Supermicro Among First to Unveil NVIDIA BlueField-4 STX Storage Server to Improve AI Inference Performance

Rhea-AI Summary

Supermicro (NASDAQ: SMCI) unveiled one of the first NVIDIA BlueField-4 STX Context Memory (CMX) storage servers to accelerate AI inference and long-lived query workloads. The CMX server pairs NVIDIA Vera CPU and ConnectX-9 SuperNIC, and uses NVIDIA Dynamo for Key Value (KV) cache management.

The system builds on Supermicro's Petascale JBOF experience with BlueField-3 and is shown at NVIDIA GTC 2026, with partner SSD testing planned with Micron, Samsung, Phison and others.

News Market Reaction – SMCI

-1.10%

1 alert

-1.10% News Effect

On the day this news was published, SMCI declined 1.10%, reflecting a mild negative market reaction.

Data tracked by StockTitan Argus on the day of publication.

Key Figures

Current price: $31.86 RTX PRO 6000: RTX PRO 6000 BlueField-4: BlueField-4 +5 more

8 metrics

Current price $31.86 Pre-news market context

RTX PRO 6000 RTX PRO 6000 GPU used in AI Data Platform solutions

BlueField-4 BlueField-4 Generation of NVIDIA DPU in new STX storage server

BlueField-3 BlueField-3 Prior Petascale JBOF all-flash array platform

Booth number #1113 Supermicro booth at NVIDIA GTC 2026

Conference dates March 16–19, 2026 NVIDIA GTC 2026 event window

52-week high $62.358 Pre-news 52-week high level

52-week low $27.60 Pre-news 52-week low level

Market Reality Check

Price: $30.35 Vol: Volume 28,682,205 vs 20-d...

normal vol

$30.35 Last Close

Volume Volume 28,682,205 vs 20-day avg 25,024,817 (relative 1.15x). normal

Technical Shares at 31.86, trading below 200-day MA at 41.06 despite AI product news.

Peers on Argus

SMCI up 3.61% while peers are mixed: HPQ and PSTG down, WDC, STX, LOGI modestly ...

SMCI up 3.61% while peers are mixed: HPQ and PSTG down, WDC, STX, LOGI modestly up. Moves do not show a unified hardware-sector trend.

Previous AI Reports

5 past events · Latest: Mar 02 (Positive)

Same Type Pattern 5 events

Date	Event	Sentiment	Move	Catalyst
Mar 02	AI infra expansion	Positive	-1.7%	Expanded AI-RAN and sovereign AI systems with new NVIDIA-based platforms.
Feb 25	AI data platform	Positive	+7.9%	Launched CNode-X enterprise AI data platform with NVIDIA and VAST Data.
Feb 25	AI demand response	Positive	+7.9%	Demonstrated GPU AI compute for fast grid demand response with partners.
Jan 06	Client/edge AI launch	Positive	+1.6%	Unveiled desktop, edge and client AI systems including Super AI Station.
Jan 05	Rubin AI platforms	Positive	+1.6%	Announced support for NVIDIA Rubin platforms and expanded liquid-cooled capacity.

Pattern Detected

AI-tagged announcements have mostly led to positive price reactions, with four gains and one decline, suggesting investors often reward SMCI’s AI product and platform updates.

Recent Company History

Over recent months, SMCI repeatedly highlighted AI-centric platforms and NVIDIA-aligned solutions. Early January news on Rubin-based clusters and client/edge AI systems saw modest gains. Late February AI data platform launches and demand-response demonstrations coincided with ~7.9% rises. The March 2 AI-RAN and sovereign AI expansion produced a -1.73% reaction, showing occasional divergence. Today’s GTC-linked BlueField-4 STX CMX storage server fits this ongoing narrative of deepening NVIDIA collaboration and AI factory infrastructure focus.

Historical Comparison

+3.5% avg move · Across recent AI-tagged announcements, SMCI’s average move was 3.45%. Today’s 3.61% pre-news gain si...

AI

+3.5%

Average Historical Move AI

Across recent AI-tagged announcements, SMCI’s average move was 3.45%. Today’s 3.61% pre-news gain sits close to that norm, suggesting a typical-sized reaction.

AI news has moved from announcing NVIDIA Rubin support and client/edge systems to full AI data platforms and now BlueField-4 STX CMX storage, reflecting a broadening AI infrastructure stack.

Market Pulse Summary

This announcement adds another NVIDIA-based AI infrastructure component, introducing a BlueField-4 S...

