Frore Systems, a manufacturer of cooling solutions, has presented findings on reducing temperatures in next-generation AI accelerators. The company's white paper suggests that improvements across the entire cooling stack, including GPU packaging, thermal interface materials (TIMs), coldplates, and coolant temperatures, can increase token generation per watt by over 30%.
Based on an analytical thermal model, Frore Systems projects its LiquidJet coldplate technology can lower Nvidia Rubin GPU junction temperatures by up to 12°C. This temperature reduction is estimated to translate into a 10% to 25% improvement in tokens per Watt, with a 10°C reduction specifically linked to a 15% increase in token generation.
Modern AI accelerators, such as the upcoming Nvidia Rubin, can dissipate up to 2,400 W, leading to die temperatures that can exceed 95°C. While 95°C may not immediately impact silicon longevity, high temperatures significantly increase leakage current, which doubles approximately every 10°C. This also reduces transistor switching efficiency.
Elevated temperatures necessitate higher voltages to maintain clock speeds, which often triggers Dynamic Voltage and Frequency Scaling (DVFS) to reduce clocks to stay within thermal limits. This ultimately leads to decreased performance and lower token output. Therefore, better cooling directly correlates with higher performance and token generation in AI data centers.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's briefNew every morning, and the back catalogue is archived by date.
Frore Systems published a white paper suggesting its LiquidJet coldplate technology can lower Nvidia Rubin GPU junction temperatures by up to 12°C, potentially increasing token generation by 15% with a 10°C reduction. This development addresses the critical need for improved cooling in high-power AI accelerators to maximize performance and efficiency in data centers.