Wiwynn, a major server and data-center hardware maker, is telling clients to brace for a long haul. The company expects AI infrastructure bottlenecks to persist through 2028, warning that shortages of key components and assembly capacity won't ease anytime soon.
What Wiwynn is saying
The warning came directly from Wiwynn's leadership. They point to a supply chain that simply can't keep up with the explosion in AI compute demand. The company, which builds servers for some of the biggest cloud providers, sees no quick fix. The shortages, they say, will stretch for years — not months.
Why the bottlenecks are happening
Several factors are colliding. High-end GPUs, advanced memory, and specialized cooling systems are all in tight supply. Wiwynn's own production lines are running near capacity. On top of that, the lead times for critical components like power management chips and networking gear have stretched. Every new AI data center needs thousands of those parts. The industry just didn't build enough factories fast enough.
What this means for AI adoption
Companies racing to deploy large language models and other generative AI tools will face delays. New projects may get pushed back. Smaller firms without existing contracts with suppliers like Wiwynn could find themselves at the back of a very long queue. Even the big players — the hyperscalers — are having to plan their roadmaps years in advance. The shortage isn't just about chips; it's about the entire rack-level system.
The timeline ahead
Wiwynn's outlook runs through 2028. That's more than four years from now. In tech terms, that's an eternity. If the company is correct, AI infrastructure costs will stay high and expansion will be constrained. It also means that any new fabrication plants or assembly lines announced today won't come online soon enough to change the picture before then.
The big unanswered question is whether the bottlenecks will force AI users to shift to less demanding models — or whether the industry will find ways to squeeze more out of existing hardware. For now, Wiwynn's message is clear: plan for scarcity.
