Korean
At VMware Explore 2026, we spoke with Stuart McRae, Executive Director of Lenovo’s Enterprise Storage, Software, and Solutions Offering Group. Unlike typical product-focused interviews, this wide-ranging conversation covered AI-driven shifts in enterprise buying trends, the ongoing RAM shortage, and a lighthearted debate over Texas vs. North Carolina BBQ. Below are key takeaways from the discussion.
The High-Stakes Return to On-Premises Infrastructure
Early data center development prioritized custom-built infrastructure, with architects assembling bespoke environments from discrete compute, storage, and networking components. Today’s enterprise landscape has shifted fundamentally: skyrocketing generative AI costs and volatile public cloud pricing are driving a rapid return to on-premises deployments. Businesses now prioritize fast, predictable operational outcomes over customized architectural tweaks.
As McRae emphasized, the industry mission has evolved. Organizations are moving beyond piecemeal component assembly to adopt integrated, turnkey “AI factory” infrastructure—designed to be stable, low-maintenance, and seamlessly supported in daily operations.
RAM Shortage Drives Critical Memory Innovation
The industry faces a widespread “RAM apocalypse”, with DRAM and flash supply shortages inflating prices and creating major bottlenecks for data center expansion. This crisis has accelerated long-overdue memory architecture innovation.
Memory tiering is not a new technology, but early implementations relied on complex OS or application-layer configurations, making them unstable and hard to deploy at scale. Lenovo and VMware are solving this by embedding tiering intelligence directly into the VMware Cloud Foundation (VCF) hypervisor layer, offloading data to NVMe storage at the software-defined level to decouple performance from DRAM costs.
This approach delivers two core advantages: broad compatibility with existing systems, requiring no application overhauls or OS replacements; and up to 45% DRAM cost savings via optimized tiering. McRae noted, “The best time to buy a server was a year ago; the next best time is today.”
From Discrete Components to Pre-Validated Hybrid AI Factories
Enterprise IT priorities have shifted drastically: the era of experimental, customized infrastructure building has given way to demand for reliable, plug-and-play solutions. Sizing and validating AI infrastructure independently has become too complex for most organizations, fueling demand for standardized hybrid AI factories.
Lenovo integrates its hardware portfolio with the VCF Private AI stack to deliver fully pre-validated environments, ensuring seamless compatibility across SSDs, network cards, and firmware. Beyond hardware compatibility, the unified stack delivers robust lifecycle management and consistent stability through ongoing updates and patches. Enterprises are transitioning from component assemblers to operators of automated, unified infrastructure production lines.
Liquid Cooling Goes Mainstream for Enterprise Data Centers
Long limited to HPC and specialized research scenarios, liquid cooling has become a mainstream enterprise requirement. Dense deployments of high-performance CPUs and power-heavy GPUs have pushed traditional air cooling to its thermal limit.
Lenovo offers three tailored cooling architectures for modern data centers: Neptune warm water cooling, a highly efficient node-direct solution that eliminates energy-intensive chillers; closed-loop cooling, a flexible middle option for facilities without water infrastructure, with fully chassis-internal liquid circulation; and baseline air cooling, now reserved exclusively for low-density workloads.
Fast Logistics & Price Stability Become Competitive Advantages
Amid volatile component markets, deployment speed and cost predictability have become critical infrastructure priorities. Lenovo’s “Top Choice Express” and “AI Express” programs deliver VCF-ready, AI-optimized servers in just 10 days. Complementing fast delivery is a 30-day price guarantee—a rare stability advantage in a market where typical quotes expire in 72 hours, especially valuable for budget-constrained public sector enterprises.
McRae highlighted unprecedented hardware demand at the software-focused VMware Explore event, with infrastructure cost control sessions drawing full crowds and heavy booth traffic.
AI Transforms System Administration with Natural Language Operations
A defining industry shift is emerging: AI is evolving from a standalone workload to a native hypervisor interface. Following Broadcom’s launch of VMware Private AI Cloud and VCF AI Assistant, paired with third-party management pack support, enterprise infrastructure is moving toward natural language-driven administration.
Similar to Tassient Aipex’s cross-OS AI remediation capabilities, VCF’s integrated AI assistant will automate complex IT workflows. Administrators will simply query system issues in plain language—such as identifying crash causes—enabling the AI to pinpoint driver faults, locate patches, and execute remediation with user approval. This eliminates decades of manual, labor-intensive troubleshooting and provisioning work, marking a major leap in operational efficiency.
Three-Year Private Cloud Development Blueprint
McRae outlined a clear three-year roadmap for private cloud evolution centered on full infrastructure integration. VCF and vSAN will become enterprise default standards, with GPU resources fully clustered as core infrastructure rather than specialized auxiliary hardware. Driven by data sovereignty and operational efficiency demands, the integration-first model will overtake complex public cloud deployments.
Ultimately, AI will render infrastructure “invisible” by automating troubleshooting, remediation, and provisioning. Enterprise IT will shift from building customized component stacks to running efficient, reliable AI infrastructure factories—letting technology serve business needs seamlessly.
On the lighthearted closing BBQ debate, McRae offered a neutral take, praising North Carolina’s tangy vinegar-style BBQ and Texas’ slow-smoked brisket equally, with no comment on Cincinnati-style BBQ.
Beijing Qianxing Jietong Technology Co., Ltd.
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!