“In the era of generative AI, Red Hat's strategy is to become an AI platform that supports a wide range of hardware--including GPUs, NPUs, and CPUs--and serves as the foundation for the models and agents built on top of them.”
In a recent interview, Daniel Au, Vice President and General Manager for Asia Pacific (APAC) at Red Hat, emphasized that the company's key competitive advantage in the generative AI era lies in its open-platform strategy, which avoids dependence on any specific technology or vendor.
“During the early stages of AI adoption, some providers promoted all-in-one solutions that bundled GPUs, inference engines, and models together,” Au said. “Red Hat, however, remained committed to its open-platform philosophy. The market is increasingly moving in the direction we have long believed in.”
According to Au, Red Hat's differentiation stems from its commitment to a “100% open-source” approach. Having started as a provider of server operating systems, Red Hat expanded into a data center platform company in the cloud era through Kubernetes. In the AI era, it is evolving into an open platform that connects semiconductor resources such as GPUs and NPUs with AI models and agents.
“Not every company that uses open source is truly an open-source company,” Au said. “Red Hat remains deeply committed to its open-source roots and strives to keep everything 100% open source.”
As generative AI moves beyond experimentation into real-world business applications and services, cost optimization has become a major concern. Citing the case of BC Card, Au explained that Red Hat's AI platform is designed to reduce the total cost of ownership (TCO) of AI deployments.
“To lower AI platform TCO, organizations need to utilize GPU resources more efficiently,” he said. “Red Hat supports GPU and NPU slicing, as well as large language model (LLM) quantization and compression, reducing the number of GPUs required. Through distributed inference based on llm-d, multiple applications can also share a common GPU pool.”
Au also stressed the growing importance of AI governance. Red Hat AI 3.4 supports a “Model-as-a-Service” approach, enabling organizations to leverage models deployed across both on-premises and cloud environments. A key feature is its treatment of user prompts as first-class data assets rather than simple inputs. This allows enterprises to track the entire workflow, from prompt submission and model interaction to agent analysis and execution.
On open-source security, Au highlighted the role of enterprise platforms in an era where AI can identify open-source vulnerabilities and newly disclosed Common Vulnerabilities and Exposures (CVEs) more rapidly than ever. He argued that companies need standardized open-source governance frameworks and should work with trusted providers that possess strong security capabilities.
Regarding South Korea, Au sees significant growth potential in both AI and open source. “Korea has historically been a market with strong adoption of commercial Unix systems, while Linux and open-source technologies are also widely embraced,” he said. “Many organizations rely on community-driven or locally developed open-source solutions. As security risks increase, this creates opportunities for Red Hat to provide enterprise-grade open-source platforms.”
