Red Hat, a global open-source solutions company, held the "Red Hat Tech Day 2026" event on March 11 at Yangjae L Tower.

At the event, held under the theme of an "AI-ready hybrid cloud platform building strategy", Red Hat introduced practical ways to implement enterprise AI that go beyond simple hardware expansion. It presented approaches spanning an operable platform, a secure security structure and a scalable architecture.

In a keynote speech, Choi Won-young (최원영), executive director at Red Hat Korea, presented Red Hat's integrated operations platform for enterprise-grade AI. "A corporate AI operating environment should be based on the same technologies and operating methods as existing IT environments, which is advantageous in terms of costs and governance," Choi said. "Red Hat OpenShift makes it possible to manage existing environments and new AI environments on a single platform," he added. Choi also said it was time to focus on AI inference and stressed the importance of an inference architecture that enables optimal use of GPU performance.

In subsequent sessions, various presentations were delivered focusing on customer cases and technology strategies. KEPCO KDN shared a cloud-native transition case based on Red Hat OpenShift, including challenges in the build process, solution strategies and tangible results.

In technical sessions, an open-source-based AI inference architecture strategy was introduced.

Red Hat presented ways to raise GPU utilization efficiency by using vLLM PagedAttention and Continuous Batching technologies. It also suggested ways to implement scalable AI services in enterprise environments through quantization and distributed inference.

It also held various sessions on topics including a proactive OS security lifecycle built with Red Hat solutions from detection to action; adoption and migration strategies for OpenShift Virtualization to enable stable service transitions; ways to apply platform security that had been delayed; and a shift in DR architecture for business continuity in the cloud-native era.