Singapore's Minister for Foreign Affairs Dr Vivian Balakrishnan used his 16 May 2026 keynote at AI Engineer Singapore to walk an audience of 2,000-plus builders through a fully local, personal AI stack he runs on a Raspberry Pi — Ollama for model serving, whisper.cpp for voice transcription, NanoClaw as an agentic interface, and Obsidian for the underlying note graph. The demo was unusual for a cabinet-level keynote: it skipped the macro framing common at policy speeches and instead put a working developer stack on stage as the argument.
The substantive point was sovereignty without protectionism. By showing that frontier-capable workflows can be run on commodity hardware with open-weight models, Balakrishnan framed the local AI agenda as one of capability ownership — individuals, agencies, and enterprises being able to inspect, replicate, and modify the stack — rather than restricting access to foreign models. The full transcript was published by Singapore's Ministry of Foreign Affairs.
The keynote landed on the same conference day OpenAI and the Singapore government signed a S$300 million MOU at ATxSummit committing to OpenAI's first Applied AI Lab outside the United States, and Minister for Digital Development Josephine Teo's separate AI Engineer talk reinforced the policy direction. The pairing — a cabinet endorsement of local, open-source AI workflows alongside a flagship deal with a US frontier lab — is consistent with the dual track Singapore has been signalling through IMDA and the National AI Strategy 2.0: anchor frontier partnerships while ensuring local engineers and procurement officers retain the option to deploy fully self-hosted stacks for sensitive workloads.
