Sheetak, a thermoelectric cooling manufacturer, is expanding production and hiring engineers to keep up with the demand for chip-scale cooling inside AI data centre optics
The demand traces to a heat problem that room-level and rack-level cooling was not designed to solve. It lives inside the optical components themselves, not the server rack or the room around them.
The lasers and optical engines inside network transceivers move data between racks of GPUs and work correctly only within a narrow temperature window. A distributed-feedback laser drifts on the order of 0.08 to 0.12nm for every degree Celsius of temperature change, enough to push a channel off its assigned wavelength and corrupt the link. As millions of these components are in data centres, each one needs localised cooling that a chilled-water loop or a rack fan cannot reach efficiently.
Sheetak’s µCENTUM™ micro-TEC line is designed with this in mind. The modules start at 3.6 x 4.7mm with a 1.45mm profile, mount directly onto transceiver and photonic packages, and hold temperature to within ±0.01°C under closed-loop control, with thermal time constraints measured in seconds for fast correction.
Sheetak also works with developers of optical transceivers and data centre optical infrastructure to design custom thermoelectric solutions matched to each customer’s package, power budget, and performance targets. The company runs thermoelectric materials, device engineering, and manufacturing in its Austin facility, a vertically integrated model that supports ITAR-aligned programs and customers’ domestic-sourcing requirements. The coolers are also applicable in aerospace, defence, medical devices, and industrial systems where an optical engine needs active thermal control.
Kurt Devlin, chief revenue officer at Sheetak, said, “Our customers can’t afford a thermal failure at the component level, so we work directly with their engineering teams to design and manufacture parts that fit their exact package and performance requirements. That kind of close partnership is why we’re growing here in Austin and bringing on more local engineers to keep up with demand.”
To meet the demand for AI infrastructure, Sheetak is adding engineers to its Austin headquarters across thermal design and manufacturing, expanding a US-based workforce that builds and ships thermoelectric coolers, generators, controllers, and integrated thermal assemblies into photonics, telecommunications, aerospace, defence, computing, and medical markets.
