Your AI bottleneck isn’t compute. It’s the file system
When AI platforms underperform in production, most teams reach for more GPUs. But the real bottleneck is often the storage...
We are here to help
Have a question or need guidance? Whether you’re searching for resources or want to connect with an expert, we’ve got you covered. Use the search bar on the right to find what you need.
I have been going to NAB for a long time. Long enough to know that the feeling of a show is often more telling than the official numbers.
The DPP published a detailed analysis of NAB 2026 this week, and a lot of it matched what I observed: quieter in places, footfall uneven, and fewer international faces than I have seen in previous years. All of that is fair. But here is where my experience diverges from the broader narrative: for us, it was our best NAB yet.
Our booth was not packed wall to wall, but it was busy in the way that really matters. Steady, purposeful, and full of conversations that are going somewhere. The people who found us knew why they were there, and that made a real difference to the quality of what we discussed: no tyre-kicking, just people with genuine problems looking for genuine answers.
I spoke at the BEIT Conference on Sunday morning, which is not the easiest slot to fill – 9:30 AM, the first day of the show – but the room was engaged, and the discussion that followed was exactly the kind I find most useful: technical, specific, and a little challenging in the best sense.
The session was built around a question I think about a lot: why are facilities spending serious money on storage hardware and high-speed networks, then watching that investment get strangled by protocol software that was never designed for what they are asking it to do today? It is not a new question, but it is one that the industry is finally taking seriously.
IAMT came to the booth and interviewed us during the show. They made a point of saying how clear our stand graphics were, and how easy it was to understand exactly who we are and what we do. Honestly, that feedback meant as much to me as any of the technical conversations. We have spent a lot of time thinking about how to explain this stuff without hiding behind complexity.
The performance numbers we were showing this year are worth sitting with for a moment. At NAB 2025, we showed 5.3 GB/s to a Mac client. This year: 10.5 GB/s, achieved on a Mac Pro with ATTO hardware on a single 100GbE port. Windows SMB Direct is now scaling beyond 45 GB/s. Linux is hitting 48.8 GB/s at 400GbE. Fusion NFS is scaling past 40 GB/s. A year ago, some of those figures would have seemed implausible. That is the whole point.
What struck me most was the shift in how people were asking questions. A year ago, there was a lot of curiosity. This year, people wanted to know about timelines, about testing, about what it would take to get Fusion into their environment.
We left Las Vegas with a lot of follow-ups to do. It’s a good problem to have!
Ready to talk timelines, testing, or what it would take to get Fusion running in your environment?
Get in touch with me and the Fusion teamSuggested content for: