4 surprising things we learned from IBM Research

While AI and quantum computing were the two big themes at this year's IBM Research Europe media event, Tim Danton reveals there was also one surprise lurking in the lab

-  Tim Danton

Perched on a hill overlooking Lake Zurich, there sits a group of inauspicious buildings. This is the campus for IBM Research Europe, and inside you'll bump into some of our continent's brightest minds. Some of their research areas are predictable: artificial intelligence, quantum computing, integrated circuits. But it's also working on areas that might surprise you.

I was there for the faculty's annual media event, titled “IBM: Shaping the Future of Computing". Based on briefings, demonstrations and visits to the labs, this article focuses on four initiatives of particular note.

Understanding the Earth

On 6 December, the day before the event, IBM announced that in partnership with NASA - it had released PrithviEO-2.0, an open-source geospatial foundation model. Such models are trained to recognise anything from cattle to water coverage to foliage; feed it satellite data and it will produce meaningful information without human analysis.

"The new model that we have is excelling at all these tasks to a point that [we can say] our models are the leading ones in this space," said Juan Bernabé-Moreno, director IBM Research Europe/UK & Ireland. "But what do we do with the models?"

The first example was all too topical, following the flooding in the Valencia region in late October 2024. "Here, the beauty of the model is it doesn't only work with Sentinel 2 data, so with images, it works also with radar data. The difference is with images only then when you have flooding, you have clouds, and if you have clouds, it's very difficult to understand the impact of the flooding. But... with radar data, you can look through the clouds and you see all the flooding."