Delivering Olympic coverage across Europe requires production teams to manage hundreds of feeds, multiple editorial teams and simultaneous live events. For Warner Bros. Discovery, the Olympic Winter Games Milano Cortina 2026 provided another opportunity to operate and refine a production platform, driven by Grass Valley’s AMPP using Framelight X and Playout X, designed to support that scale.
Milano Cortina was my seventh Olympics working with Eurosport and Warner Bros. Discovery. Each Games brings different operational challenges, but this one highlighted how much our production infrastructure has evolved in recent years.
Over several Olympic cycles we have moved towards a shared production platform connecting editorial teams, operators and engineers across our operations. The aim has been to support both major live events and day-to-day sports production through the same infrastructure.
Earlier Olympic workflows were organised differently. Production environments were often built around individual events or teams, which meant systems and capacity were spread across multiple locations. Over time, we moved towards a unified platform – built on Grass Valley’s AMPP and underpinned by Framelight X and Playout X – that was capable of supporting Olympic coverage alongside our regular sports output, as well as adding capacity for new channels and capabilities.
Today, through Grass Valley’s Framelight X, teams can access the same production environment from multiple locations. During the Milano Cortina Games we saw around 650 concurrent users working on the platform at peak, with an average of more than 600 users per day across Europe and the US.
Supporting that level of collaboration requires infrastructure that can evolve without interrupting live production. The core platform powered by Grass Valley’s AMPP operates on a microservices architecture, with the Framelight X workflow engine connecting ingest, content management, central LiveTouch replay management, playout (Playout X) as well as post production. This allowed engineers to update individual components without affecting the wider system.
In the past, system upgrades were often lengthy processes requiring updates to each individual user’s computer, but changes can now be introduced gradually while production continues. With the GV team, we reconfigured the core AMPP platform through software-defined production for the event instead of rebuilding it as a one-off workflow. This also changed how we introduced new services, by extending existing resources rather than deploying new infrastructure.
During Milano Cortina, this approach supported the launch of additional pop-up channels alongside our core services.
Central to all of this was the Framelight X story-centric, continuous media production environment, which is specifically designed for high-demand and distributed operations like ours. Framelight X federated all of our content into a single system, giving us streamlined access to everything we needed – from ingest through to editing and publishing – from any location.
Scaling ingest and managing content
One of our priorities ahead of Milano Cortina was ensuring the platform could handle the volume of content we expected to record. Unlike a summer Olympics where most other major competitions are no longer running, the Winter Games sit right in the middle of an extremely busy global sports calendar. So in addition to our Winter Olympics coverage, we were also delivering our regular sports output across Eurosport and TNT Sports, which meant managing multiple events simultaneously. That complexity is added to by the Winter Games being more susceptible to weather delays and last minute rescheduling, as we experienced this year with events being postponed by snow delays.
Across the Games, the platform handled more than 62,000 hours of recorded content, with peaks of over 4,000 recording tasks in a single day.
Working with Grass Valley, we completed a migration project that modernised our ingest workflow and expanded recording capacity, moving us away from legacy GV I/O recording workflows using CMF formats and towards an architecture based on MXF and Elastic Record X.
With this model, the ingest layer remains local while the resources required to run the recording services and connect to Framelight X can scale through distributed infrastructure. At the same time, the platform operates as a single production environment, with media asset management, playout, production tools and post-production capabilities all connected to the same core storage. Content doesn’t need to be moved between systems, and is accessible from any tool, even during ingest, across 22 supported languages.
This simplifies operations and support, while ensuring content is immediately available to editorial teams. It also reduces duplication across systems and helps teams deliver more content, more efficiently. This allowed us to add 80 additional recording ports to cope with the recording peaks during the Games.
Alongside ingest, we also identified specific requirements around playout and commercial break handling. Working with Grass Valley, we developed dedicated user interfaces within Playout X to support live operations and simplify how commercial breaks are managed. One of these is the Sports UI, which allows operators in the PCR gallery to manage commercial breaks directly during live production, without needing to be connected to the central playout system. This gives production teams more flexibility during live events while allowing central teams to remain focused on international and master playlists.
We also extended our approach to regionalization. Alongside a primary international feed, we operate multiple regional channels that can either follow the main feed and replace breaks with local content, or opt out entirely to deliver rights-specific programming. This allows us to maintain a single master playlist while supporting multiple localised outputs in a more efficient and scalable way.
At the same time we reviewed how operators interacted with the platform itself. Following the Paris 2024 Games, feedback from users highlighted the difficulty of locating specific clips within large volumes of recorded content.
When hundreds of feeds are recorded simultaneously, search and discovery become critical parts of the workflow. Together with Grass Valley we focused on improving Framelight X’s search capabilities and metadata handling within the system, and during Milano Cortina the difference was noticeable. Operators were able to locate relevant material much faster, which helped editorial teams publish content more efficiently across both digital and linear platforms.
Resilience and distributed production
Operational resilience is another essential requirement for large-scale live production. For the Winter Games, Paris and London served as our main operations centres for Olympic coverage. The system is designed so that either location can assume control if required; for instance, if an operational issue occurred in Paris, our teams in London could immediately manage recording operations through the same scheduling interface. And vice versa.
Rather than relying on disaster recovery processes that activate only during emergencies, the platform operates in a balanced mode where redundancy is already built into day-to-day operations.
The same infrastructure also supports more distributed production models. Across the Games, we delivered 42 linear channels in 22 languages, alongside more than 100 digital channels, supporting over 4,000 live events across linear and digital platforms.
For some workflows it is no longer necessary to move large numbers of people or equipment to every event location. Infrastructure can remain in central data centres, cloud environments or any hybrid, while production teams connect remotely. For the Milano Cortina Games, I was based in Paris rather than being on site in Italy. From a production perspective this approach allows us to coordinate resources more efficiently across multiple events.
Collaboration and continuous improvement
Technology provides the framework for these workflows, but large productions ultimately depend on the teams operating them. Our engineering and operations groups are stationed in Paris, London, the US and across Europe with each region handling its own dedicated editorial requirements, and collaboration between those teams is essential during major events.
Engineers also need to think from the perspective of operators and editorial teams. A workflow that appears simple from a technical point of view may become inefficient if it needs to be repeated many times during a production day. Feedback from operators after each major event helps guide improvements ahead of the next production cycle.
Looking ahead, our focus now moves to upcoming productions such as [the French Open Tennis Championship at] Roland Garros, while we continue to evolve the platform itself. One area we are exploring is the increased use of 1080p HDR as a mezzanine format for production workflows. We are also building additional media platforms for other parts of the Warner Bros. Discovery portfolio, including free-to-air channels in Nordic markets.
Each event adds experience that feeds into the next stage of development, and the AMPP-powered platform – driven by the innovative capabilities of the Framelight X and Playout X systems – continues to evolve alongside the production requirements it supports.
