TDK — leaderboard ()

Supermicro, Krambu and Endor team on liquid-cooled AI data centers with waste heat recovery

Krambu is collaborating with Endor Development and Supermicro on the development and deployment of what the companies describe as next-generation, sustainable “AI Factories,” with a focus on direct liquid cooling and waste-heat reuse.

The work is centered on integrating Supermicro’s AI server platforms with facility designs that incorporate “industrial symbiosis” concepts, including capturing waste heat from AI data centers and reusing it for applications such as agriculture, residential heating, and industrial processes.

Who’s doing what

Supermicro’s role covers “end-to-end AI server solutions,” including system design, architecture, integration, and delivery of fully configured, rack-level compute systems. Supermicro will also support testing and deployment, plus lifecycle services such as maintenance, upgrades, and technical support.

Krambu will act as channel partner and infrastructure integrator, coordinating procurement, logistics, and installation of AI systems. Krambu is also slated to lead the design and engineering of advanced cooling solutions—specifically direct liquid cooling (DLC-2)—and sustainable infrastructure systems that incorporate waste-heat recovery and energy reuse. The company will also support customer engagement and help facilitate relationships with “anchor tenants.”

Endor Development will lead facility design and operations work, including site development, permitting, and construction. Endor will also oversee integration of AI systems into the data center facilities, with the stated goal of building AI ecosystems that are “efficient” and “resilient.”

Why data center engineers should care

For high-density GPU deployments, thermal design and heat rejection often become the gating factors well before floor space does. Direct-to-chip liquid cooling can shift the heat-transfer problem from room air management to liquid loops and heat exchangers, which can open up higher rack densities—if the facility-side plumbing, redundancy approach, and maintenance model are engineered correctly.

But waste-heat reuse is where projects can get hard fast. The concept is straightforward; the execution hinges on matching temperature grade, seasonal demand, distribution distance, and who owns and operates the downstream heat network. Any “reuse” plan that can’t survive those realities ends up as a nice diagram instead of a reliable operating mode.

“This collaboration with Supermicro and Endor allows us to bring together best-in-class AI compute, advanced direct liquid cooling, and industrial symbiosis to redefine how data centers are built and operated,” said Steve Wood, CEO of Krambu.

Source: Krambu

Get Data Center Engineering News In Your Inbox:

By subscribing, you agree to our Privacy Policy and Terms of Use. You can unsubscribe at any time.

Popular Posts:

Why data center power validation is moving to real-time simulation
Why data center power validation is moving to real-time simulation
CPC-TRM-Tech-Guide-Liquid-Cooling-Reliability-(1)-4
In liquid cooling reliability, tubing choice creates another layer of protection
Warm-Water-Ready-Accelsius-Whitepaper-3
Accelsius two-phase cooling runs NVIDIA B200 9° C cooler than single-phase
Vertiv-Tognana-Innovation-Lab
Vertiv to double chiller production capacity for data center cooling by 2026
6a4f3014091bd43fb15eee94_6a4f2dee21d1dee3314b8959-1783574215391
Lightpath to build AI-grade fiber for two hyperscale data center campuses

Share Your Data Center Engineering News

Do you have a new product announcement, webinar, whitepaper, or article topic? 

Get Data Center Engineering News In Your Inbox:

By subscribing, you agree to our Privacy Policy and Terms of Use. You can unsubscribe at any time.