- Published on
Report: Amazon AWS Faces Data Center Power Bottleneck, AI Compute Racks Delayed
- Authors
- Name
- DeepHorus
Beijing time, October 19: foreign media outlet Business Inside reported that some Amazon Web Services (AWS) data centers are facing power supply capacity bottlenecks, causing the rack-level AI compute systems they ordered to be unable to be commissioned on schedule. The relevant equipment has therefore been forced to delay deployment. The news was disclosed on October 21.
In recent years, AWS has continued to expand its AI compute infrastructure. In August this year, AWS expected to receive racks worth about $3 billion per month. This scale shows strong demand for its AI compute deployment. Power supply capacity has become a key link constraining the commissioning of equipment.
People familiar with the matter revealed that some equipment that has already arrived can only be temporarily stored in temporary warehouses due to insufficient power supply. AWS is closely monitoring the power supply progress of data centers. The company is using a combination of various means to accelerate power readiness. The above situation means that there is a time gap between equipment arrival and power support.
AWS has also adjusted its "cash-on-delivery" convention. The company acquired ownership of racks at an early stage of the supply chain. This arrangement helps AWS ensure the supply of compute resources. But at the same time, it also means that it must bear the losses caused by delays in rack commissioning.
The power supply bottleneck directly affects the pace of commissioning rack-level AI compute systems. For AWS, equipment arriving early but being unable to be commissioned is directly related to insufficient power supply capacity. Reports show that power readiness progress has become an important variable in its compute expansion.
AWS is closely monitoring the power supply process and accelerating power preparation. Whether it can commission racks as planned in the future depends on improvement in the power supply capacity of data centers. The market will pay attention to the impact of this bottleneck on AI compute supply.
