We started as GPU based blockchain enthusiasts and grew to a HPC cloud provider!
We have been working on professionalising HPC infrastructure since the beginning of
2016. Our most important ambition is to bring HPC tasks to everyone.
Our enthusiasm is what helped to create BlockTechSystems.
Our expertise lies within building, operating and orchestrating large scale
GPU clusters, aiming to be the worlds most energy-efficient
and cost effective. We recently set up a large scale GPU cluster
of 46.000 GPUs in Norway.
BlockTechSystems is one of the largest HPC operations in Europe.
We build and develop data centres for large scale GPU clusters. A large part of the energy required to operate IT infrastructures for "Big Data", "Deep Learning", "Neural Networks" or mining is used exclusively for cooling the IT components. Our solution is to convert the energy used for cooling into usable waste heat and return it to the community. This way we can create new and self-contained ecosystems. Therefore, we are rethinking every aspect of building GPU clusters, cooling infrastructure and datecentres in general.
A big problem in the data centre landscape is that local
infrastructures in the area of energy and bandwidth are
approaching the performance limits. It is now also no longer
possible to further expand the networks. This alone sometimes
increases planning times for larger data centres to more than
Datecentres are currently not build to accomodate the most advanced hardware, they are build to fascilitate as much traditional rack space as possible for colocation puposes. It leads to unflexible and only standardised datacentre designs. Our client is the IT hardware itself, so we are able to provide disruptive, purpose build datacentre designs, to bring cost-effective computational GPU power to the cloud.
We follow the approach of setting up our GPU clusters in places where energy is generated and cannot be used adequately.
There are numerous places in Europe – and worldwide – where immense amounts of energy are generated (e.g. wind farms in Northern Germany, biogas in rural areas, solar energy in the desert, hydropower in the mountains, etc.).
The problem is that the energy cannot always be transported to places where it is needed. We want to set up decentralised data centres in places where energy is generated but cannot be used efficiently – at the same time we want to supply the environment (residential, industrial areas, etc...) with our waste heat. This way we can make sure that an energy-intensive process, such as an HPC purpose, does not waste energy unnecessarily, but instead uses the energy that is available but cannot currently be consumed.
We are setting up decentralised HPC centres all over Europe.
We do not consume energy – we use energy in places where it is generated!
Our systems are easy to set up. All we need is access to water, electricity and a data line.
We are designing systems to minimize the energy used for cooling.
Through our energy efficient cooling techniques,
we eventually create the possibility to give back heat to the community.
We are developing data centre solutions for the future for "Deep Learning", "Rendering", "Neural Networks" and other complex application areas.