JetCool–Broadcom Team Up on Liquid Cooling for Next-Gen AI XPUs
JetCool, a Flex Inc. (NASDAQ: FLEX) company, has partnered with Broadcom Inc. (NASDAQ: AVGO) to deliver production‑ready liquid cooling solutions for next‑generation AI XPUs, targeting multi‑kilowatt‑class ASIC platforms in hyperscale data centers. Backed by Flex’s global manufacturing footprint, the collaboration is designed to move advanced thermal architectures from lab concepts to high‑volume deployment.
What the collaboration delivers
JetCool has developed a single‑phase direct‑to‑chip liquid cooling solution that integrates with Broadcom’s mechanical and thermal reference architecture, enabling sustained multi‑kilowatt ASIC operation at heat flux levels up to 4 W/mm² per device. By aligning silicon design, advanced packaging, mechanical integration and thermal engineering early, the companies aim to de‑risk performance, reliability and deployment timelines for AI systems.
The partnership combines JetCool’s direct‑to‑chip cold plate technology, Flex’s global manufacturing and integration capabilities, and Broadcom’s custom AI silicon expertise to create a scalable thermal foundation for hyperscale AI infrastructure. JetCool’s offering spans cold plates, manifolds and coolant distribution units (CDUs), enabling end‑to‑end liquid cooling infrastructure for high‑density AI environments.
Strategic angle for FLEX and AVGO investors
For Flex, JetCool’s role in Broadcom’s AI XPU ecosystem strengthens its positioning as a manufacturing and systems integration partner in high‑growth AI infrastructure markets. For Broadcom partnering with JetCool supports its strategy to lead at hyperscale by tightly coordinating silicon architecture, power delivery and thermal design as AI power densities push higher.
The collaboration is explicitly geared toward supporting future generations of AI silicon as power densities continue to increase, signaling a multi‑cycle opportunity in advanced cooling rather than a one‑off design win. While execution and customer adoption remain key variables, the announcement reinforces the importance of thermal innovation as a gating factor for scaling next‑gen AI compute.
