IBM eyes APAC’s demand for sovereign AI with new Foxconn-backed platform

Foxconn’s head of AI and quantum computing Jesse Chao predicts the ‘big trend’ of data sovereignty ‘will happen soon’ in the region and sees enterprises going for a hybrid cloud play.

IBM and Visionbay.ai, Foxconn’s subsidiary for supercomputing and cloud-AI operations, will collaborate to develop a full-stack AI platform for enterprises and regulated industries across Asia Pacific.

The integrated platform aims to support organizations from model development, fine-tuning to deployment, as well as provide other AI services. Built on Visionbay’s AI infrastructure and cloud operations, Red Hat OpenShift will provide the hybrid cloud foundation for managing AI workloads and IBM’s watsonx will support AI development and governance capabilities.

The collaboration was motivated by growing demand for data sovereignty from enterprises across the region.

“We believe that data sovereignty or data centre sovereignty, even all the way until the data supply chain, will happen soon,” said Jesse Chao, head of AI and Quantum Computing Chairman Office at Foxconn during a media briefing at IBM Think Singapore.

“It’s a top priority for most of the countries because no matter [the sector] you want to have ownership of all these data. […] Then you want control. You want to fit it into your own models, so that’s where we [see] more of the hybrid cloud play. Even hyperscalers feel like they have to localize in a way, so we see this big trend happening.”

Currently demand is coming mostly from model makers and hyperscalers, he said, but there is growing interest from enterprises in regulated industries like financial services, defense, and telecommunications — customers he counts as ‘first adopters’ of advanced technologies.

“Most of the enterprises [are] still trying to figure out how to best leverage these large language models (LLMs) to become one of their use cases,” said Chao. “[It will] still take a little bit of time to help these LLMs to kind of converge into a smaller one, which can also fit into the hybrid cloud or the local cloud, and [for enterprises] to build everything on their own.”

The AI platform can be accessed through Visionbay’s AI app store, which will offer ready-to-use AI agents and AI applications tailored for businesses while maintaining three critical layers — data residency, localized operational control, as well as governance and compliance with regulatory and security standards.

IBM and Visionbay intend to adopt a build-operate-localize model to support enterprises scaling their AI operations. It will begin with deployments in Taiwan and expand across APAC.

“What I hear a lot across Asia Pacific is the need for technology sovereignty, and when we look at that through the lens of IBM, sovereignty means four different things,” said Hans Dekkers, general manager at IBM APAC.

“First is data needs to be localized. Normally, when people talk about technology sovereignty, it’s all about the data but there are three other things that are equal [importance].

“The second one is: Are you technology sovereign? […] Are you building on open source? If the vendor is not there, will the systems operate? Open source is a very important ingredient of that.

“The third one is skills. Do you have the skills to operate your technology landscape locally, or do you need vendors or external influences. Can you operate it on your own or not?

“The fourth one is governance. Can you govern your technology landscape locally, or do you need someone else to do it? These four elements actually give you a perfect answer on technology sovereignty.”