Reliability Evaluation of Advanced GaN HEMTs Under Al Data Center Power Applications
Gallium nitride (GaN) high electron mobility transistors (HEMTs) offer significant performance advantages for power delivery in AI data centers through improved efficiency and increased power density. As an emerging technology, the stability and reliability of GaN HEMTs under AI data center mission profiles remain critical considerations for adoption. In this work, a test-to-fail methodology is employed to assess the dominant intrinsic wearout mechanisms in GaN HEMTs under representative mission profiles in AI Data Center Power Applications. Based on the identified wearout mechanisms, physics-based lifetime models are developed to project failure rates over a 10-year target lifetime. The combined lifetime projections indicate that GaN HEMTs are a robust solution for AI data center applications.
Authors
- Shengke Zhang (ORCID: https://orcid.org/0000-0003-3846-1665)
- Siddhesh G. Gajare (ORCID: https://orcid.org/0000-0002-5666-8371)
- Ángel Espinoza (ORCID: https://orcid.org/0009-0002-0489-485X)
- Ricardo Garcia (ORCID: https://orcid.org/0009-0008-9472-4144)
Publication Details
- Journal
- IEEE Power Electronics Magazine
- Published
- 2026-11-20
- DOI
- https://doi.org/10.1109/mpel.2026.3733195
- Primary Topic
- GaN-based semiconductor devices and materials
- Type
- article
- Field-Weighted Citation Impact
- 0.00