We offer the latest AI accelerators from NVIDIA, Intel and Cerebras Systems – we strive to match your specific AI compute need with the optimal compute solution.
Sustainable Computing – Thanks to renewable energy sources and ingenious technology that takes advantage of the heat generated with, we are excited to offer you a CO2 -negative AI cloud service.
We offer the lowest rates on the market. No transfer costs, no extra hidden fees. Fully transparent and predictable monthly pricing. Contact us for your quote!
Green AI Cloud is the first player in Europe to provide AI cloud compute power with the groundbreaking Cerebras CS-2 system. This system features the world’s largest and fastest AI processor – WSE-2, Wafer-Scale Engine.
With 850,000 cores and extreme bandwidth memory, WSE-2 offers extreme speed and capacity within one single processor.
This heavily reduces the initial programming work and enables easy integration with your existing work flow.
Green AI Cloud’s solution is suitable for basically all kinds of AI and ML applications, data science and simulation workloads.
Create your account and subscribe on the one and only 100% fossil-free, 100% EU-compliant AI cloud compute service today.
Our localization in Sweden means that your company will avoid sending valuable and sensitive data outside the EU, and your operations will be completely in line with Schrems II, the new guidelines from the EU regarding data security. At present, all other AI cloud services are based outside the internal European market. Green AI Cloud is therefore the only player in Europe that guarantees compliance with the EU directives on data storage and data security within the internal market. We make it possible for large Nordic and European companies to live up to these directives and ensure a high level of data storage security.
The challenge for the digital medical industry in relation to computer vision
Today’s medical sensors generate imagery with resolutions at an amazing level.
However, the traditional GPU-based accelerators can process only a fraction of the imagery, due to the limited memory of GPU-based compute. At the moment, there are no apparent solutions to this problem. Maneuvers like downsampling and patching in order to fit samples on device substantially reduces the visual context available to an AI model to make predictions, which radically reduces the accuracy and leads to serious risks.
We are constantly affected by the movement of fluids – they are all around us all the time, regardless of what we do. Water and other fluids are constantly affecting our lives.
However, it’s very difficult to simulate how fluids act and move, due to the complexity of the process as well as the massive amount of compute needed for this type of simulation.
But Green AI Cloud’s Super Compute Cloud platform, which uses the CS-2 WSE, has the potential to change this. The platform makes it possible to simulate how fluids act, at a much greater scale, with enormous gains to be made in a large number of areas.
The challenge for the digital medical industry in relation to computer vision
Today’s medical sensors generate imagery at resolutions beyond what can be processed by traditional ML algorithms.
For example, whole-slide-images (WSI) are often 50k x 50k or beyond, i.e. significantly larger than natural images from datasets like ImageNet.
However, the traditional GPU-based accelerators can process only a fraction of the imagery, due to the limited memory of GPU-based compute. At the moment, there are no apparent solutions to this problem. Maneuvers like downsampling and patching in order to fit samples on device substantially reduces the visual context available to an AI model to make predictions, which radically reduces the accuracy and leads to serious risks.
Now there is a very exciting new solution that will improve the end result dramatically
Cerebras trained a UNet model with 28 layers on 5k x 5k imagery, successfully achieving a near state-of-the-art mIOU on the eval set of the Inria Aerial Imagery Dataset, without any hyperparameter tuning, network architecture modifications, or other tricks. Among the many areas where the benefits are obvious, are medical institutions operating with large-volume segmentation in large data types, such as MRI and EM scanning.
Medicine – radiology and pathology
Examples: X-Rays, CT-Scans, MRI, retinal imagery, elector microscopy, PET in knees, lungs, soft-tissues, bones, brains, retina, skin
Geographic Information Systems (GIS)
Examples: SAR, infrared, optical, aerial
(30cm -> 3k x 3k; 10cm -> ~10k x 10k)
Remote Sensing and Earth Observation
Agriculture
Manufacturing
Monthly | Quarterly | Yearly Pricing