Sheetak expands Austin thermoelectric cooling for AI data center optics

Sheetak is scaling U.S. manufacturing and hiring engineers in Austin as demand rises for micro-coolers that keep optical transceivers in AI data centers within a tight temperature range. The push targets a component-level heat problem that traditional rack and room cooling cannot solve.

Why it matters: - AI data centers depend on optical transceivers to move data between GPU racks, and those components need precise thermal control to stay on wavelength and keep links stable. - Sheetak’s expansion adds U.S.-based capacity for chip-scale cooling at a time when domestic sourcing and specialized optics hardware are becoming more important. - The company is also growing local engineering jobs in Austin.

What happened: - Sheetak, Inc. is expanding production in Austin, Texas, and hiring engineers to meet demand for thermoelectric cooling inside AI data center optics. - The company designs and builds its thermoelectric devices in Austin. - AI infrastructure is now Sheetak’s fastest-growing business segment.

The details: - Sheetak’s µCENTUM micro-TEC line starts at 3.6 by 4.7 mm with a 1.45 mm profile. - The modules mount directly onto transceiver and photonic packages. - Closed-loop control holds temperature to within ±0.01°C. - Thermal time constants are measured in seconds, allowing fast correction. - The cooling targets the lasers and optical engines inside network transceivers, not the server rack or room environment. - A distributed-feedback laser can drift about 0.08 to 0.12 nm for every degree Celsius of temperature change. - That drift can push a channel off its assigned wavelength and corrupt the link. - Sheetak works with major developers of optical transceivers and data center optical infrastructure to design custom thermoelectric solutions. - Those designs are matched to each customer’s package, power budget and performance targets. - Sheetak runs thermoelectric materials, device engineering and manufacturing under one roof in Austin. - The vertically integrated model supports ITAR-aligned programs and domestic-sourcing requirements. - The same coolers are also used in aerospace, defense, medical devices and industrial systems. - The company also builds and ships thermoelectric coolers, generators, controllers and integrated thermal assemblies for photonics, telecommunications, aerospace, defense, computing and medical markets. - Sheetak says it has more than 25 issued patents. - The company says prototype lead times run three to six weeks. - More information is available in the company’s announcement.

Between the lines: - The expansion reflects a shift from facility-level cooling toward component-level thermal control as AI hardware density increases. - Sheetak’s focus on localized cooling suggests that optical performance is becoming a bottleneck in AI networking, not just compute power. - The Austin buildout also fits a broader push for U.S.-based manufacturing in advanced electronics supply chains.

What's next: - Sheetak is adding engineers at its Austin headquarters across thermal design and manufacturing. - The company is expected to keep scaling output as demand for AI data center optics grows. - Sheetak will continue serving photonics, telecommunications, aerospace, defense, computing and medical customers from its U.S. manufacturing base.

The bottom line: - Sheetak is betting that AI data centers will need precision micro-cooling close to the chip, and it is expanding Austin manufacturing to meet that need.