Memory shortages are challenging one of enterprise IT’s biggest assumptions

This article is part of our Opinions section, where we invite industry professionals to share their views on the most pressing technology questions of our time.

For years, enterprise IT has operated on the fairly comfortable assumption that when more infrastructure is needed, it can be bought. Servers, memory and other hardware components have been treated as readily available commodities, with predictable lead times, stable pricing and supply chains that, while occasionally disrupted, have largely been taken for granted. 

This assumption has quietly shaped everything from infrastructure refresh cycles to major transformation programmes. IT strategies have been built around future capacity requirements on the expectation that the technology needed to support them will be available at broadly predictable prices and within manageable lead times. 

But that assumption no longer holds true.  

Memory shortages, rising costs, manufacturing concentration, geopolitical uncertainty and surging demand driven by AI are all exposing just how dependent organisations have become on a small number of semiconductor manufacturers and just-in-time procurement models. 

And the global memory market is just the latest warning. Demand is rising, supply is constrained, and pricing is moving accordingly, but memory is not a niche concern buried somewhere in the semiconductor supply chain. It sits inside the servers, storage systems and devices that underpin modern organisations. When its availability changes, the economics and timelines of enterprise infrastructure change with it.

The question for IT leaders, therefore, is not: when will memory prices return to normal? It is whether the assumption of predictable infrastructure availability still makes any sense at all.

Memory is the warning, not the whole problem

It would be easy to dismiss today’s memory shortages as another semiconductor cycle. Technology markets have experienced shortages before. Supply tightens, prices rise, manufacturers invest, and eventually the market finds a new equilibrium. But waiting for the cycle to turn risks missing the much bigger issue.

Memory is just one example of the physical constraints influencing technology strategy. Semiconductor manufacturing remains highly concentrated. Global supply chains are exposed to geopolitical disruption. Power availability is influencing where data centre capacity can be built. Higher rack densities are creating new cooling challenges. Specialist engineering expertise itself is a finite resource. Individually, none of these challenges is new. What has changed is their impact on enterprise technology planning.

Historically, many of the physical constraints affecting infrastructure sat far enough down the supply chain that IT leaders rarely needed to consider them as part of wider business strategy. But that distance is narrowing.  A transformation programme can be strategically sound, fully funded and backed by the board, but if the underlying infrastructure is delayed, costs significantly more than forecast or relies on components in short supply, the outcome is the same: the programme stalls.

So what begins as an infrastructure availability issue can quickly become a business problem, delaying transformation, increasing costs and limiting an organisation’s ability to respond to growth. In this context, scarcity can no longer be treated solely as a procurement inconvenience. It needs to become part of infrastructure strategy, at board level.

The traditional refresh cycle deserves another look

Predictable hardware availability has also shaped how organisations manage infrastructure lifecycles. Hardware is purchased, operated for a defined period of time and replaced as it approaches an internal refresh date or an OEM support milestone. For many organisations, three-to-five-year refresh cycles have become an established part of infrastructure planning.

There are good reasons for this approach. Ageing hardware can introduce performance, security and resilience concerns, while newer infrastructure may provide greater capacity or efficiency. However, standardised refresh cycles also assume that replacement infrastructure will be available, affordable and necessary at exactly the point an existing asset reaches a predetermined date. In a more constrained market, that assumption requires greater scrutiny.

This is not an argument for retaining ageing infrastructure indefinitely. There will always be systems that need replacing because they no longer meet the performance, capacity or resilience requirements of the business. But age alone should not determine infrastructure strategy. The more important consideration is whether an asset remains fit for purpose. Is it performing reliably? Does it have sufficient capacity? Can it continue supporting the workloads the organisation requires? Can it be maintained effectively?

If the answer is yes, automatically replacing infrastructure because a standard refresh date has arrived may not always represent the best use of capital, particularly when new hardware is more expensive or difficult to source.

Equally, extending an asset’s lifecycle without understanding its condition and criticality creates a different form of risk. There is a clear distinction between deliberately retaining infrastructure that remains fit for purpose and simply postponing a necessary investment. The former can provide strategic flexibility. The latter creates technical debt.

