RAM shortages force partners to rethink infrastructure projects: Research

Partners are adopting a range of mitigation efforts as memory shortages risk derailing customer projects.

Partners are bearing the brunt of RAM shortages to protect customers as supply chain constraints and price volatility threaten to derail projects, Westcon-Comstor research has found.

Memory shortages risk project delays created by need to redesign or revalidate solutions when components become unavailable or specifications change, adding time and complexity to project delivery.

Almost 90 percent of partners are adopting mitigation measures, according to the survey of its global Tech Xpert community, which include switching vendors or platforms, increased use of managed services and optimisation techniques to reduce reliance on memory-intensive workloads.

Patrick Aronson, APAC Executive VP and CMO, Westcon-Comstor, said partners are taking a pragmatic approach when customer outcomes are at risk.

“This isn't necessarily about replacing established vendors, but about building more resilient technology ecosystems and reducing dependency on a single source of supply,” Aronson said.

Should partners plan for ongoing hardware shortages?

The rapid uptake of AI is driving demand across the entire AI infrastructure stack, and that includes components, memory, processing capacity and specialised hardware.

Yet AI isn't the only factor, Aronson noted. Supply constraints, changing demand patterns, and broader economic and geopolitical uncertainty are creating a perfect storm of market volatility.

“We're seeing demand across the entire AI infrastructure stack, including advanced networking, high-performance storage, next-generation cybersecurity platforms and the underlying compute environments needed to support increasingly data-intensive workloads,” he said.

These conditions are expected remain, and Aronson’s advice to partners is to plan earlier, maintain flexibility and build contingency options into projects wherever possible.

“Strong collaboration with distributors and vendors is also becoming more important because visibility into supply trends and alternative solutions can help prevent disruptions before they impact customers,” he told CRN Australia.

Even if conditions change, Weston recommends partners position themselves to offer a hybrid approach, offering cloud, managed services and on-premises infrastructure to suit customer needs.

“The winners will be those partners who can guide customers across all of those environments and help them make informed choices, rather than assuming one model fits every situation,” he said.

Partners face the frontier challenge

One of the most significant challenges facing partners is the rapid emergence of new security risks associated with frontier AI models, according to Aronson.

“We're seeing models such as Anthropic's Mythos and other LLMs create and expose new attack vectors and infrastructure vulnerabilities that many organisations simply weren't planning for a year ago,” he said.

“As AI capabilities advance, customers are having to reassess the resilience of existing networks, security architectures and operational environments against a new generation of threats,” he added.

Partners are facing a difficult balancing act, with organisations being asked to invest in upgrading or refreshing legacy infrastructure to improve security and resilience, while simultaneously funding new AI initiatives. In many cases, those investments are competing for the same budget.

“They’re increasingly having conversations that weren't common 12 months ago: how to modernise legacy environments, mitigate emerging AI-related security risks, and build the infrastructure required to support AI adoption, all at the same time,” he said.

“We're also seeing the knock-on effect in project planning. Infrastructure decisions that were once relatively straightforward now require additional validation, testing and security assessment, increasing complexity across the project lifecycle,” he said.