Analysis

This announcement adds another NVIDIA-based AI infrastructure component, introducing a BlueField-4 STX CMX storage server aimed at long-lived inference and agentic workloads. It follows earlier AI data platforms, Rubin system support, and client/edge AI launches, reinforcing SMCI’s positioning around full AI factory stacks. Investors may watch how quickly STX-based solutions and AI Data Platform offerings gain traction, alongside broader trends in AI server demand and the stock’s position relative to its 200-day MA.

Key Terms

context memory (cmx), nvidia stx reference architecture, dpu, key value (kv) cache, +1 more

5 terms

context memory (cmx) technical

"unveiled one of the industry's first context memory (CMX) storage server as part"

Context memory (cmx) is a running record of past statements, data points and events that helps interpret new corporate disclosures; think of it as a company’s notebook that keeps track of what was said, promised and shown before. For investors it matters because comparing new news to that history reveals consistency, progress or surprises—similar to checking a player’s past performance before betting on them.

nvidia stx reference architecture technical

"built on the NVIDIA STX reference architecture for AI storage."

A reference architecture is a published technical blueprint showing how a company’s high-performance chips, software and compatible hardware can be combined, installed and upgraded in data center servers. Think of it like an automotive platform or building code that lets many manufacturers build parts that fit together reliably; investors care because such standards speed customer adoption, lower deployment costs and expand partner sales channels, which can boost revenue and market reach.

dpu technical

"we proved the feasibility of a JBOF powered by NVIDIA BlueField-3 DPUs, we have"

Distributions per unit (DPU) is the amount of cash or income paid to each unit holder of a trust, real estate investment trust (REIT) or similar pooled investment for a given period. It tells investors how much cash income they received per unit, like getting a fixed slice of a pie for every share you own, and helps compare yield and judge whether the payout level is steady, growing or at risk.

key value (kv) cache technical

"This storage of tokens, called Key Value (KV) cache, is managed by NVIDIA"

A key value (kv) cache is a small, fast storage system that keeps frequently used pieces of information paired with a simple label or “key,” so programs can retrieve them instantly without redoing slow work. Think of it as a labeled sticky-note drawer that helps a website or service respond faster and handle more users. For investors, a kv cache matters because it reduces operational costs, improves user experience and reliability, and enables platforms to scale efficiently, all of which can affect revenue and valuations.

inference orchestration layer technical

"KV cache, is managed by NVIDIA Dynamo, NVIDIA's inference orchestration layer."

An inference orchestration layer is the software that manages when, where and how machine learning models are used to produce predictions or decisions; think of it as a traffic controller that routes requests to the right model, ensures responses are fast and consistent, and applies safety or audit checks. For investors, it matters because this layer affects reliability, speed, regulatory traceability and cost of any automated forecasts, trading signals or risk assessments that a company relies on to drive revenue or meet compliance.

AI-generated analysis. Not financial advice.

• Supermicro illustrates leadership with one of the first Context Memory (CMX) storage servers, built on the NVIDIA STX reference architecture for AI storage.
• The BlueField-4 STX storage server combines NVIDIA Vera CPU and NVIDIA ConnectX-9 SuperNIC.
• Supermicro's storage server builds upon last year's introduction of the Petascale JBOF all-flash array powered by NVIDIA BlueField-3.

SAN JOSE, Calif., March 17, 2026 /PRNewswire/ -- Supermicro, Inc. (NASDAQ: SMCI), a Total IT Solution Provider for AI, Cloud, Storage, and 5G/Edge, today unveiled one of the industry's first context memory (CMX) storage server as part of NVIDIA STX reference architecture announced at NVIDIA GTC 2026. STX is a new modular reference architecture from NVIDIA which is designed to accelerate the full lifecycle of AI.

"Supermicro continues to be first to market with new rack scale architectures designed to exceed the needs of a rapidly evolving AI Factory customer base," said Charles Liang, president and CEO of Supermicro. "Building upon last year's introduction of the Petascale JBOF (Just a Bunch of Flash), where we proved the feasibility of a JBOF powered by NVIDIA BlueField-3 DPUs, we have developed the CMX storage server. Our prototype of the latest storage architecture demonstrates the level of our collaboration with NVIDIA, and our commitment to be first-to-market with game changing technologies."