Selective lifecycle extension can give organisations greater control over when they invest. Non-essential refreshes can be deferred, capital can be directed towards higher-priority systems, and IT teams are less exposed to sourcing replacement infrastructure at a specific point in the market cycle.

In an environment where availability is becoming less predictable, that flexibility has increasing value.

Support strategy is becoming part of infrastructure resilience

More flexible infrastructure lifecycles also require organisations to reconsider the relationship between vendor support milestones and the useful life of an asset.

End-of-service dates are often treated as fixed infrastructure deadlines. When an OEM changes or withdraws support, organisations can feel compelled to refresh equipment regardless of whether it continues to perform effectively. In a market where replacement infrastructure is readily available, this may be relatively straightforward to accommodate. But when components are constrained and costs are rising, it becomes a more significant strategic decision.

IT leaders need a clear understanding of the support options available across the entire infrastructure lifecycle. This means assessing where OEM support remains essential, where alternative support models may allow infrastructure to remain operational for longer and which assets genuinely need replacing. The objective should not be to avoid refreshes; it should be to maintain choice.

The cloud should not automatically be viewed as the answer to infrastructure scarcity either. Enterprise IT has spent much of the past two decades moving towards an on-demand mindset, with cloud computing reinforcing the idea that capacity is elastic and technology resources are infinitely scalable. But behind that apparent elasticity sits the same physical infrastructure: servers, processors, memory, storage and networking equipment. Cloud providers remain dependent on finite manufacturing capacity and global supply chains. Moving infrastructure to the cloud may create distance from the hardware, but it does not remove the physical constraints affecting its availability.

A flexible infrastructure strategy may involve refreshing an asset, extending its lifecycle, upgrading selected components, reallocating existing capacity or changing the way an environment is supported. Different systems will require different approaches depending on their criticality and the requirements of the business. This optionality becomes particularly important during periods of supply constraint.

Organisations that rely on a single vendor, support model or refresh timetable have fewer alternatives when availability changes. By contrast, those with a clear understanding of their infrastructure estate and the options available to them can make decisions based on business priorities rather than external deadlines.

Infrastructure resilience is therefore about more than keeping systems operational. It is also about retaining control over when and how technology decisions are made.

Treat infrastructure availability as a strategic risk

Traditionally, hardware availability has been treated as an operational concern. If a component is difficult to source, procurement looks for another supplier. If a delivery is delayed, the project timeline is adjusted. However, when infrastructure underpins customer services, revenue-generating operations and major transformation programmes, availability can have much wider consequences.

IT leaders need to understand where infrastructure scarcity would create the greatest business impact and ensure that exposure is visible beyond the technology function.

That starts with four questions:

1. Where are our critical infrastructure dependencies? Which systems, components or vendors would be hardest to replace if availability tightened?
2. Which refreshes are genuinely essential? Are lifecycle decisions based on performance and business requirements, or established policies and vendor dates?
3. Where do we have flexibility? Can assets be supported for longer, upgraded selectively or capacity reallocated if new infrastructure is delayed?
4. How far ahead can we see? Does infrastructure planning reflect the organisation’s three-to-five-year growth strategy, or only the next budget cycle?

The aim is not to eliminate uncertainty. That is unrealistic. Instead, organisations need to understand where they are exposed and create options before disruption occurs.

Memory shortages are an important warning because they challenge one of the assumptions that has shaped enterprise IT planning for years: that infrastructure will be available when the business needs it. Supply may improve, prices may stabilise, and new manufacturing capacity will eventually come online. However, the wider pressures affecting infrastructure are unlikely to disappear.

Infrastructure strategy now needs to be about more than deciding what to buy next. It must consider what can be optimised, what can remain in service, how assets will be supported and, crucially, where flexibility exists when Plan A is no longer available.

Because in a constrained technology landscape, the greatest infrastructure risk is not scarcity itself. It is building a strategy that assumes availability is guaranteed.

Related articles

• How your business can beat the PC memory crisis
• Rupert Bedell, Managing Director at Fasthosts: “IT teams will need to adopt a more customer-centric mindset”
• Beyond greenwashing: A tech leader’s guide to decoding ESG reports