By: Masahisa Kawashima, NTT, Technology WG Chair, the IOWN Global Forum
At the Open Compute Project (OCP) EMEA Summit 2026 in Barcelona, I presented the IOWN Global Forum perspective in the AI Computing Continuum (AICC) session, alongside Sean Varley from Ampere and the OCP AICC project. One of the strongest impressions from this year’s summit was that the AI Computing Continuum is rapidly emerging as a major strategic topic within the OCP community.
This post outlines what AICC is, why it matters, and how related discussions are evolving.
What is the AI Computing Continuum? Why does it matter?
AI infrastructure technologies are evolving rapidly. Over the past several years, terms such as Scale-Up, Scale-Out, and Scale-Across have become common throughout the industry. This reflects how strongly the industry has focused on expanding scale — larger GPU clusters, larger AI data centers, and larger distributed training environments. This direction was natural because the primary users of AI infrastructure have so far been AI providers developing large-scale foundation models and LLMs.
However, the next phase of AI adoption will be different. As AI increasingly transforms industries and society, the primary users of AI infrastructure will expand beyond hyperscalers to include enterprises, public institutions, universities, telecom operators, factories, logistics operators, and many other organizations.
To support this broader adoption of AI, infrastructure must evolve to satisfy a new set of requirements. Most importantly, it is critical to provide high-bandwidth, low-latency connectivity between AI infrastructure and the physical spaces where industries and societies operate — factories, campuses, offices, cities, hospitals, transportation systems, and edge environments.
In other words: Scale is not everything. Connectivity is equally important. This is one of the key motivations behind the AICC.
The AI Computing Continuum is a collaborative initiative between OCP and the IOWN Global Forum that aims to create a unified computing space spanning cloud to edge, built on high-bandwidth and low-latency networking infrastructure.
The goal is not simply to interconnect data centers, but to make geographically distributed computing resources operate as a coordinated and manageable infrastructure fabric.
While the industry has begun discussing Scale-Across — connecting multiple data centers to further expand AI infrastructure scale — AICC considers that inter-data-center connectivity provides benefits beyond simple scaling. Based on this perspective, AICC defines three “Across” models as follows:
- Scale-Across — expanding compute capacity across geographically distributed regions
- Sync-Across — synchronizing data and system states across distributed environments
- Reach-Across — enabling real-time access between enterprise/edge environments and cloud resources
Together, these three models aim to make AI infrastructure not only scalable, but also distributed, coordinated, and deeply connected to the environments where AI will actually be used.
Looking ahead, this unified computing space is also expected to evolve into a platform that enables application developers to more easily build and deploy AI-enabled services across distributed cloud, edge, and industry environments.
AICC Discussions at OCP EMEA 2026
Throughout OCP EMEA 2026, several OCP executives presented the “Open Data Center Ecosystem Vision,” which positioned the future data center ecosystem around three major domains: AI Data Centers, Cloud Data Centers, and AI Computing Continuum (AICC). Within this vision, AICC was consistently recognized as a key growth area for the next phase of infrastructure evolution.
One particularly notable phrase repeatedly used during discussions was “Hyperscaler Adjacents.”
This expression suggests that AICC is expected to complement and extend the computing spaces provided by hyperscalers, connecting them with enterprise, edge, telecom, sovereign, and industrial environments through high-bandwidth and low-latency infrastructure.
During the summit, it also became clear that collaboration between organizations such as OCP and the IOWN Global Forum can play a highly complementary role in advancing AICC.
The IOWN Global Forum focuses on developing vertical use cases together with leaders from industries such as manufacturing, finance, logistics, telecommunications, and public infrastructure. Based on these use cases, the Forum develops end-to-end architectures, validates them through Proof of Concept (PoC) activities, and analyzes their techno-economic feasibility.
Meanwhile, OCP provides the open hardware ecosystem and product standardization framework needed to realize such systems at scale.
OCP has recently launched the AICC project, which already includes active subprojects such as:
- AI-native Wireless Access
- AI-native Edge Modular Hardware System (MHS)
These activities indicate that AICC is now evolving from a high-level architectural vision toward a concrete ecosystem initiative spanning networking, compute platforms, edge infrastructure, and AI service deployment models.
Conclusion
AI infrastructure is entering a new phase. The next challenge is no longer only about building larger AI clusters, but also about connecting distributed computing environments into a unified and operationally manageable infrastructure fabric.
AI Computing Continuum represents one possible architectural direction for that future.
At OCP EMEA 2026, it was clear that interest in AICC is growing rapidly across the industry. The next step will be to collaboratively develop concrete problem-solution narratives, practical use cases, and deployable architectures that can turn this vision into reality. This discussion is expected to continue at upcoming OCP events, including the OCP APAC Summit in Taipei.
