70 terabytes. That’s how much image data of pollen and particles Professor Carsten Ambelas Skjøth has collected. He leads a project aimed at delivering automatic pollen measurements and forecasts for Denmark. Automatic instruments are now being used in Denmark, and these instruments continuously produce more data. Even with powerful computers, it takes many hours to fully analyze 70 terabytes of data.

That’s why the news that landed in Carsten Skjøth's inbox on Monday afternoon was incredibly welcome. He had received a positive response to his application to the Novo Nordisk Foundation for computing time on the new AI supercomputer, Gefion.

Gefion was launched in October 2024, with H.M. King Frederik X among those who symbolically helped power up the machine. The supercomputer gives Denmark renewed strength to accelerate innovation across a range of fields—from quantum computing and new drug development to solving societal challenges such as the green transition.

It’s no surprise that there has been great interest in accessing its technological muscle from the very beginning. Large volumes of data can be analyzed many times faster on Gefion than on regular computers, opening up entirely new and exciting possibilities.

For Carsten Skjøth, the task is to analyze and classify millions of images. Gefion will be used both to further develop the mathematical model used to analyze the images and to improve the quality of the underlying data that the model draws on.

More concretely, the opportunity to use Gefion will help accelerate the rollout of an automatic pollen network in Denmark, as the improved mathematical model will deliver real-time data to users.

In the application process to the Novo Nordisk Foundation, it was emphasized that applicants should be as ready as possible to begin using Gefion. And Carsten Skjøth is ready. Both the data and infrastructure are designed to be transferable to a supercomputer, and the underlying mathematical model is prepared to take advantage of a supercomputer like Gefion.

Carsten Skjøth says: “When the pollen project with the automatic instruments was first formulated, we knew that at some point we would run out of computing resources, so we prepared from the start for a move to a supercomputer. What we didn’t know was that the development of AI and artificial intelligence would move so fast and have such a big impact on both society and the research we’re doing in the pollen project. We also didn’t know that Denmark would get an AI supercomputer that is almost perfectly designed to analyze the vast amounts of data we produce.”

Carsten Skjøth is therefore thrilled about the opportunity to use Gefion and has high expectations for the supercomputer’s role in the future.