CyberSecurity news

FlagThis - #nvidia

@tomshardware.com //
Nvidia has unveiled its next-generation data center GPU, the Blackwell Ultra, at its GTC event in San Jose. Expanding on the Blackwell architecture, the Blackwell Ultra GPU will be integrated into the DGX GB300 and DGX B300 systems. The DGX GB300 system, designed with a rack-scale, liquid-cooled architecture, is powered by the Grace Blackwell Ultra Superchip, combining 36 NVIDIA Grace CPUs and 72 NVIDIA Blackwell Ultra GPUs. Nvidia officially revealed its Blackwell Ultra B300 data center GPU, which packs up to 288GB of HBM3e memory and offers 1.5X the compute potential of the existing B200 solution.

The Blackwell Ultra GPU promises a 70x speedup in AI inference and reasoning compared to the previous Hopper-based generation. This improvement is achieved through hardware and networking advancements in the DGX GB300 system. Blackwell Ultra is designed to meet the demand for test-time scaling inference with a 1.5X increase in the FP4 compute. Nvidia's CEO, Jensen Huang, suggests that the new Blackwell chips render the previous generation obsolete, emphasizing the significant leap forward in AI infrastructure.

Recommended read:
References :
  • AIwire: Nvidia’s DGX AI Systems Are Faster and Smarter Than Ever
  • www.tomshardware.com: Nvidia officially revealed its Blackwell Ultra B300 data center GPU, which packs up to 288GB of HBM3e memory and offers 1.5X the compute potential of the existing B200 solution.
  • BigDATAwire: Nvidia's GTC 2025 conference showcased the new Blackwell Ultra GPUs and updates to its AI infrastructure portfolio.
  • www.laptopmag.com: Blackwell Ultra and Rubin Ultra are Nvidia's newest additions to the growing list of AI superchips
  • BigDATAwire: Nvidia used its GTC conference today to introduce new GPU superchips, including the second generation of its current Grace Blackwell chip, as well as the next generation, dubbed the Vera The post appeared first on .
  • venturebeat.com: Nvidia's GTC 2025 keynote highlighted advancements in AI infrastructure, featuring the Blackwell Ultra GB300 chips.
  • Analytics Vidhya: An overview of Nvidia's GTC 2025 announcements, including new GPUs and advancements in AI hardware.
  • AI News: NVIDIA Dynamo: Scaling AI inference with open-source efficiency
  • www.tomshardware.com: Nvidia unveils DGX Station workstation PCs with GB300 Blackwell Ultra inside
  • BigDATAwire: Nvidia Preps for 100x Surge in Inference Workloads, Thanks to Reasoning AI Agents
  • Data Phoenix: Nvidia introduces the Blackwell Ultra to support the rise of AI reasoning, agents, and physical AI
  • The Next Platform: This article discusses Nvidia's new advancements in AI, and how the company is looking to capture market share and the challenges they face.

info@thehackernews.com (The Hacker News)@The Hacker News //
A critical security vulnerability, CVE-2024-0132 (CVSS score: 9.0), has been identified in the NVIDIA Container Toolkit, potentially allowing attackers to completely compromise the host system. Researchers at Wiz have uncovered a new exploit, CVE-2025-23359 (CVSS score: 8.3), that bypasses the original patch for CVE-2024-0132. This vulnerability allows malicious actors to escape the container's isolation, gaining full access to the underlying host, posing a significant risk to sensitive data and infrastructure.

Security researchers Shir Tamari, Ronen Shustin, and Andres Riancho found that file paths used during mount operations could be manipulated using a symbolic link, making it possible to mount from outside the container into a path within "/usr/lib64." This allows an attacker to mount the host's root file system into a container, granting unrestricted access to all files and enabling the launch of privileged containers and achieve full host compromise via the runtime Unix socket. It is strongly recommended that users immediately update to NVIDIA Container Toolkit version 1.17.4 and NVIDIA GPU Operator version 24.9.2, which addresses both vulnerabilities, and to not disable the "--no-cntlibs" flag in production environments.

Recommended read:
References :
  • The Hacker News: Researchers Find New Exploit Bypassing Patched NVIDIA Container Toolkit Vulnerability
  • Wiz Blog | RSS feed: Critical NVIDIA container vulnerability (CVE-2024-0132) allows full host compromise. Update to NVIDIA Container Toolkit 1.17.4 immediately and restrict access to privileged runtime sockets. The post appeared first on .
  • BleepingComputer: Information about an NVIDIA vulnerability and its bypass.
  • Open Source Security: CVE-2025-23359: Nvidia-container-toolkit: GPU Container Escape (CVE-2024-0132 fix bypass)
  • Security Risk Advisors: NVIDIA Container Toolkit Vulnerability Enables Full Host Compromise (CVE-2024-0132 & CVE-2025-23359)

Fred Oh@NVIDIA Newsroom //
NVIDIA's CUDA libraries are increasingly vital in modern cybersecurity, bolstering defenses against emerging cyber threats like malware, ransomware, and phishing. Traditional cybersecurity measures struggle to keep pace with these evolving threats, especially with the looming risk of quantum computers potentially decrypting today's data through "harvest now, decrypt later" strategies. NVIDIA's accelerated computing and high-speed networking technologies are transforming how organizations protect their data, systems, and operations, enhancing both security and operational efficiency.

CUDA libraries are crucial for accelerating AI-powered cybersecurity. NVIDIA GPUs are essential for training and deploying AI models, offering faster AI model training, enabling real-time inference for identifying vulnerabilities, and automating repetitive security tasks. For example, AI-driven intrusion detection systems, powered by NVIDIA GPUs, can analyze billions of events per second to detect anomalies that traditional systems might miss. This real-time threat detection and response capability minimizes downtime and allows businesses to respond proactively to potential cyberattacks.

Recommended read:
References :
  • NVIDIA Newsroom: CUDA Accelerated: How CUDA Libraries Bolster Cybersecurity With AI
  • TechPowerUp: NVIDIA Explains How CUDA Libraries Bolster Cybersecurity With AI

@cybersecuritynews.com //
NVIDIA has released a critical security update to address vulnerabilities in its GPU Display Driver and vGPU software, affecting both Windows and Linux systems. The update, issued in January 2025, aims to mitigate risks such as information disclosure, denial of service, data tampering, and potential code execution. Users are strongly advised to install the latest updates via the NVIDIA Driver Downloads page or the NVIDIA Licensing Portal.

Among the key vulnerabilities addressed is CVE-2024-0149, a flaw in the NVIDIA GPU Display Driver for Linux that could allow attackers unauthorized access to files. This low-severity vulnerability could lead to limited information disclosure. The security update also addresses high-severity vulnerabilities like CVE-2024-0150, a buffer overflow in the GPU Display Driver for Windows and Linux which could lead to information disclosure, denial of service, or data tampering.

Recommended read:
References :
  • cyberpress.org: NVIDIA GPU Driver Flaw Allow Attackers to Access Files Remotely
  • gbhackers.com: NVIDIA GPU Display Drivers Vulnerability Lets Attackers Access Files Remotely
  • Cyber Security News: NVIDIA GPU Driver Flaw Allow Attackers to Access Files Remotely
  • gbhackers.com: NVIDIA GPU Display Drivers Vulnerability Lets Attackers Access Files Remotely
  • Cyber Security News: NVIDIA GPU Display Driver Vulnerability Lets Attackers Steal Files Remotely – Update Now
  • Pyrzout :vm:: Report on NVIDIA GPU Display Driver Vulnerability