Andrew Chen, VP – Head of Platform & Products at Magna AI, discusses the shift towards production-ready AI, the importance of governance and accountability, and the partnerships driving sovereign AI capabilities across global markets
Enterprises and governments are increasingly shifting their focus from AI experimentation to deploying scalable systems that deliver measurable business value. Sovereign AI, governance, security and operational accountability are becoming critical priorities as organisations integrate AI into core business processes.
Since its launch in October 2025, Magna AI has positioned itself as an end-to-end AI transformation partner, securing USD 1.5 billion in signed commitments across five markets in its first year. Its strategy combines AI infrastructure, platforms, applications and governance to help organisations transition from isolated pilots to secure, scalable production environments.
With expanding partnerships and projects across the Gulf, Southeast Asia and Europe, Magna AI is also focusing on strengthening local AI capabilities, developing sovereign-ready AI factories and building sustainable ecosystems.
Andrew Chen, VP – Head of Platform & Products at Magna AI, shares insights into the lessons shaping the company’s growth, the evolving definition of AI sovereignty and the role of strategic partnerships in translating AI investments into tangible business outcomes.
What lessons from Magna AI’s first year have shaped your approach to helping customers move from AI experimentation to measurable business outcomes?
We launched Magna AI to move organisations from pilots to production-grade, sovereign-ready AI operations, and secured signed commitments of USD 1.5 billion across five markets in our first year. The first of three lessons is that technology is rarely the barrier, since scaling AI depends on the operating model, including ownership, integration, governance and security. Outcomes must be agreed before anything is built, with AI measured by productivity, cost or risk improvements rather than usage.
Customers want one accountable partner, so our forward-deployed engineers work alongside them from first workshop to managed run-state.
How can organisations bring infrastructure, data, security, governance and applications together into an accountable AI environment?
Moving AI into production means aligning infrastructure, data, platforms, security, governance and operating models around business outcomes, so they work as one environment, not separate layers. Compute provides capacity, platforms make it manageable across workloads, data needs governance over residency and access, and applications must connect AI to business and government processes. Accountability breaks down when these layers are procured separately without a common architecture, because each is optimised for its own layer while no one owns how the system performs. The objective is not fewer partners, but making them operate as one system with clear accountability across the lifecycle.
What should sovereign AI encompass beyond data residency, particularly as models and autonomous agents take on greater operational responsibility?
Sovereign AI goes well beyond keeping data or infrastructure within national borders, and means meaningful control across data, models, compute, applications, agents and operational decision-making. Agents add a new dimension, namely what an autonomous system may do on a nation’s behalf, where its actions are recorded and who can stop it. Organisations therefore need clear boundaries on what agents can access and do, with runtime guardrails, monitoring and policy enforcement keeping them in check. Infrastructure can be acquired and models licensed, but the expertise to operate, govern and improve AI must be developed, turning sovereign infrastructure into genuine sovereign capability.
What makes Malaysia and Thailand strategic markets for Magna AI, and which lessons can the Gulf and Southeast Asia exchange?
Southeast Asia is one of the world’s most dynamic digital infrastructure markets, with upcoming data centre capacity nearly four times current levels, led by Malaysia at over 6 GW and Thailand at around 3.5 GW. Our collaboration with Global Telecommunications and Zchwantech, signed at Computex 2026, initially focuses on both markets, including a USD 700M Malaysian programme built around Sarawak. We are also delivering NACX, Malaysia’s National AI Compute Exchange.
The Gulf can teach Southeast Asia to design for sovereignty from day one, while Southeast Asia’s speed in bringing power, connectivity and land together offers the Gulf a useful reference.
How will partnerships help Magna AI scale production deployments and build locally developed AI capabilities for international markets over the next 12–24 months?
Partnerships are central to our strategy because no single organisation can build the AI economy alone. Our full-value-chain model gives each partner a defined role, with infrastructure partners strengthening the foundation and systems integrators driving local delivery. Over the next 12–24 months, we will expand sovereign-ready AI factories and deploy industry platforms for finance, healthcare, energy, manufacturing and government, including a USD 500M AI factory partnership with GMI Cloud in Taiwan and further projects in Malaysia, Belgium and Romania. Through Centres of Excellence, workforce upskilling and MagnaGRID, our AI compute exchange, we are building lasting local capability rather than dependency.
Source: Tahawul Tech

