Infiniflux

Infiniflux exists to make data centres more efficient than ever before. We do this with groundbreaking and patent-pending technology. We look at this not just in narrow terms such as PUE, but also in terms of Data Centre Functional Efficiency (DCFE). After all, getting more computational capacity out of a server not only benefits the efficiency of the data centre, but even more its economics. Our system keeps the temperature difference between GPUs and our proprietary refrigerant smaller than any other system. For instance, for 1000W chips it’s within 10˚C. Not by running our refrigerant at a high temperature, but by keeping the chip temperature below 30˚C. The computational capacity of the chip is determined by heat spikes at hot spots on the chip. The latent heat buffer in our 2-phase D2C liquid cooling system removes these heat spikes, allowing GPUs to perform at their maximum level all the time. This can give you up to 40% extra compute capacity. By removing these heat peaks and maintaining unprecedented low chip temperatures, our system extends the lifespan of AI and HPC chips by at least a factor of 8X compared with the best-in-class 1-ph D2C water cooling systems. This will also significantly increase the 2nd life value of servers. With our 1000W/cm2 heat removal capacity we enable the next generation of high-density computing. In our cooling system temperature variations are not only kept to as little as 3˚C in the secondary but also in the primary cooling loop. Furthermore, the pressure drop in our primary cooling loop is so much lower than in other D2C cooling systems that the required pump input power is 5X – 21X lower! All of this translates into at least 30% lower cooling costs. Our technology also reduces the water consumption of data centres with more than 80% and our proprietary refrigerant has a GWP of zero and contains no environmentally dangerous substances such as PFAS. Liquid cooling by Infiniflux: clean, green and best for your bottom-line.