Brad Turner, CEO of EditShare, reflects on the conversations that stood out at IBC 2026, from practical AI and hybrid production to cost, security and the changing way teams access media.
One of the things I like most about IBC is that you very quickly get into real conversations with customers, partners and other vendors about what is actually changing in the industry.
You spend a few days talking to people, and you start to hear the same questions coming up. This year, a lot of those questions were about AI, but not in the way they were a year or two ago. People are still interested in what AI can do. What has changed is that they now want to know where it genuinely saves time, what it costs, where their content has to go, and whether it actually makes their workflow easier.
That is a much more practical conversation.
AI has to solve a real problem
At EditShare, we have taken a fairly practical approach to AI.
With FLOW AI, the aim is to help people get through huge volumes of media faster. If you have thousands of hours of content, you should be able to search transcripts, on-screen text, people, logos and scenes and get to the material you need without somebody having to manually tag everything first. Practical, cost controlled AI has a real business case: reduce search time and potentially monetize content better.
That is the kind of AI people understood immediately at IBC because the problem is obvious.
The on-prem processing part of that story also got a lot of attention. Not every organisation wants to send all of its media to the cloud for AI processing, particularly when you start talking about security, governance and potentially very large usage bills.
There was also a lot of talk at the show about what comes next, particularly agentic AI and natural-language interfaces.I think that is going to be interesting. We are going to see more people interacting with production systems by asking for what they need rather than using the native UI. Is a traditional asset management system dead? If you mean a traditional system with strict media governance, highly structured permission based workflows, and manual metadata: yes it’s been dead in all but a few markets for years. It lives on in parts of Europe and the Middle East, especially in government owned or sponsored broadcasters. But we are betting on the part of the market that wants fast and flexible workflows.
Hybrid storage and workflow is the new normal
The other thing that came up a lot was where content actually lives. For most media companies today, the answer is not one place. You might have production media on shared storage, older content in an archive, material in cloud or object storage, and people working remotely who still need access to all of it. That is normal now.
I don’t think the answer is to tell customers they have to move everything into one environment or have to index it all with a system of record. The better question is how you let people get to the content they need, wherever it happens to be. That is why I was pleased with the response to MediaSilo Media Drives at the show (and we won a TVB Europe “Best in Show” Award).
The idea behind it is pretty straightforward. You can browse, search and preview media on supported S3-compatible storage, and then bring selected content into MediaSilo or our on premise EFS system. Then do something with it, whether its edit the footage or send it somewhere for review or final delivery. AND, as I mentioned, we do it without forcing you to wait while we index the content and create proxies unlike other heavy weight asset management solutions.
You are moving what you need, when you need it, rather than copying everything around as a matter of course. That feels much closer to the way people really want to work in today’s world..
Cost is becoming a bigger part of the conversation
We also had a lot of conversations about overall cost and hardware cost in particular.
There is a huge amount of investment going into AI infrastructure right now, and that is putting pressure on the wider hardware market. At the same time, nobody in media is being given a blank cheque to keep adding servers, applications and infrastructure. Customers are looking pretty hard at what they actually need.
That thinking is behind the new all-in-one EFS configurations we showed at IBC. On EFS 310 and EFS 410, we can now bring the shared storage, FLOW and the GPU processing needed for FLOW AI into the same system.
For some customers, that means they don’t need another server to start using AI-assisted media discovery. For others, cloud will be the better answer.
There isn’t one architecture that’s right for everybody. What matters is giving people options and letting them make that decision based on the way they work and what makes financial sense.
Openness matters
Choice came up in a lot of conversations as well.
Media companies already have a lot invested in their editing systems, storage, cloud services and other production tools. They should not have to throw all of that away because one part of the workflow changes. That is why openness continues to matter to us.
EditShare has always had a well documented API to allow customers to integrate with our systems. We do have an integrated ecosystem with our ARK archive product (which some customers love, but not everyone). Recently we decided to embrace third party on premise archive products like Spectra, Diva, and Kumulate. We now support archive and restore workflows from within FLOW and can search these archives using FLOW AI.
What I came away thinking
IBC is always useful because you get a pretty good sense of which ideas are getting traction and which ones are mostly noise.
For me, the direction is becoming clearer. People want AI that does something useful. They want easier access to media without constantly moving and duplicating it. They want the freedom to choose between on-prem, cloud and hybrid depending on the job. And they want technology that works with the production environment they already have.
The opportunity for our industry is to keep adding capability while making it easier for people to put that capability to work.
Estimated reading time: 5 minutes
