Friday, November 20, 2026

Microsoft: "Sovereign AI Is Not Domestic AI... We Must Secure Choice Across the Ecosystem" [Global AI Briefing]

Input
2026-10-06 08:12:13
Updated
2026-10-06 08:12:13
[Financial News] Amid the accelerating race for AI supremacy, experts have pointed out that AI sovereignty, or Sovereign AI, does not simply mean developing a domestic AI model. The true meaning of Sovereign AI lies not only in directly developing a specific AI model, but also in who controls the entire ecosystem—including data, AI models and computing infrastructure—and whether the necessary components can be selected when needed.
On the 5th (local time), Microsoft (MS) explained the concept of AI sovereignty, or Sovereign AI, and related technologies in a cloud blog post jointly announced with NVIDIA.
Didier Ongena, Microsoft's public-sector and sovereign lead, emphasized in a report, "Sovereign AI encompasses not only where data is stored, but also which AI model processes the data, who manages the infrastructure that runs the AI, and whether the AI can continue operating when its connection to external environments is cut off."
On the 5th (local time), Didier Ongena, Microsoft's public-sector and sovereign lead, emphasized in a cloud blog post, "Sovereign AI encompasses not only where data is stored, but also which AI model processes the data, who manages the infrastructure that runs the AI, and whether the AI can continue operating when its connection to external environments is cut off." (Source: Yonhap News Agency)

The report particularly emphasized that having the freedom to choose across the entire AI ecosystem is central to Sovereign AI.
He pointed out, "Businesses and governments should not be dependent on a single AI model. They must be able to choose a model suited to their work and switch to another when a better model becomes available." The assessment is that excessive dependence on a particular model or cloud service makes it difficult to respond to rapid advances in AI technology.
He also emphasized that where AI is run is important. Even if general tasks are processed on a public cloud, sensitive information held by financial institutions and public agencies must be capable of being processed at their own data centers or on a private cloud.
He added that in places such as manufacturing sites, where connecting to external networks is difficult, AI must also be capable of running directly on-site.
In this regard, voices are growing within South Korea's domestic industry that the country, which is concentrating its policy efforts on securing its own foundation model, must secure control over the entire ecosystem—including data, GPUs, data centers, cloud services, security, model choice and operational capabilities—not just the model itself in order to achieve AI sovereignty.
An industry source pointed out, "The government and industry need to be alert to the fact that dependence on a particular global company's cloud and platform has actually been increasing in recent domestic AI development efforts."
[email protected] Lee Gu-soon Reporter