NVIDIA Vera Rubin as a new standard for artificial intelligence infrastructure

Dmitry Namiot, Vladimir Sukhomlin

Abstract


This article analyzes the new NVIDIA Vera Rubin platform, introduced in 2026 and positioned by the company as a qualitative leap in building computing infrastructure for artificial intelligence. Unlike traditional approaches focused on individual chips, the platform is considered as a holistic system, combining Vera (Arm) processors, Rubin graphics accelerators, high-speed NVLink interfaces, ConnectX-9 network adapters, and programmable BlueField-4 DPUs. Particular attention is paid to hardware support for agent-based AI, including spatial multithreading, distributed key-value caching, and the scalable NVL72 rack-mount architecture. A separate section is devoted to the use of digital twins based on Omniverse DSX for the design and operation of large-scale AI factories. The authors conclude that the Vera Rubin platform marks a shift from performance evaluation by peak FLOPS to system-wide optimization of memory and network bandwidth, setting new standards for infrastructure solutions in the field of artificial intelligence.


Full Text:

PDF (Russian)

References


Hsu, Kuan-Chieh, and Hung-Wei Tseng. "Simultaneous and heterogenous multithreading: Exploiting simultaneous and heterogeneous parallelism in accelerator-rich architectures." IEEE Micro 44.4 (2024): 11-19.

Li, Qingyuan, et al. "Flash communication: Reducing tensor parallelization bottleneck for fast large language model inference." arXiv preprint arXiv:2412.04964 (2024).

Mu, Siyuan, and Sen Lin. "A comprehensive survey of mixture-of-experts: Algorithms, theory, and applications." arXiv preprint arXiv:2503.07137 (2025).

Namiot, Dmitry, and Eugene Ilyushin. "On Architecture of LLM agents." International Journal of Open Information Technologies 13.1 (2025): 67-74.

Durante, Zane, et al. "Agent ai: Surveying the horizons of multimodal interaction." arXiv preprint arXiv:2401.03568 (2024).

Krishnan, Naveen. "Ai agents: Evolution, architecture, and real-world applications." arXiv preprint arXiv:2503.12687 (2025).

Kupriyanovsky, Vasily, et al. "Digital Economy and the Internet of Things-negotiating data silo." International Journal of Open Information Technologies 4.8 (2016): 36-42.

Namiot, Dmitry, and Eugene Ilyushin. "On the Cybersecurity of AI Agents." International Journal of Open Information Technologies 13.9 (2025): 13-24.

Buscemi, Alessio, et al. "Towards Sandboxes for the Internet of Agents." Available at SSRN 5801322 (2025).

Zheng, Yusheng, et al. "AgentCgroup: Understanding and Controlling OS Resources of AI Agents." arXiv preprint arXiv:2602.09345 (2026).

Ahn, Seok-hyun, et al. "Studying the Universal Scene Description (USD) file format from a Digital Twin Convergence Perspective." International journal of advanced smart convergence (2025): 224-241.


Refbacks

  • There are currently no refbacks.


Abava  Кибербезопасность Monetec 2026 СНЭ

ISSN: 2307-8162