How This Startup’s Metal Stacks Could Help Solve AI’s Massive Heat Problem
This startup’s metal stacks could help solve AI’s massive heat problem as next-gen GPUs demand extreme energy. With Nvidia’s upcoming Rubin Ultra chips projected to draw up to 600 kW per rack by 2027, data centers are searching for advanced cooling solutions. U.S. startup Alloy Enterprises is pioneering additive-manufactured copper plates to cool GPUs and surrounding hardware more efficiently—tackling heat from memory, networking, and other chips often ignored by traditional cooling systems.
Image Credits:Arctic-Images / Getty Images
How This Startup’s Metal Stacks Could Help Solve AI’s Massive Heat Problem
Alloy Enterprises produces solid copper cold plates designed to fit into tight spaces inside AI servers. Unlike traditional cold plates, these metal stacks are built layer-by-layer, allowing channel designs that improve heat transfer and withstand the high pressures required for liquid cooling.
With AI racks now jumping from 120 kW to 480 kW and soon 600 kW, legacy cooling cannot keep up. Peripheral chips—responsible for nearly 20% of cooling load—are no longer optional. Alloy’s metal stacks aim to solve this gap for GPUs, memory, and networking hardware, helping data centers stay efficient at massive scale.
Why AI Data Centers Need Metal Stacks Cooling
The explosive growth of AI models means GPUs are getting hotter and more numerous. Liquid cooling is now essential, and Alloy’s metal stacks are engineered to survive high-pressure coolant environments that traditional designs struggle with. Their slim form factor makes them easier to integrate into dense hardware layouts, improving thermal efficiency without major infrastructure changes.
Post a Comment