JetCool and Broadcom Innovate AI Cooling Solutions for Advanced Computing Infrastructures

JetCool and Broadcom Partner to Innovate AI Cooling Solutions

Flex’s subsidiary, JetCool, pioneers next-generation thermal management solutions as it collaborates with Broadcom to enhance liquid cooling systems for advanced AI computing infrastructures. Announced on March 11, 2026, this partnership addresses the challenges posed by increasing power densities in silicon devices, which are reportedly reaching multi-kilowatt ranges. By utilizing JetCool’s direct-to-chip liquid cooling technology, the collaboration aims to optimize performance and reliability in AI workloads, which require sophisticated thermal management to prevent overheating and ensure efficient operation.

JetCool's unique single-phase cooling solution is engineered to align seamlessly with Broadcom’s thermal architecture, effectively facilitating multi-kilowatt ASIC operations at heat flux levels of up to 4 W/mm² per device. This integrated approach highlights the significance of early-stage collaboration between silicon design, advanced packaging, and thermal engineering domains. By combining JetCool’s innovative cooling technology with Flex's robust global manufacturing capabilities, the partnership establishes a formidable platform for developing high-performance solutions vital for modern hyperscale AI infrastructures. As AI workloads continue to expand, such advancements are essential in pushing silicon technology towards sustainable thermal performance.

Key stakeholders underscore the potential impact of this collaboration on the AI sector. Ken Kutzler, VP of AI Systems Development at Broadcom, notes that a unified approach leveraging silicon architecture and thermal engineering accelerates the timeline from concept to market for AI XPUs. Meanwhile, Bernie Malouin, PhD and VP of Liquid Cooling at Flex, comments on JetCool’s capabilities to meet escalating thermal challenges faced by AI silicon designs. This partnership not only strengthens JetCool's role within Broadcom’s extensive network of partners but also drives the evolution of direct-to-chip thermal architectures, establishing a foundation for future generations of AI silicon technology.

In related news, Broadcom introduces a high-speed networking chip designed to enhance data movement within AI data centers, showcasing its commitment to addressing the growing demand for rapid data processing capabilities. This advancement is anticipated to significantly improve data flow and latency, crucial for the efficient handling of extensive datasets utilized in AI applications across various industries, including automotive and healthcare. As companies increasingly rely on advanced solutions for AI workloads, Broadcom's new technology positions them as a significant player in the burgeoning AI infrastructure market, reinforcing their commitment to innovation and competitiveness in the tech industry.