtual machines and leaving VMware. The company is moving its entire virtualization estate over to StorMagic’s SvHCI platform, and the decision came down to a simple reason: Broadcom created too much uncertainty.
The migration is a useful case study in what happens when a major vendor changes its licensing model, and why a company with thousands of distributed sites decided the risk was no longer worth carrying.
Why Sheetz is leaving
The trigger is Broadcom’s acquisition of VMware, which closed in late 2023 and was followed by an aggressive reworking of how VMware products are sold. Broadcom dropped perpetual licenses in favor of subscription bundles, simplified the product line into core bundles, and raised prices substantially for many customers. The changes generated pushback across the industry, and a fair number of organizations concluded that staying on VMware was a long-term budgeting gamble.
Sheetz infrastructure leadership made the same call. In the company’s telling, projected price increases and the shift to a subscription model created real uncertainty around long-term budgeting and around how dependent the company would be on a single vendor going forward. When the cost curve and the control you have over your own platform both become uncertain, staying put stops looking like the safe option. For Sheetz, that uncertainty was enough to justify a rewrite of its virtualization layer.
The scale of the move
This is not a small migration. Sheetz runs about 11,000 virtual machines spread across its operations, with 838 stores in the footprint. The stores themselves are the distributed part: each location runs its own infrastructure for point-of-sale, back office, and store-level applications, which is classic edge computing that needs to keep working even if the link back to headquarters drops.
Moving an estate that size means touching nearly every workload the company runs. It is the kind of project that takes months and is easy to underestimate, which is precisely why the company has been marching through it in stages rather than trying to do it all at once.
How the migration is going
According to reporting on the project, Sheetz had completed roughly 600 stores at a pace of about 200 per month. The remaining stores follow at the same rate. A store roughly every few hours is a steady, repeatable cadence, and it suggests the team built a migration playbook early and has been running it on loop rather than reinventing the process for each site.
The choice of StorMagic’s SvHCI is notable. StorMagic is known for lightweight hyper-converged infrastructure aimed at exactly this kind of distributed, edge-heavy deployment. Its storage and virtualization stack is designed to run on modest hardware at remote sites, which fits a convenience store footprint far better than a large centralized cluster. For Sheetz, which needed something that could live in a store closet and survive without constant attention, a purpose-built edge HCI product makes sense.
What this says about the broader market
Sheetz is far from alone. Since Broadcom took over VMware, a steady stream of organizations has announced plans to move off the platform, and the pattern shows up across universities, enterprises, and even public sector bodies. Most cite the same drivers: subscription pricing that is harder to predict, consolidation of products they depended on, and a sense that the roadmap is no longer aligned with what they need.
For the virtualization market, that has opened the door to alternatives. VMware still holds a dominant share of on-premises virtualization, but the exodus has created real runway for competitors, and StorMagic’s win at Sheetz is a concrete data point in that shift. Every large customer that leaves is both a lost account for VMware and a reference customer for whoever takes the workload.
Lessons for other enterprises
There are a few takeaways for anyone running a large on-premises workload who is watching this from the sidelines.
First, evaluate lock-in before it becomes a crisis. If you depend on a single virtualization vendor, know what it would take to leave, what alternatives would meet your needs, and where your biggest integration costs sit. That groundwork makes a forced migration far less painful than starting from zero in a panic.
Second, sizing the alternative to the actual deployment matters. Sheetz did not move to another big centralized platform. It chose a tool built for distributed edge sites. Matching the replacement to how your workloads are actually spread out will save you from overbuilding and from ongoing operational pain.
Third, plan a repeatable migration path. The store-per-cadence approach worked because the team standardized the process. Automate the repetitive parts, document the playbook, and treat the migration like a factory line rather than a series of one-offs. Across 838 sites, that discipline is what turns a daunting project into a schedule you can actually hit.
For Sheetz, the upshot is a virtualization stack it controls on hardware that fits its stores, without the pricing and licensing uncertainty that prompted the move in the first place. It is a big bet to rebuild 11,000 VMs on a smaller vendor, but it is also a bet on predictability, and right now that is exactly what the company says it was missing.