Super Micro Stock Spotlighted After NVIDIA AI Blueprint
Super Micro stock drew focus after a DCBBS blueprint based on the NVIDIA Vera Rubin platform, spurring AI backlog bets, heavy trading and a GF upgrade.
KEY TAKEAWAYS
- Supermicro published a DCBBS blueprint built on NVIDIA Vera Rubin NVL4 for liquid-cooled, rack-scale AI servers.
- Blueprint defines a 3.2 MW Scalable Unit: eight liquid-cooled racks, 36 nodes per rack, up to 1,152 GPUs.
- GF Securities upgraded to Buy with a $48 target, citing the $7 billion capital raise and AI-server demand.
HIGH POTENTIAL TRADES SENT DIRECTLY TO YOUR INBOX
Add your email to receive our free daily newsletter. No spam, unsubscribe anytime.
Super Micro Computer, Inc. (NASDAQ: SMCI) drew investor attention on June 22, 2026, after unveiling a Data Center Building Block Solutions (DCBBS) blueprint based on NVIDIA’s Vera Rubin NVL4 platform for liquid-cooled, rack-scale AI servers. Analysts cited the company’s recent $7 billion capital raise as context for a GF Securities upgrade.
DCBBS Blueprint Details
Supermicro said in a June 22 press release that it delivered an end-to-end DCBBS blueprint designed as a rack-scale solution for converged high-performance computing (HPC) and AI workloads. The package is described as a “complete, modular AI infrastructure built from validated components and subsystems.”
The blueprint defines a 3.2-megawatt “Scalable Unit” composed of eight liquid-cooled compute racks. Each rack contains 36 NVL4 nodes within a 362-kilowatt power envelope. A full Scalable Unit supports up to 1,152 NVIDIA Rubin GPUs and 576 NVIDIA Vera CPUs. Each rack includes eight 72-kilowatt power shelves delivering busbar power.
Cooling and power delivery are central to the design. The system uses Supermicro’s DLC-2 direct-to-chip liquid-cooling architecture with three in-row cooling distribution units per Scalable Unit arranged in 2+1 redundancy. Copper cold plates and proprietary SMC PG25-A coolant circulate through the racks. Networking relies on NVIDIA’s Quantum-X800 InfiniBand via dedicated switch racks, with options for fully liquid-cooled networking.
The blueprint offers native FP64 (double-precision) performance targeting combined HPC and AI tasks such as climate modeling, drug discovery, materials science, and energy research. This NVL4 blueprint follows earlier DCBBS designs for NVIDIA’s NVL72 and NVL8 platforms revealed at Computex.
Analyst Upgrade and Financial Context
GF Securities upgraded Supermicro from Hold to Buy on June 22, setting a $48 12-month price target. The firm cited the recent $7 billion capital raise, which implied roughly a 15% increase in share count and pressured sentiment, contributing to a 28–30% stock pullback. GF argued that strong demand for AI servers warranted a more positive outlook.
The analyst projects NVL72 rack shipments of about 7,200 units in fiscal 2026 and 12,000 units in fiscal 2027. This translates into potential NVL72-related revenues of roughly $24 billion in fiscal 2026 and $51 billion in fiscal 2027. These figures are sell-side estimates, not company guidance.
Supermicro describes itself as an “Application-Optimized Total IT Solutions” provider for enterprise, cloud, AI, and 5G/edge customers. Its portfolio includes servers and AI systems, storage, IoT and switch systems, software, support services, and modular rack-scale solutions. Fiscal 2025 revenue was about $22 billion, with trailing 12-month revenue near $33.7 billion. Trailing net income stood around $1.3 billion, with trailing EPS of about $1.91 and approximately 647 million shares outstanding.
On June 18, 2026, the Federal Register posted a “Proposed Production Activity” notice for Supermicro’s facilities in San Jose, Fremont, and Milpitas, California, tied to production authority in a Foreign-Trade Zone.
