When operating in the cloud, a single workstation specification doesn’t fit every creative user. Historically, in traditional on-premise environments that approach needed to be the default setting, because any unused capacity is largely invisible once the hardware has been acquired. Media infrastructure always needs to be built for peak capacity when it’s in a physical environment.
In the cloud, additional CPU, GPU, memory, and storage is metered. What was previously hidden inside a capital expenditure becomes visible as an operational cost that scales across the business. For media organisations, designing cloud workstations is a continuous balancing act between creative performance, operational flexibility, and commercial efficiency.
Why “one-size-fits-all” doesn’t work
In an organisation with a variety of different creative roles, if every user receives the same heavyweight workstation configuration then that company is effectively paying for the most demanding workflow on every seat, whether it is used or not.
The storage impact alone can be significant, add to that the unused compute, additional applications and media tools, and you increase the size of the workstation cost footprint unnecessarily. When that is multiplied across dozens or hundreds of virtual machines in a cloud environment, it quickly becomes an unsustainable recurring cost
Start with roles, before resources
A more effective approach is to define workstation personas first and resources second. Creative teams often need some flexibility, and every media organisation has unique combinations of workflows and processes. But tailored workstation templates that carry the configuration appropriate to each role will help avoid unnecessary overhead.
A typical media deployment might maintain separate workstation templates for:
- Editorial
- VFX and compositing
- Colour and finishing
- QC, compliance and versioning
- Graphics and motion graphics
- Marketing and promotion
This keeps workstation requirements proportionate to each workflow, simplifies governance, and allows companies to evolve each persona as needs change. The templated system in Lens supports this approach by enabling standardised provisioning while preserving flexibility across different creative roles, allowing changes to be made easily as workflows evolve.
The same software often behaves differently
A subtle but important point is that identical applications do not imply identical infrastructure requirements. Editors and assistants may both be using the same NLE, yet their workloads can differ dramatically. An editor working with long-form UHD material, multiple effects layers, and extended playback sessions may stress GPU and storage throughput continuously. An assistant editor performing ingest, sync, metadata, and multicam preparation may place heavier demands on memory and CPU during specific operations.
Camera formats, codecs, frame rates, multicam counts, and delivery requirements all influence workstation behaviour. So setting up performance profiles in advance, ensures that cloud resources are deployed in the most efficient and intelligent way possible.
Use workload data to right-size continuously
The most successful cloud usage adapts to requirements, and leverages operational analytics. By analysing real workload behaviour across different personas, infrastructure teams can observe CPU utilisation, GPU utilisation, memory pressure, storage activity, and session performance over time. After auditing their workflows they can then adjust workstation sizing and measure the effect on user experience.
An editorial Pool may be increased from a mid-range GPU tier to a higher tier if playback metrics consistently show dropped frames during peak activity. Conversely, a graphics Pool may be reduced if utilisation data shows substantial unused headroom. Lens is particularly valuable here, because it provides workload and utilisation visibility across workstation Pools rather than isolated devices. That allows management teams to evaluate trends across departments and productions and adjust accordingly. The key shift is from provisioning by assumption to provisioning by evidence.
Flexibility matters as much as optimisation
There is a temptation to optimise cloud usage aggressively, but media production is rarely static. A team that has been working on HD factual programming may suddenly begin a UHD HDR project with very different requirements.
Role-based workstation design should therefore preserve operational flexibility. The Image Library, standardised workstation templates, and automated provisioning in Lens make it possible to introduce a new workstation profile quickly without rebuilding the entire environment. This is one reason many media organisations maintain a small number of higher-specification Pools for exceptional workloads while keeping the majority of seats aligned to typical production demand.
The commercial conversation is ongoing
Cloud usage in media workflows is not a quest for the absolute best technical experience, nor the absolute lowest possible cost. It is a continuous negotiation between creative expectations and commercial reality. If budgets were unlimited, every user could receive the highest available GPU tier and the largest memory configuration. In practice, organisations must decide where additional performance genuinely improves productivity, and where it simply creates unused capacity.
The most effective infrastructure teams make those decisions transparently. They use workload data, session analytics, and utilisation reporting to explain the trade-offs between performance and cost, enabling production leadership to make informed decisions rather than relying on anecdotal feedback. They also take advantage of the cloud’s ability to provide rapid access to newly released hardware, including the latest GPU instance types and compute families, without waiting for a hardware refresh cycle.
Lens makes it straightforward for administrators to test these newer instance families with selected workstation Pools, evaluate the impact using the same utilisation and performance analytics, and then update virtual workstation templates and Pools through Lens tooling if the new hardware delivers a measurable benefit.
Designing for the role improves the experience
Counterintuitively, giving every user the same powerful workstation often produces a worse overall experience than giving each user a workstation designed for their actual role. Purpose-built workstation personas reduce unnecessary clutter, shorten provisioning times, simplify support, improve governance, and direct infrastructure investment toward the workflows that genuinely need it. They also create a clearer operational model for both post-production and IT teams.
Cloud infrastructure strategy that is tailored to the media industry, is about building the right resources, for the right creative task, at the right moment. Organisations that adopt that mindset tend to achieve both better user satisfaction and better control of operating costs.