Radixia

HPC and AI in Saudi Arabia: From Infrastructure to National Impact

SessionTuesday · 14:15–15:30 · Hall 4 - Ground Floor · ~1,459 words

Speakers: Carla Osthoff (National Laboratory of Scientific Computing Brazil)

Session summary

In this session, introduced by Carla Osthoff (National Laboratory of Scientific Computing, Brazil), Saber Feki provides an update on the status and national impact of HPC in Saudi Arabia, focusing on KAUST and partner institutions. He traces KAUST's HPC lineage from a 2009 IBM Blue Gene with 64,000 cores, through the Cray XC Shaheen II with 200,000 CPU cores, to the heterogeneous Shaheen III combining 880,000 AMD Genoa CPU cores with 2,800 Grace Hopper GPUs, noting each generation delivered top-20 Top500 rankings and roughly fivefold energy efficiency gains, with a nine-to-tenfold improvement moving from CPU to GPU. He also describes the smaller, agile IBEX cluster for rapidly evolving research needs. Scientific successes include an ACM Gordon Bell Prize in climate change modeling run on Shaheen III CPUs, an HPCwire award for a rice genome project scaling toward 20,000 genomes to support sustainable agriculture, and an oil-spill weather forecasting model built with Saudi Aramco. Early GPU projects span AI for materials science (solar panels and greenhouse coatings), fundamental AI model research led by Jurgen Schmidhuber, kilometer-scale climate modeling with ICON, and CFD acoustic simulations of urban cities. Feki surveys other national systems, including Saudi Aramco's Dammam 7 and Ghawar systems for seismic imaging plus a new $370 million project, the National Center of Meteorology's weather, dust, and flash-flood models, the well-funded Humain sovereign AI cloud initiative, and Tahakom's newly deployed public-safety system. He closes by announcing the Saudi HPC and AI Conference at KAUST.

Topics: hpc infrastructure · sovereign ai · heterogeneous computing · energy efficiency · genomics · climate modeling

AI-generated summary of an auto-generated transcript (~1,459 words in full). Details may be imprecise — verify against the session recording.

Auto-generated captions from ISC 2026 session recordings · transcription errors likely, verify quotes against the video · timestamps are offsets into each recording · independent tool, not affiliated with ISC · a Radixia Labs experiment