📡 Breaking news
Analyzing latest trends...

NVIDIA has launched the NVIDIA DGX SuperPOD, based on the Rubin platform, each containing 600TB of RAM.

-
NVIDIA has launched the NVIDIA DGX SuperPOD, based on the Rubin platform, each containing 600TB of RAM.
Following NVIDIA's unveiling of the Rubin architecture details today, they also introduced the NVIDIA DGX SuperPOD, a server cluster deployment model for each chip variant.

Two versions of the NVIDIA DGX SuperPOD for Rubin are available: a DGX Vera Rubin NVL72 rackmount version using the Vera ARM architecture CPUs, and a DGX Rubin NVL8 rackmount version using the x86 architecture.

The DGX Vera Rubin NVL72 rackmount SuperPOD features 576 Rubin GPUs, 600TB of RAM, a total processing power of 28.8 exaflops in FP4 mode, and NVLink connectivity with a maximum data transfer rate of 260TB/s. The DGX Rubin NVL8 SuperPOD incorporates 512 Rubin chips.

Both SuperPOD models are expected to begin shipping in the latter half of this year.

 


 

💬 AI Content Assistant

Ask me anything about this article. No data is stored for your question.

Comments

Post a Comment

Popular posts from this blog

[Rumor] Meta Secret Model Avocaco Slips to May Following Underwhelming Benchmarks.

-
Image
  Meta Secret AI "Avocaco" Faces Delays: Internal Tests Struggle to Outperform Google’s Gemini 3.0 Reports late last year suggested that Meta was developing a formidable new AI model codenamed "Avocaco." Intended to succeed the Llama series, Avocaco was rumored to be a proprietary, closed-source model designed to bridge the gap with industry rivals. While an initial launch was expected within Q1 2026 , new information suggests the project has hit a significant roadblock. Performance Gap and Internal Friction According to The New York Times , citing three internal sources at Meta, the launch of Avocaco has been pushed back to at least May 2026 . Despite being Meta’s most powerful model to date, benchmark results have reportedly been underwhelming when compared to current market leaders. Sources indicate that while Avocaco successfully outperforms Google’s older Gemini 2.5 , it has failed to surpass the capabilities of Gemini 3.0 (released late last year). This per...
Read more

From Startup to Security Standard Promptfoo Joins OpenAI to Bolster LLM Protection.

-
Image
OpenAI Acquires AI Security Startup "Promptfoo" to Bolster LLM Vulnerability Defense OpenAI has officially announced the acquisition of Promptfoo , a prominent AI security startup specializing in tools for identifying vulnerabilities within Large Language Models (LLMs). While the financial terms of the deal remain undisclosed, the move signals OpenAI’s intensifying focus on enterprise-grade security. A Proven Track Record in Enterprise Security Founded by Ian Webster and Michael D’Angelo , Promptfoo has quickly become a cornerstone of AI safety. Its tools are currently utilized by over 25% of Fortune 500 companies and are widely adopted within the open-source community to evaluate and stress-test models for potential exploits. Securing the Automated Workspace As organizations increasingly integrate AI into their core workflows and automated safety evaluations, the surface area for potential attacks has expanded. OpenAI noted that with the rise of automated AI systems, mali...
Read more

The "Forced" Upgrade Windows 10 User Outraged After PC Automatically Installs Windows 11 Without Consent

-
Image
Outsmarted by an Update: Reddit User Claims Microsoft Forced Windows 11 During a 30-Minute Shower. A heated debate has erupted on Reddit after a user claimed his PC was forcibly upgraded from Windows 10 to Windows 11, despite him repeatedly declining the invitation for months. The user reported that for the past two months, he had been bombarded with persistent pop-ups and "Install Update" prompts nestled next to the power and restart buttons. Despite his consistent refusals, the upgrade allegedly occurred while he was away from his desk. After leaving his computer on to take a 30-minute shower, he returned to find the system performing an in-place upgrade to Windows 11. The incident left the user so infuriated that he reportedly placed a call to Microsoft support to voice his extreme dissatisfaction. The Push for the New Era This incident follows the end of mainstream support for Windows 10 last October. Currently, Windows 10 users only receive security updates through the ...
Read more

Adobe Settles FTC Lawsuit for $150M Over "Difficult-to-Cancel" Subscriptions.

