G42 consortium anchors Vietnam’s sovereign AI push

Abu Dhabi-based technology group G42 has signed a $1bn framework cooperation agreement with a Vietnamese consortium led by FPT Corporation and Viet Thai Group, marking one of the largest cross-border technology collaborations aimed at building sovereign artificial intelligence and cloud infrastructure in Southeast Asia.

The agreement sets out a multi-year plan to deploy large-scale cloud capacity across three data centre locations in Vietnam, alongside the development of domestic AI platforms, data governance frameworks and sector-specific applications. The initiative is positioned as a cornerstone of Vietnam’s ambition to reduce reliance on foreign hyperscalers while accelerating digital transformation across government and industry.

Under the framework, G42 will provide core AI models, cloud architecture and operational expertise, while the Vietnamese partners will contribute local infrastructure, regulatory coordination and market access. FPT Corporation, Vietnam’s largest listed technology company, is expected to lead software development, systems integration and enterprise deployment, while Viet Thai Group will support infrastructure build-out and financing structures.

The project aligns with Vietnam’s national strategy to develop domestic AI capacity as part of its broader digital economy agenda, which targets technology-driven growth in manufacturing, finance, healthcare and public services. Officials familiar with the discussions say the emphasis on “sovereign AI” reflects growing sensitivity around data localisation, model governance and national security, particularly as AI systems become embedded in critical infrastructure.

Cloud deployment will focus on hyperscale-grade facilities designed to host high-performance computing workloads required for large language models and advanced analytics. Industry executives involved in the talks indicate that the data centres will be optimised for energy efficiency and high-density compute, addressing one of the main constraints facing AI expansion in emerging markets. Vietnam’s relatively young data centre sector has expanded quickly, but demand from AI training and inference has begun to outpace available capacity.

For G42, the agreement extends its strategy of exporting AI and cloud capabilities through government-aligned partnerships rather than pure commercial expansion. The group has built a portfolio spanning healthcare AI, geospatial analytics, fintech and large-scale cloud services, and has increasingly focused on markets seeking alternatives to US and Chinese technology ecosystems. Southeast Asia, with its fast-growing digital economies and complex regulatory environments, has emerged as a priority region.

FPT Corporation has framed the partnership as a step change in Vietnam’s ability to move up the AI value chain. While domestic firms have built strong capabilities in software outsourcing and enterprise IT, large-scale model development and cloud infrastructure have largely remained the domain of global players. By anchoring compute and model development locally, the consortium aims to enable Vietnamese enterprises and public agencies to build applications without transferring sensitive data offshore.

Early use cases under discussion include smart manufacturing systems, digital government platforms, financial risk analytics and healthcare diagnostics tailored to local languages and datasets. Executives say an important focus will be Vietnamese-language AI models, which have lagged behind global counterparts due to limited training data and compute resources. Locally hosted models are expected to improve accuracy while meeting data residency requirements.

The agreement also reflects intensifying competition among emerging markets to secure AI infrastructure investment. Governments across Asia are racing to attract capital and expertise to avoid being locked into external platforms, while balancing concerns over cost, skills shortages and power supply. Vietnam’s appeal lies in its political stability, expanding renewable energy capacity and strong pipeline of engineering talent, though grid constraints and land availability remain challenges for large data centre projects.

Analysts note that the $1bn headline figure is likely to cover phased investments tied to capacity milestones and demand uptake rather than a single upfront outlay. Such structures are common in large technology infrastructure projects, allowing partners to scale spending as workloads migrate onto the platform. Financing is expected to combine equity contributions, long-term debt and potential support from development-linked institutions, though detailed terms have not been disclosed.