DigiFarm is one of Norway's leading agri-tech startups specializing in the development of deep neural networks for automatic detection of field boundaries and sown areas based on high-resolution satellite data.
This is primarily used by their customers for verification of subsidies, crop forecasts, variable fertilization/spraying, carbon mapping, sustainability and traceability verification, grain species detection and regenerative agriculture.
"We are dependent on a technological partner who can assist and support us through this process. We could not find a better partner than the NCC, who has been the optimal partner for us.
This makes us more competitive in international environments."
Nils Helset, CEO and Founder, DigiFarm
DigiFarm´s core operation consists of the development of technology and model training of deep neural networks that can automatically detect field boundaries, sown area, and other objects within land changes such as trees, bogs, shade (from trees), telephone poles, and irrigation divots. The models are trained in several international regions including Europe, South America, APAC (Thailand, Vietnam, and India) as well as the USA and Canada.
This requires a large amount of image data as well as annotation data for training models as they process from over 300 million hectares of satellite data with 1 m resolution every month from 3 different dates of data, as well as to detect 30 million hectares with land changes.
According to CEO and Founder, Nils Helset, this requires a stable, secure, and high-performance HPC infrastructure with GPUs. DigiFarm has been part of several cloud-based startup programs including Google Cloud for Startups (A100 alpha program), AWS Activate, AWS Open Earth Initiative, Microsoft for Startups, NVIDIA Inception, and Oracle for Startups, but each of these platforms and programs has several challenges:
GPU costs: The cost of using GPUs for the development work has risen to an average of $2.75 per hour, particularly for NVIDIA V100 and P100 GPUs. This cost has become increasingly burdensome for the operations. Compounding this issue is the limited availability of certain GPU types through free cloud credits and startup programs. For example, in Azure's Microsoft for Startup program, you can only access K80 GPUs. Meanwhile, in AWS, Google, and Oracle, the A100 GPU is not an option, leaving us with only V100 and P100 GPUs to choose from. This situation emphasises the cost and availability challenges DigiFarm faces when utilising GPUs for development across different cloud platforms.
Workshop and technical support: due to a complex computer process infrastructure with many variables and different elements, it is particularly useful to get help with the optimisation of processes.
DigiFarm strongly believes in having a national network (HPC) and program for artificial intelligence (AI) startups that will contribute to a stronger ecosystem in Norway for the development of new and sustainable technologies.
Dette prosjektet mottar finansiering fra Forskningsrådet i Norge og European High-Performance Computing Joint Undertaking (JU) under tilskuddsavtale nr. 101101903. JU mottar støtte fra Digital Europe Programme og Tyskland, Bulgaria, Østerrike, Kroatia, Kypros, Tsjekkia, Danmark, Estland, Finland, Hellas, Ungarn, Irland, Italia, Litauen, Latvia, Polen, Portugal, Romania, Slovenia, Spania, Sverige, Frankrike, Nederland, Belgia, Luxembourg, Slovakia, Norge, Tyrkia, Republikken Nord-Makedonia, Island, Montenegro, Serbia.