ASRock Rack Unveils Next-Gen AI Infrastructure Powered by NVIDIA Vera CPU at COMPUTEX 2026

Key Takeaways

• 2UXGM-VERA2 uses NVIDIA Vera CPU, 50% faster than traditional CPUs
• Vera combines Olympus cores, LPDDR5X, and Scalable Coherency Fabric
• Liquid‑cooled AI servers integrate NVIDIA HGX Rubin NVL8 for high density
• Edge model 2UXGI‑Thor leverages NVIDIA IGX Thor for real‑time sensor AI
• ASRock Rack targets AI factories, cloud inference, and industrial robotics markets

Pulse Analysis

The rise of agentic AI—systems that act autonomously and coordinate with other AIs—has shifted the performance bottleneck from GPUs to CPUs. NVIDIA’s Vera processor, introduced in ASRock Rack’s 2UXGM-VERA2, tackles this challenge with custom Olympus cores and a high‑bandwidth LPDDR5X memory subsystem, delivering roughly 50% faster execution on complex reinforcement‑learning tasks. This architecture signals a broader industry move toward CPUs that can handle the rapid decision‑making loops required for autonomous agents, complementing existing GPU‑centric models.

ASRock Rack’s hardware rollout emphasizes advanced liquid‑cooling to sustain the dense compute loads of modern AI factories. By pairing Vera CPUs with NVIDIA HGX Rubin NVL8 GPUs in fully liquid‑cooled 2U and 5U chassis, the company offers a turnkey solution that maximizes thermal efficiency while maintaining high throughput. The inclusion of rack‑scale cooling units and in‑row CDUs further supports hyperscale deployments, reducing energy costs and footprint. For edge and industrial use, the 2UXGI‑Thor platform leverages the NVIDIA IGX Thor system‑on‑chip, enabling real‑time sensor processing and functional safety for robotics and medical devices.

From a market perspective, ASRock Rack’s announcement could reshape competitive dynamics among server OEMs. The integration of a purpose‑built AI CPU with proven NVIDIA GPU technology provides a compelling value proposition for cloud providers and enterprises seeking end‑to‑end AI infrastructure. As data‑center operators prioritize performance per watt and rapid AI model iteration, the liquid‑cooled, AI‑optimized servers may accelerate adoption of agentic AI workloads. Analysts will watch early deployment metrics to gauge whether the promised 50% speed uplift translates into measurable cost savings and faster time‑to‑insight for AI‑driven businesses.