NAB Show Perspectives: When the hardware market changes, everything downstream follows

There is a conversation happening in broadcast engineering that is not on any conference agenda. While AI use cases or cloud-native workflows are being discussed at length, it’s random access memory (RAM) that is quietly dismantling assumptions the industry has held for a decade. As the industry heads toward NAB Show in April, that switch in hardware economics is starting to influence the infrastructure decisions being planned for the next cycle.

For years, video processing infrastructure followed a simple model. A small number of channels per server. Clean, isolated, easy to troubleshoot. When hardware was cheap and memory was cheaper, it worked. Operators could build predictable business cases, plan hardware refresh cycles with reasonable confidence, and sleep soundly knowing that the cost curve bent in their favor over time. But now, those assumptions are starting to break down.

AI infrastructure is changing the hardware market

The AI infrastructure buildout has begun to reshape the component market. Advanced dynamic RAM (DRAM) and high-performance storage are flowing toward AI deployments at a scale the industry has not seen before. This demand is affecting supply dynamics for other industries that rely on the same components. Pricing validity windows are shorter. Lead times are longer.

According to IDC, AI infrastructure is projected to reach $758 billion globally by 2029, reflecting sustained investment in AI servers and supporting infrastructure. This level of demand is increasingly shaping the broader hardware market that other industries depend on.

The pressure this creates is not abstract. If you are running low-density deployments, each underutilized server compounds the problem. You are not just paying for the channels running on that box. You are paying for idle memory, unused compute, excess power draw, rack space and cooling. Across large deployments, those inefficiencies can materially affect the economics of operating broadcast infrastructure.

The cost of under-utilized infrastructure

As a result, infrastructure efficiency is becoming a more strategic consideration.

The best hedge against hardware volatility is needing less hardware in the first place. But efficiency is only one side of the equation, and infrastructure teams are increasingly asking a related question: if hardware costs are rising and becoming less predictable, should buying servers still be the default model?

The cloud debate in broadcast infrastructure stalled for a simple reason. For steady workloads, running services on-premise was often cheaper. And when the numbers were close, they were rarely far enough apart to justify the disruption of changing the status quo. Operational inertia is a powerful force, and technology alone has never been enough to overcome it. What is changing now is the economics behind that calculation.

Advertisement

The impact of rising hardware costs is already visible across the infrastructure market. Rising component costs are affecting infrastructure providers across the market, but those increases tend to appear differently depending on operating scale. For infrastructure planners, this introduces a new set of considerations. Hardware purchases lock organizations into long refresh cycles, while cloud services allow capacity to scale more dynamically. For operators planning a hardware refresh, the impact can be immediate, while cloud platforms often absorb those changes more gradually.

Rethinking the default infrastructure model

This does not mean cloud is automatically the right answer. Nor does it mean on-premise is finished. What it does mean is that the default assumption — that new services should begin with dedicated hardware — is being properly challenged.  The most important change is less about technology itself and more about how infrastructure decisions are framed.

Teams responsible for deploying new services are no longer asking “why would we move to the cloud?” They are asking something simpler and more honest: “Why are we assuming we need to buy hardware?” That reframing allows new services to be evaluated on current economics rather than legacy assumptions. It also creates space for more flexible infrastructure strategies, where workloads can move between environments as requirements change.

The organizations that navigate this shift successfully will be those that recognize how the underlying economics have changed before the next hardware refresh forces the issue. AI infrastructure demand is unlikely to slow in the near term, and its influence on component markets will continue to shape the hardware landscape.

As operators gather at NAB Show this year, many of the infrastructure conversations will reflect that changing reality. For broadcast organizations planning their next infrastructure cycle, the real question may not be which architecture wins, but how quickly infrastructure decisions can adapt as the economics underneath them continue to change.

Paul O'Donovan, MediaKindPaul O'Donovan is the senior director of product management for MediaKind.