For more information about the new Supermicro storage server built on the NVIDIA STX reference architecture please visit: www.supermicro.com/en/solutions/ai-storage

Leveraging the STX architecture, the CMX server is designed to address the challenge of long-lived AI queries and multi-stage chain-of-thought agentic workloads, which require the prior and intermediate tokens associated with the user's query to be accessed. This solution both accelerates the results and reduces the power which would otherwise be required to recompute the results when the local storage required for the tokens is exceeded. This storage of tokens, called Key Value (KV) cache, is managed by NVIDIA Dynamo, NVIDIA's inference orchestration layer.

As the STX solution comes to market, Supermicro will be working with these software partners and others on porting and validation. Additionally, Supermicro long-standing relationships with leading SSD providers such as Micron, Samsung, Phison, and others will enable testing for the specific STX architecture requirements.

At GTC 2026, Supermicro also announced seven AI Data Platform solutions based on the RTX PRO 6000 Blackwell Server Edition GPU with NVIDIA and storage partners such as Cloudian, DDN, Everpure (formerly Pure Storage), IBM, Nutanix, VAST Data, and WEKA. The AI Data Platform enables enterprises to process their data for AI workloads. The CMX server is being shown in Supermicro booth #1113, and at the NVIDIA exhibit, at the NVIDIA GTC 2026 March 16-19.

About Super Micro Computer, Inc.
Supermicro (NASDAQ: SMCI) is a global leader in Application-Optimized Total IT Solutions. Founded and operating in San Jose, California, Supermicro is committed to delivering first-to-market innovation for Enterprise, Cloud, AI, and 5G Telco/Edge IT Infrastructure. We are a Total IT Solutions provider with server, AI, storage, IoT, switch systems, software, and support services. Supermicro's motherboard, power, and chassis design expertise further enables our development and production, enabling next-generation innovation from cloud to edge for our global customers. Our products are designed and manufactured in-house (in the US, Taiwan, and the Netherlands), leveraging global operations for scale and efficiency and optimized to improve TCO and reduce environmental impact (Green Computing). The award-winning portfolio of Server Building Block Solutions® allows customers to optimize for their exact workload and application by selecting from a broad family of systems built from our flexible and reusable building blocks that support a comprehensive set of form factors, processors, memory, GPUs, storage, networking, power, and cooling solutions (air-conditioned, free air cooling or liquid cooling).

Supermicro, Server Building Block Solutions, and We Keep IT Green are trademarks and/or registered trademarks of Super Micro Computer, Inc.

All other brands, names, and trademarks are the property of their respective owners.

View original content to download multimedia:https://www.prnewswire.com/news-releases/supermicro-among-first-to-unveil-nvidia-bluefield-4-stx-storage-server-to-improve-ai-inference-performance-302716157.html

SOURCE Super Micro Computer, Inc.

FAQ

What is the Supermicro CMX storage server announced by SMCI on March 17, 2026?

The CMX storage server is a context memory storage system pairing NVIDIA Vera CPU and ConnectX-9 SuperNIC for AI inference acceleration. According to the company, it implements KV cache managed by NVIDIA Dynamo to support long-lived queries and multi-stage chain-of-thought workloads.

How does the SMCI BlueField-4 STX server improve AI inference performance?

It reduces recomputation by storing prior and intermediate tokens in a KV cache, accelerating responses and lowering power needs. According to the company, the STX architecture and NVIDIA Dynamo enable efficient access to long-lived AI query data for inference workloads.

Where and when did Supermicro (SMCI) showcase the BlueField-4 STX CMX server?

Supermicro showed the CMX server at NVIDIA GTC 2026, March 16–19, including its booth #1113 and the NVIDIA exhibit. According to the company, the prototype demonstration highlights its collaboration with NVIDIA and STX reference architecture integration.

Which NVIDIA and storage technologies power SMCI's new CMX storage server?

The CMX server combines NVIDIA Vera CPU, ConnectX-9 SuperNIC and BlueField-4 STX reference architecture components. According to the company, KV cache orchestration is provided by NVIDIA Dynamo and SSD testing involves Micron, Samsung, and Phison among partners.

Will Supermicro (SMCI) validate software and SSD compatibility for the STX CMX server?

Yes. Supermicro plans to work with software partners to port and validate on STX while testing SSDs for architecture requirements. According to the company, established relationships with leading SSD vendors will support specific STX validation and interoperability testing.

What workloads is the SMCI CMX server designed to address for AI customers?

The CMX server targets long-lived AI queries and multi-stage chain-of-thought agentic workloads that need prior and intermediate token access. According to the company, this design aims to accelerate inference results and reduce power from unnecessary recomputation.