Dell has drastically shortened the delivery time of its AI servers. While we had to wait 39 weeks in December, with a bit of luck buyers can now get started after just eight weeks.

Dell shortens the delivery time of its AI servers by more than half a year. While we reported in December that AI servers were taking up to 39 weeks to ship due to Nvidia component shortages, the delivery time has now been reduced to eight to 12 weeks.

These faster deliveries are a direct result of a larger influx of Nvidia GPUs. Nvidia struggled with limited production capacity last year. Specifically, TSMC’s CoWoS service was at capacity and manufacturers like Nvidia had to look for alternative providers. TSMC said it could only fulfill 80 percent of orders placed. The manufacturer itself has now expanded its production capacities.

Good for everyone

The situation isn’t just improving for Dell. Other manufacturers such as HPE can now also deliver at somewhat more realistic conditions. For HPE the current delivery time would be around twenty weeks. Given the delivery difficulties, all server specialists have a huge backlog of orders. This in itself is not a catastrophe for the providers involved, as GPU servers are expensive devices with the necessary margin.

For AI servers in general, there are now more and more options available without Nvidia. For example, servers with AMD Instinct MI300 are becoming increasingly important. Lenovo recently saw demand for this peak.