Two hundred people in Asia-Pacific by 2027. Mistral AI set this goal when announcing on 20 November 2026 the opening of its regional headquarters in Singapore. Presented as an extension of the company's scientific hub, the office will develop models for the public sector, high-tech industry and robotics. That is more than double the current headcount in the region.
Arthur Mensch, co-founder and CEO of Mistral, sees it as a long-term commitment "to building sovereign AI capabilities in the region, for the region". Governments will be able to adapt and run its open-weight models in their own environments, without their data leaving a defined perimeter, the company says.
Mistral has been working with Singapore's defence and home affairs ministries since March 2025
The public sector is the most advanced of the three areas. The office builds on two public partnerships. In March 2025, Singapore's Ministry of Defence, its technology agency DSTA and DSO National Laboratories announced they were developing generative AI models with Mistral to support decision-making for the armed forces. DSO, the country's leading defence research organisation, wanted to train its own multilingual models and run them on its premises. The two teams co-developed, with language data obtained from AI Singapore, a mixture-of-experts model, an architecture that activates only part of its parameters for each item processed. The second phase focuses on small reasoning models. They are intended for autonomous systems deployed in the field.
HTX, the science and technology agency of the Ministry of Home Affairs, which works notably for the police, prison service and immigration authorities, has been collaborating with Mistral since March 2025. It wanted models that understand the Singaporean context and run on infrastructure isolated from the internet. Phoenix emerged from this. The Small and Medium versions of this language model series, pre-trained on Mistral's base models and then fine-tuned on HTX data, run on NGINE, the agency's AI infrastructure. Unveiled in April, Phoenix-VL 1.5 Medium processes text and images with 123 billion parameters.
A memorandum of understanding signed in November 2025, followed by a framework agreement concluded on 28 April 2026, extended this work to robotics, video analytics and cybersecurity. In the spring, Singtel, NCS and ST Engineering also signed memoranda of understanding with Mistral.
Mistral's first navigation model guides a robot with a single camera
In high-tech industry, Mistral wants to accelerate engineering simulations and optimise processes. In robotics, the release highlights Robostral Navigate, unveiled on 8 July 2026. This 8-billion-parameter model, designed in-house and trained entirely in simulation, drives a wheeled, legged or flying robot from a natural language instruction and images from a single colour camera, without depth sensors. It points in the image to the next point to reach. To access it, Mistral refers to its teams.
In June, DSTA extended its research with Mistral to drone navigation guided by vision-language models. HTX, for its part, expects robots that understand natural language commands and find their way alone in cities, in unfamiliar places. For Pee Beng Kong, executive vice president of the Economic Development Board, Singapore's economic development agency, this investment will create "opportunities for Singaporeans to work on cutting-edge AI technologies".
Singapore headquarters comes less than two months after Munich hub
On 28 September, Mistral opened a hub in Munich where teams work on physical AI and industrial AI research, as well as engineers serving corporate clients. Deutsche Telekom announced on 19 November it had invested 100 million euros in the company, which had raised 3 billion euros in September in a round led by South Korea's Samsung Electronics.
The Singapore project dates back to January 2025. Mistral was then planning to open an office there to cover Asia-Pacific, Bloomberg reported, picked up by TechCrunch.
Our articles will then appear first in Google Top Stories.
