HomeTechnology
Evening edition
OpenAI and Singapore signed a memorandum of understanding for an 'OpenAI for Singapore' partnership that will establish an applied AI lab — OpenAI’s first outside the United States — and expand a local technical team to about 200 roles, with the company committing over S$300 million. The lab will focus on applied deployments, talent programmes and collaborations with public‑sector and industry partners.
OpenAI and the Singapore government signed a memorandum of understanding on May 20 to establish “OpenAI for Singapore,” the company’s first applied AI lab outside the United States, and to expand its local technical team to roughly 200 roles. Singapore’s Economic Development Board said the partnership will focus on applied deployments, talent programmes and collaborations with public‑sector and industry partners, and noted a company commitment of more than S$300 million (press release: https://www.edb.gov.sg/en/about-edb/media-releases-publications/openai-launches-applied-ai-lab-in-singapore.html). Bloomberg reported the commitment with a different figure, headlining it as $234 million (https://www.bloomberg.com/news/articles/2026-05-20/openai-commits-234-million-for-new-ai-lab-in-singapore?srnd=homepage-europe).
The Singapore announcement underscores a broader industry push this week to expand AI compute, diversify supply chains and increase commercial access to specialised hardware. In China, Alibaba’s chip unit T‑Head unveiled the Zhenwu M890, a training‑and‑inference accelerator the company says triples performance versus its predecessor and is optimised for so‑called “agentic” workloads. Alibaba also introduced a server system that packages multiple accelerators into single nodes as part of a multi‑year roadmap to expand domestic AI compute and model infrastructure amid export curbs on U.S. suppliers (reporting: https://www.bloomberg.com/news/articles/2026-05-20/alibaba-unveils-new-ai-chip-for-training-and-inferencing?srnd=homepage-europe; ).
In the United States, Google and private equity firm Blackstone announced a joint venture to provide data‑centre capacity bundled with Google’s Tensor Processing Units (TPUs). Blackstone committed an initial $5 billion of equity to the venture, which has a stated target of bringing 500 megawatts of capacity online by 2027 — a move aimed at broadening commercial access to Google’s custom AI chips as demand for specialised infrastructure grows (https://news.bloomberglaw.com/california-brief/google-blackstone-to-create-cloud-business-to-handle-ai-demand).
Markets reacted to the flurry of AI and chip‑related developments: U.S. stock futures and semiconductor shares gained on May 20 as investors positioned ahead of Nvidia’s quarterly results, which market participants treated as a barometer of sustained demand for AI infrastructure (reporting: https://www.investing.com/news/stock-market-news/us-stock-futures-climb-as-chip-stocks-rebound-ahead-of-nvidia-results-4700711).
At the same time, policymakers are watching for risks to physical supply. South Korea’s president urged an “appropriate limit” on collective labour action after talks between Samsung Electronics and its union stalled and the union said an 18‑day work stoppage would proceed, raising concerns about potential disruptions at a major global semiconductor supplier (https://www.bloomberg.com/news/articles/2026-05-20/s-korea-s-lee-urges-limit-to-labor-action-amid-samsung-unrest).
Taken together, the Singapore lab, new domestic chip announcements, cloud partnerships and market moves this week reflect a technology landscape focused on scaling specialised compute, securing supply chains and expanding commercial pathways for AI. The initiatives announced on May 20 highlight how governments, chipmakers and cloud providers are deploying capital and policy to capture opportunities and mitigate risks as demand for large‑scale AI infrastructure accelerates.