KAUST
academicThe King Abdullah University of Science and Technology (KAUST), founded in 2009, is Saudi Arabia's graduate research university on the Red Sea coast. Its AI Initiative is directed by Jürgen Schmidhuber (co-inventor of the LSTM), and its Center of Excellence for Generative AI (launched 2024, with Bernard Ghanem as chair and Schmidhuber as co-chair) is a hub-and-spoke program for building Saudi generative-AI capacity.
Flagship lines: Mohamed Elhoseiny's Vision-CAIR long-video vision-language model line, running from MiniGPT-4 (25.7K★, ICLR 2024) through Goldfish/MiniGPT4-video to Tempo-6B; Schmidhuber's Huxley-Gödel Machine, a self-improving coding agent that operationalizes his 2003 Gödel-machine idea; and the KAUST-led Exascale Climate Emulator, which won the ACM Gordon Bell Prize for Climate Modelling at SC24.
The Shaheen III supercomputer (2,800 NVIDIA GH200, #18 on the TOP500) is slated to train Arabic-centric language models from early 2026. Attribution boundaries matter here: the national Arabic LLM ALLaM belongs to SDAIA, LongVU is Meta-led, and the OWL/CAMEL agent framework spun out to a startup — all tracked elsewhere. Elhoseiny is an incoming visiting faculty member at Stanford.
Your notes
People
- Jürgen Schmidhuber Google Scholar — Director, KAUST AI Initiative; Co-Chair, Center of Excellence for Generative AI
- Bernard Ghanem Google Scholar — Professor, ECE; Chair, Center of Excellence for Generative AI
- Mohamed Elhoseiny Google Scholar — Associate Professor; Vision-CAIR lead (formerly Incoming Visiting Faculty, Stanford)
- Peter Wonka Google Scholar — Professor, CS (3D / generative vision)
- Marco Canini Google Scholar — Professor, CS; PI, SANDS Lab (systems for large-model training)