Global AI-driven memory chip shortage

Causes of the memory chip shortage

An AI infrastructure requires enormous amounts of memory. Processors and equipment that process AI require large amounts of advanced memory. Data centers are therefore being expanded on a massive scale.

Unfortunately, production capacity is limited. Worldwide, there are only three serious memory chip suppliers: Samsung, SK-Hynix, and Micron Technologies. They supply over 80% of all memory shipments. Moreover, chip production is only possible in dedicated factories. Scaling up production isn’t a matter of simply accelerating. Producing memory chips requires specialized machinery, which can’t simply be replicated. Moreover, we’re currently facing a difficult supply of raw materials.

Moreover, the few memory chip manufacturers that do exist are becoming choosy. They prefer to sell to AI-related customers, because that’s where they can achieve the highest margins. Other companies are left empty-handed; whoever pays the most gets the goods first. Companies are even bidding against each other. As a result, some products have already become 300-400% more expensive since last year.

We’re also seeing a shift in memory vendors’ production capacity toward AI-driven hardware. This is putting pressure on memory capacity (DRAM and NAND) and leading to consolidation among the major players. They’re increasingly pulling more of their own resources, leaving smaller players to wait.

Consequences for Enterprise Mobility

In the world of Enterprise Mobility, we work with network components like access points, routers, and switches, which require memory components. These devices are all affected by the increasing shortage of memory chips.

The impact of this AI-driven shortage on Enterprise Mobility and Wi-Fi networks is significant. For starters, we’re facing higher purchasing prices in our market, which are being passed on to customers. There’s simply no other option. The price increases that all manufacturers have already announced could reach 30% or more next year.

Delivery times are getting longer and longer. Chip manufacturers are implementing a strict policy: whoever pays the most gets it first. The delivery time is normally 4-8 weeks, but will likely soon be 12-16 weeks.

What does that mean for your business?

As a business owner, it’s wise to critically review potential replacement projects. Consider your budget and don’t expect to be up and running in just a few months, as everything you need has a long lead time. Share your forecast or roadmap with us so we can advise you and help you identify suitable alternatives or models. Be prepared for price indexation: products can suddenly become 25% more expensive.

Peak Technologies is here for you

As a system integrator, we believe it’s important to keep you informed in a timely manner about what’s happening and how you can prepare. Just as importantly, we can actively advise and assist you in finding and implementing solutions.

We monitor the supply chain, inform you about price changes, advise on lifecycles and roadmap planning, and help you find alternatives when a model needs to be replaced. Sometimes, bringing a project forward can actually save money. Do you have a project planned? Let us know. We’re here to help.