-
Image
Adobe Settles $150M Lawsuit Over Deceptive Cancellation Practices Adobe has officially reached a settlement following a long-standing legal battle with the U.S. Federal Trade Commission (FTC) . The lawsuit, initially filed in mid-2024, accused the software giant of intentionally complicating the subscription cancellation process for services like Creative Cloud through hidden buttons, mandatory phone calls, and excessive early termination fees. The Settlement Details After nearly two years of litigation, Adobe has agreed to a total settlement package of $150 million . This includes: $75 million in civil penalties paid to the FTC. $75 million set aside for customer restitution to compensate affected users in the United States. Despite the significant payout, Adobe maintains its stance of non-wrongdoing, stating that while they deny any violation of the law, they are pleased to put this legal matter behind them. This case is a major victory for the FTC in cracking down on so-called ...
Read more

Microsoft Teams Up with Anthropic "Copilot Cowork" Brings Claude Intelligence to Microsoft 365.

-
Image
Microsoft Unveils "Copilot Cowork": Bringing Anthropic’s Claude Intelligence Directly into Microsoft 365 Microsoft has officially introduced Copilot Cowork , a powerful new AI agent designed to function as a seamless digital colleague within the Microsoft 365 ecosystem. By integrating Anthropic’s Claude Cowork technology directly into the platform, Microsoft is aiming to transform how users manage their daily workflows across its suite of productivity apps. Unified Security and Seamless Integration Through a strategic partnership with Anthropic , Microsoft has embedded Claude Cowork into the Microsoft 365 Copilot infrastructure. A primary advantage of this integration is the unified security layer . Users no longer need to manage separate security protocols between different apps; Copilot Cowork inherits all existing Microsoft 365 configurations, including: Identity Verification (MFA/SSO) Access Permissions Enterprise Compliance Policies Real-World Utility: Your New Executi...
Read more

Smartphone Stalemate Apple and Samsung Tie for World's Top Producer in 2025.

-
Image
Global Smartphone Market 2025: Apple and Samsung Tied for First Place Amidst Rising Production Costs According to the latest report from market research firm TrendForce , the global smartphone landscape saw a historic deadlock in 2025. Both Apple and Samsung produced approximately 240 million units each, making them joint leaders in the global smartphone market. Market Overview and Growth Drivers The total global smartphone production in 2025 reached 1.254 billion units , a 2.5% increase year-over-year. This growth was primarily fueled by two factors: Government Subsidies: Aggressive subsidy programs in China during the first half of the year boosted consumer demand. Flagship Season: The second half of the year saw a surge in production driven by major flagship launches, particularly the iPhone 17 series . Apple witnessed a dramatic 50% production spike in the fourth quarter compared to Q3, a result of the strong market reception of the iPhone 17 and a well-executed pricing stra...
Read more

AWS and Cerebras Launch Inference Disaggregation to Slash AI Latency on Bedrock.

-
Image
  Cerebras Partners with Amazon to Supercharge Bedrock with "Inference Disaggregation" Technology Cerebras Systems , the pioneer in specialized AI accelerators, has announced a strategic partnership with Amazon Web Services ( AWS ) to integrate its powerhouse CS-3 chips into the Amazon Bedrock ecosystem. The highlight of this deal is a sophisticated hybrid solution known as "Inference Disaggregation." This approach optimizes AI performance by splitting the inference workload into two distinct stages, handled by different specialized hardware: The Hybrid Architecture: Trainium + CS-3 The two-step process utilizes the strengths of both Amazon's in-house silicon and Cerebras's massive wafer-scale engine: Prompt Processing (Prefill Stage): This stage is computationally intensive as it processes the user's initial prompt in parallel. Since it requires high compute power but moderate memory bandwidth, it is handled by Amazon Trainium chips. Output Genera...
Read more