Microsoft is preparing to phase out a range of older Azure VM instance types. This summer, the company will stop offering new long-term reservations for a large number of variants, after which some of them will be permanently phased out in 2028.
Starting July 1, Microsoft will no longer accept one-year reservations for thirteen older instance types, reports The Register. These are variants that have been widely used, particularly over the past decade, but now run on outdated hardware. These systems will remain available for a few more years but will be decommissioned in May and November 2028.
In addition, four other instance types will no longer allow new one- or three-year reservations. These will remain operational after 2028, but are no longer aligned with the platform’s future development. Microsoft is directing customers to migration documentation to ensure workloads are transferred to newer alternatives in a timely manner.
Renewal of underlying infrastructure
The instance types in question largely date from the period when cloud adoption was growing rapidly. Since then, Microsoft has introduced multiple generations of infrastructure, with recent systems offering greater computing power and more efficient energy usage. This allows more virtual machines to be deployed on the same physical hardware, delivering benefits in terms of both performance and cost.
Although Microsoft does not provide a detailed explanation, it is clear that the processor generations used matter. Many of these instance types are based on older Intel Xeon processors, which have since been surpassed by newer architectures. For most workloads, migration will present few technical obstacles, as compatibility issues are limited.
For customers, however, this does mean that a transition will be necessary. Organizations that still rely on these older instance types will face a fixed end date, whereas such decisions in on-premises environments can often be planned more flexibly. At the same time, the switch offers opportunities to improve performance and more efficiently use resources.