Image source: Shutterstock

Highlights

  • Collaboration focuses on integrating Ferric IVRs into Marvell’s custom silicon platforms.
  • Ferric IVRs are over 10x smaller than conventional power regulation solutions.
  • Aimed at improving efficiency, scalability, and ROI for AI and cloud infrastructure.

Ferric, Inc., a company focused on advanced power conversion technologies, has announced a collaboration with Marvell Technology, Inc. (NASDAQ: MRVL) to co-develop integrated power solutions tailored for Marvell’s custom silicon platforms. The partnership targets efficiency and performance improvements in AI and cloud infrastructure by utilizing Ferric’s miniaturized integrated voltage regulators (IVRs).

Ferric’s proprietary thin-film magnetic technology enables the creation of high-density DC-DC power converters. These IVRs are more than ten times smaller than traditional power management units, making them suitable for in-package integration within advanced processors. This proximity reduces energy transmission losses and enhances current density, addressing key performance bottlenecks in high-power compute environments.

As hyperscale data centers demand more efficient system-level power delivery for multi-kilowatt platforms, this collaboration aims to streamline IVR integration for Marvell’s custom XPU customers. The joint solution is expected to improve return on investment by delivering optimized power performance and reducing total cost of ownership for cloud operators.

Noah Sturcken, CEO of Ferric, emphasized the importance of the partnership: “Our collaboration with Marvell enables a streamlined path to bring our integrated IVR technology into the most demanding processor environments.”

Will Chu, Senior Vice President at Marvell, noted the relevance of under- or near-package power integration. He stated that this approach enhances performance and efficiency while offering greater flexibility to cloud operators.

The partnership reflects a broader industry trend toward integrating power solutions directly into processor packaging to meet the demands of next-generation AI workloads and scalable cloud deployments. Both companies aim to address growing energy efficiency requirements and infrastructure performance challenges through this technical collaboration.