Latest / Government and geopolitics
Japan’s GENIAC programme gives seven groups government compute to train foundation models
See how this story connectsMETI’s Generative AI Accelerator Challenge (GENIAC), run with NEDO, began in February 2024 by giving selected developers time on GPU clusters to train foundation models. Sakana AI said on 2 February 2024 that it was one of seven institutions chosen to receive cluster access for several months in 2024.
Why it matters
Japanese model builders short of GPUs gained subsidised access to large clusters, and the government began funding domestic foundation models directly.
- When
- 2 Feb 2024
- ID
WTA-N-2024-0161- Where
- Japan
- Who
- METI, NEDO, Sakana AI
- Topics
- Sovereign AI, Frontier models, Data centres
Summary by wheresthe.ai, from Sakana AI and the reporting credited above. Checked against the source on 9 Oct 2026.
Source copy: download the page as saved on 18 Oct 2026 (PDF). Document the source offers: document 1 (PDF). Also held by the Internet Archive.
Data to measure this
Compare the years before 2024 with the years since in these series.
- Epoch AI: frontier AI data centresEpoch AISite-level timelines of IT power, compute (H100-equivalents) and capital cost for large AI data centres.
- IEA Energy and AI dataInternational Energy AgencyWorld and regional data-centre capacity, power usage effectiveness, load factor and electricity use.
- Epoch AI: GPU clusters (archived)Epoch AIMore than 500 GPU clusters and AI supercomputers with country, owner, performance, power and cost.
Cite this story
WTA-N-2024-0161: “Japan’s GENIAC programme gives seven groups government compute to train foundation models.” wheresthe.ai, 2 Feb 2024. https://wheresthe.ai/d/jp-geniac-launch-compute-grants-2024-02/ (source: Sakana AI).
Spotted an error? Email [email protected].