Cloud was the future. For more than a decade, broadcasters, streamers and technology vendors have been moving infrastructure, workflows and operations into cloud environments with the expectation that flexibility, scalability and software-driven architectures would fundamentally modernise the industry. Streaming growth, global distribution, remote production, faster service iteration and changing economic pressures all accelerated that transition.

And in many ways, the shift has absolutely delivered. Large-scale live streaming is now routine. Pop-up channels can launch in hours rather than months. AI-assisted localisation, dynamic scaling and software-driven operations are becoming normal parts of modern media infrastructure.

But beneath that transformation sits a more complicated question: has the industry actually become cloud-native, or have many organisations simply relocated traditional vendor systems into cloud environments? The distinction matters because cloud adoption alone does not change the operational model.

In many deployments, proprietary broadcast systems still sit at the centre of the architecture. Workflows remain tightly coupled to vendor implementations. Operational logic is embedded inside platform-specific tooling. Monitoring depends on proprietary telemetry frameworks. Some of the opaque operational behaviours that once existed inside hardware appliances can reappear inside software stacks and managed services.

Cloud-native is not just cloud-hosted

The more important transition now emerging is towards genuinely vendor-agnostic operational models. Kubernetes, containerised workloads, APIs and software-defined orchestration are beginning to create portable media platforms that are less tightly bound to individual infrastructure providers or proprietary vendor stacks.

Emerging frameworks such as the EBU Dynamic Media Facility (DMF), DPP LPX and Time Addressable Media Store (TAMS) reinforce that direction by attempting to define media workflows independently of underlying infrastructure.

The hyperscalers themselves are not the issue. Their infrastructure and managed services have unquestionably accelerated innovation across media. The industry would not have achieved the current pace of experimentation in live production, localisation, AI-assisted workflows or large-scale streaming without them.

The challenge is the default way many organisations have chosen to consume those platforms. Media organisations may physically leave the data centre while still operating cloud-hosted equivalents of traditional broadcast architectures.

The rise of multi-vendor dependency

Historically, broadcast environments became operationally dependent on relatively small numbers of specialist suppliers. Workflows, integrations, operational procedures and engineering knowledge accumulated around tightly coupled systems that were difficult to evolve or replace. Modern cloud deployments can recreate many of those same conditions, except the new dependency model is often multi-vendor rather than single vendor.

At first glance, this can appear healthier. Organisations may no longer rely on one or two major broadcast suppliers. Instead, they assemble complex ecosystems spanning hyperscaler services, orchestration platforms, observability vendors, AI providers, workflow engines, transport layers, SaaS products and proprietary middleware.

But simply increasing the number of suppliers does not automatically create openness. Each platform introduces its own APIs, telemetry models, orchestration logic and specialist knowledge requirements. Integration itself becomes a permanent engineering discipline; a workflow may technically span five or six different vendors while still remaining deeply locked into the combined behaviour of that ecosystem.

When integration becomes immovability

This is particularly important in media because workflows are unusually stateful, timing-sensitive and operationally interconnected. A broadcaster may technically be able to redeploy workloads elsewhere, but practical migration becomes increasingly difficult once observability, orchestration, media transport and AI tooling are distributed across multiple tightly integrated platforms that evolve independently.

The observability challenge becomes even more complicated in these fragmented environments. Infrastructure metrics may still look healthy, but understanding why a live workflow degraded, why SCTE signalling drifted, or why a multilingual audio chain failed becomes significantly harder once telemetry itself is fragmented across multiple vendor ecosystems.

Read more: Observability: The missing discipline in cloud-native media operations

The danger is not merely commercial dependency, but operational dependency created through accumulated complexity. A media organisation may discover that its workflows no longer truly exist independently of the surrounding vendor ecosystem. Engineers become experts not in transferable operational models, but in managing the interactions between multiple proprietary systems. 

None of this means every workflow should prioritise maximum portability at all costs. Deep platform integration can simplify operations, accelerate deployment and reduce engineering overhead. For many organisations, accepting a degree of dependency is an entirely rational trade-off.

The problem is not integration itself. It is losing sight of where operational dependency accumulates. The goal is not purity or total abstraction, but retaining enough operational ownership to preserve strategic flexibility as technologies and business pressures evolve.

Platform ownership versus platform consumption

True platform ownership is not simply about avoiding a single dominant supplier. It is about retaining control of operational intent, workflow logic and data structures independently of the surrounding vendor landscape. That requires deliberate architectural decisions.

This is precisely why initiatives such as DMF, LPX and TAMS matter beyond pure technical modernisation. Their long-term value lies in attempting to separate media workflows from specific infrastructure providers and vendor implementations.

Rather than being tied to proprietary systems, these models treat media as observable and exchangeable entities that can move across environments while preserving operational continuity and metadata integrity. Without that abstraction layer, many so-called ‘cloud-native’ deployments risk becoming sprawling collections of interconnected proprietary systems across disparate cloud providers and on-premise installations.

The operational consequences are already becoming visible. A live sports workflow may integrate AI-driven real-time translation services directly into the production chain, with orchestration logic, prompt structures and monitoring pipelines all deeply tied to a specific vendor’s tooling. Elsewhere, observability itself can become fragmented across multiple platform-native systems, while signalling and workflow logic, including SCTE-35 handling, become embedded across proprietary automation layers that are difficult to inspect, correlate or reproduce independently.

Where control actually matters

This creates difficult long-term questions that many organisations are only beginning to confront.

Can operational workflows survive a platform migration? Can observability data move coherently between environments? Can media supply chains continue functioning if commercial relationships or geopolitical realities shift? Can AI-enhanced workflows be reproduced independently elsewhere?

Across Europe, issues around sovereignty, resilience and strategic dependency are increasingly shaping infrastructure discussions. Broadcasters and media companies are beginning to recognise that platform ownership is not simply a technical architecture question. It is becoming a strategic operating model question.

Mature cloud-native thinking increasingly means treating infrastructure providers and installations as interchangeable execution environments rather than permanent operational identities. That means investing in portability, open interfaces, transferable operational models and observability layers that describe media intent rather than platform-specific behaviour.

Cloud-native media was never supposed to mean assembling a larger collection of systems that are individually flexible but collectively immovable. The organisations that truly succeed over the next decade will be the ones that recognise that distinction.

Cloud hosted or cloud-native: What’s the difference?