
Don't let the heat slow your chips down
AI processor chips (GPUs) are becoming ever more powerful, constantly driving up their Thermal Design Power. This makes it ever more challenging to keep these processor chips within the recommended temperature range. Super fast heat removal avoids overheating of these precious chips. Until now single-phase Direct Liquid Cooling has done a decent job to achieve adequate cooling of these processor chips but can’t keep up with where chip design is now going. Infiniflux™ achieves this with its self-contained, maintenance-free 2-phase Direct Liquid Cooling system that makes your GPUs last 3x longer and work 15% faster.

ChipCool™: the clean solution to a growing problem
ChipCool™: the clean solution to a growing problem




The rapid expansion of AI is driving an unprecedented surge in data‑centre resource use. Global electricity demand from AI‑specific facilities is projected to soar from around 50 billion kWh in 2023 to over 550 billion kWh by 2030, contributing to a doubling of total data‑centre energy consumption over the decade. About 40% of this energy consumption is used for cooling. Cooling these high‑density systems will also require vast amounts of water, with usage forecast to almost quadruple to 664 billion litres in the same period. A growing number of data centres are being built in water-stressed regions.
It doesn’t have to be that way, and it shouldn’t! With ChipCool of Infiniflux much of these increases can be avoided. The time to act is now, to avoid locking in even more environmental and grid‑capacity pressures. By using ChipCool not only in new data centres but also retrofitting it in existing ones, the overall energy and water consumption of data centres can be stabilised whilst continuing the expansion of data centre capacity.
Our Mission
Infiniflux exists to make AI and HPC data centres much more resource efficient, cutting energy use by a third and largely eliminating water use.
More value for Money
ChipCool™ keeps your chips 10-15ºC cooler, boosting IT workload capacity and prolonging chip lifespan
2-phase cooling without PFAS and with Zero Global Warming Potential and Zero Ozone Depleting Potential
Planet Friendly Solution
Water-free refrigerant and passive flow mean zero maintenance, giving you complete peace of mind
Zero Maintenance


We measure our success in more than PUE
We are fanatical about resource efficiency so PUE is a very important metric. Our PUE is less than 1.03, but that is only half the story. We affect not only what’s below the line, but also what is above the line. By keeping GPUs cooler, we reduce the amount of throttling, leading to a higher IT workload capacity. That is extra revenue generated with the same cost base. A lower die temperature also extends the lifespan of GPUs, which are by far the most expensive components in AI data centres. This translates into a significantly lower depreciation rate, boosting data centre profitability.
For us the key metric is Performance Per Watt (PPW), which is measured in TFLOPS/Watt.
Supported by
We wish to acknowledge the support of the Henry Royce Institute for Advanced Materials through the Industrial Collaboration Programme, funded from a grant provided by the Engineering and Physical Sciences Research Council EP/X527257/1.
We are grateful for the support of the UK’s innovation agency, Innovate UK.
Explore our Thinking
This will link a couple of more recent articles or blog posts
Energy Efficiency
Our technology significantly reduces energy consumption in data centers while enhancing thermal design options for GPUs.
Sustainable Design
We provide cutting-edge solutions that minimize water usage in high-performance computing environments, promoting sustainability.