Super Micro Computer (SMCI) shares fell 6.52% to $33.25 on October 22, a decline of $2.32, despite the company announcing an expansion of its NVIDIA Vera AI infrastructure portfolio. Heavy selling erased earlier gains and pushed the stock toward its intraday lows as Supermicro introduced a liquid-cooled CPU rack designed for advanced artificial intelligence workloads. The session underscored that product announcements, even significant ones, do not necessarily dictate short-term share performance, which can be shaped by trading flows and broader market conditions.

Supermicro Expands Its NVIDIA Vera Infrastructure Portfolio

Super Micro Computer introduced the NVIDIA Vera CPU rack as part of its Data Center Building Block Solutions portfolio. The system is designed for agentic AI, reinforcement learning, and large-scale data processing workloads, expanding Supermicro’s infrastructure offerings for companies building high-performance AI data centers. Agentic AI, in which software agents carry out multi-step tasks with limited direct supervision, has emerged as one of the more resource-intensive workloads, prompting infrastructure vendors to design purpose-built racks around high core counts and concurrency.

The rack contains 256 NVIDIA Vera CPUs, providing 22,528 processing cores and 45,056 threads. It can support up to 22,500 simultaneous agent sandboxes in a single rack, enabling businesses to run multiple AI tasks while managing complex computing operations.

Supermicro also integrated NVIDIA BlueField-4 processors to enhance networking performance and hardware security. The system uses the company’s DLC-2 cooling technology for fully liquid-cooled operation across the rack. The design addresses the power and cooling requirements associated with expanding AI infrastructure — a concern that has become increasingly central as AI racks pack more compute into denser footprints.

Rack Supports NVIDIA Vera Rubin NVL72 Systems

Supermicro designed the new CPU rack to operate independently or alongside NVIDIA Vera Rubin NVL72 systems. The configuration allows data center to expand computing capacity based on their workload requirements while using shared infrastructure across large AI facilities.

The system is based on NVIDIA’s third-generation MGX architecture, which is intended to simplify system integration and infrastructure management. Customers can choose between conventional AC power and an 800-volt DC power configuration. The alternative power design supports battery backup units and can reduce equipment space requirements. Power delivery choices of this kind have become a key design decision for AI facilities as operators balance efficiency, space, and reliability at scale.

Supermicro also added support for NVIDIA OpenShell to strengthen security during autonomous AI operations. OpenShell applies access controls and monitors agent activity while tasks are running. NVIDIA Sentry uses BlueField-4 hardware to identify policy violations and isolate unauthorized activity. Security controls for autonomous agents have drawn growing attention as enterprises deploy AI systems that operate with less direct human oversight.

SMCI Declines Despite New AI Product

SMCI shares declined sharply despite the product announcement and Supermicro’s continued collaboration with NVIDIA. Selling pressure dominated the trading session and reversed gains recorded earlier in the day, overshadowing the company’s expansion into high-density CPU infrastructure for AI workloads.

Supermicro develops infrastructure for enterprise computing, AI processing, storage, and edge applications. Its latest platform also supports database management, data analytics, software development, and cluster scheduling. Compatibility with Kubernetes and Slurm extends its potential applications beyond dedicated AI factories.

The NVIDIA Vera launch adds to Supermicro’s product lineup as demand for specialized data center equipment expands. However, the announcement did not prevent a substantial decline in SMCI shares, creating a contrast between the company’s infrastructure expansion and its latest stock performance. Going forward, attention is likely to center on how the new rack line translates into customer deployments and disclosed order activity, and on whether the market’s focus returns to Supermicro’s expanding role in NVIDIA-based AI infrastructure.

Source: Blockonomi