BELLEVUE, Wash.
GemaTEG™ introduces DaTEG 1.0, a thermal management solution that significantly reduces energy costs and enhances chip performance in AI infrastructures.
Why It Matters
- As the AI industry’s appetite for computing power grows, so does the strain on data center infrastructure.
- Traditional cooling systems are energy-intensive and often lack efficiency.
- GemaTEG’s DaTEG 1.0 helps top chip designers in the HPC industry, hyper-scalers, data center providers, and systems integrators transcend AI industry limitations by enhancing capacity without increasing resource demands.
What’s new
- DaTEG 1.0 is an integrated and modular Thermal Management solution that takes far less server space than traditional, less efficient cooling technologies.
- The breakthrough lies in its ability to maintain peak operational conditions for GPUs and CPUs, thereby extending their longevity and enhancing their output.
Deep Background
- Developed by astroparticle physicist Maurizio Miozza (Co-Founder and CTO) from decades of experience in silicon properties, DaTEG 1.0 results from over five years of research in materials science and fluid dynamics.
- The technology incorporates advances from aerospace to nuclear science, providing an integrated cooling solution.
By the Numbers
- DaTEG 1.0 supports chips consuming up to 500 watts, with future enhancements to accommodate chip designs of up to 1000 watts.
- Testing shows potential performance reaches advertised GHz levels without the risk of overheating—critical for maintaining uptime for AI in data centers.
Expert Voice
- From Maurizio Miozza: “With DaTEG, we’re not just preventing chips from overheating; we’re optimizing their operational conditions to push the boundaries of what’s possible in AI computing.”
What’s Next
- GemaTEG is collaborating with global academic and industry leaders to refine and expand the capabilities of DaTEG, ensuring it meets the future needs of AI infrastructure.
