Big Data and computing are driven by consumer growth in High Performance Computing (HPC)-as a service, cloud gaming, and ADAS deep learning. Servers are expected to grow with an expected CAGR of 8% through 2025 in order to support the world€™s computing needs. AI hyperscalers are expected to grow 50% through 2025, which have especially challenging power requirements trending up to 3500 Watts. Drawing over 3500W of power the on board real estate on CPU€™s decreases. This increases the current draw, but decreases the availability for Smart Power Stage devices which supply current to the hyperscaler, resulting in an interesting power management challenge. One of the solutions is to move towards more efficient technologies such as GaN in order to deliver less loss. Dynamic RDSON is a critical specification for GaN which can provide many unique test challenges. Once GaN technology stabilizes, the industry can transition to a more reliable power management solution to support AI-Hyperscalers long term.
