SHI International: RAM Shortages and the Supply Chain Impact

As the world's supply chains are navigating ongoing cyber threats, leaders are needing to optimise their software in order to ensure simultaneous growth and resilience.

SHI International Corp connects teams with IT solutions and services needed grow operations and make impactful decisions.

Cheryl Slater is the Director for End User Compute in the UK at SHI, overseeing strategic partner relationships and ensuring alignment between vendor priorities and business goals. 

She spoke to Supply Chain Digital about the RAM shortage taking place and what that means for global supply chains.

Can you talk a little about SHI and what it does? 

We’re a global IT solutions provider, helping our customers select, source and manage IT. One way to think of us is that we combine the scale and portfolio of a global systems integrator with the resources and personal touch of a local VAR as we have more than 30 offices around the world. We also have integration centres located across the major continents, including the Nexus Integration Centre in Barnsley which is our centre of excellence for data centre hardware configuration, server racking and end user computing set-ups here in the UK. Our integration centres, AI and Cyber Lab, Technology Centres and Customer Innovation Centres allow our customers to plan, test, configure and deploy technologies, as well as explore and try out ideas, de-risking the process and ensuring that they maximise productivity and ROI.

What is RAM and why is it vital to the supply chain?

Random Access Memory (RAM) is a catch-all term used to describe the component that acts as the temporary memory of a computing device. It stores the data and code of applications that the Central Processing Unit (CPU) then accesses and utilises, speeding up processing time. There are different types of RAM, such as Dynamic RAM (DRAM) used in PCs, laptops and smartphones, but it is essential for modern computing.

The problem we have today is that RAM prices are rocketing and organisations are struggling to source these devices, stalling the market. It is also impacting data centre capacity and costs. IT and procurement leaders are now having to throw away the rulebook when it comes to supply and demand and will need to adopt some new strategies to deal with these shortages, such as inventory hold and lifecycle planning.

What is causing the RAM shortages and how long will these last?

As AI demand accelerates, suppliers increasingly select the most profitable segments. This is leading to a shift of manufacturing resources away from conventional DRAM used in devices to HBM used in AI accelerators. HBM consumes almost three times the wafer capacity of DDR5, which worsens scarcity.

But to understand how we got to this point, we need to take a closer look at the supply chain itself. SK hynix, Micron, and Samsung control the majority of global DRAM production. These manufacturers (fabricators or ‘fabs’) produce memory wafers, which are cut into dies and either sold to independent module manufacturers or used internally to produce SDRAM products sold to server Original Design Manufacturers (ODMs), Original Equipment Manufacturers (OEMs) and hyperscalers. Wafer reallocation to AI workloads has reduced the supply of traditional DRAM dies for server Registered Dual In-Line Memory Modules (RDIMMs). In fact, back in October, SK hynix announced that it had secured demand for all of its RAM production for 2026. The situation was further compounded by OpenAI announcing its intention to invest billions in data centre sites; plans that would see unprecedented demand for HBM and wafer demands. That prompted hyperscalers to pull their purchases forward and stockpile DRAM inventory amid fears that waiting would see them face a constrained supply later on.

How long will it last? The jury’s still out on that one. Conservative estimates say until early 2027 but it really depends on a number of factors. Wafer reallocation to HBM has reduced DRAM availability. Even as hyperscaler demand eases, there will likely be lagging effects on pricing as fabs rebalance production. We may see the supply chain attempt to correct the issue, with manufacturers being barred from stockpiling and new fab production facilities coming online. Equally, we could see AI demand fall off if those datacentres do not materialise due to planning objections or concerns over whether the grid will be able to supply sufficient power to those sites. That’s the optimistic outlook.

How are businesses reacting? Are we seeing panic buying or are they stalling for time?

Waiting for prices to normalise is a risky gamble. Most of the customers we are talking to are either making an initial upfront purchase or they are shutting off workloads in the data centre to give them the capacity to meet their plans for the year. We’ve actively discouraged panic buying because it could see them buy capacity at its peak and saddle them with stock but it’s also bad for the market in that it increases demand.

If IT teams improve lifecycle management and redesign data centre architectures to use memory more efficiently, demand will reduce, possibly permanently. So, there’s a real opportunity here to strategise and reform the way in which the business procures those devices or utilises capacity. It is all about prioritising device refresh and optimising those server configurations so that they use memory more efficiently, which is good for everyone.

What opportunities do the shortages create for channel players?

It is an opportunity to reshape the market for the better. That might be by developing intelligent refresh programs (IRPs) to help their customers navigate a more mixed OEM environment. Using an IRP allow the business to share information on refresh timelines, supply risks and budget triggers as well as a runbook on what to accelerate, what to delay and what to standardise, easing pressure on procurement.

Or it might be those players who specialise in IT Asset Management (ITAM), Software Asset Management (SAM) or FinOps help their customers explore cost reduction opportunities, the rationalisation of applications, or right-size licensing. Then there’s also the option to extract, resell or redeploy memory from end-of-life devices via asset recovery drives. So there are plenty of ways for channel players to lend their expertise.

Will this fundamentally reshape the supply chain?

Undeniably. The shocks felt through the supply chain have undermined some fundamental principles such as Moore’s Law. Nothing is now certain which is why we’re likely to see organisations move towards a form of supply chain resilience, whereby they use IRP to move away from fixed refresh times/rates to a more dynamic and continuous form of procurement that will see the business is less exposed during such periods of volatility.