Blackwell GPUs

Extracted Attributes

• Type

  GPU Microarchitecture

• Developer

  Nvidia

• Named After

  David Blackwell (statistician and mathematician)

• Successor To

  Hopper microarchitecture, Ada Lovelace microarchitecture

• Key Feature 1

  Second-generation Transformer Engine

• Key Feature 2

  Confidential Computing (TEE-I/O-capable GPU)

• Power Connector

  16-pin (partially adopted)

• NVLink Bandwidth

  50GB/sec per link

• Transistor Count

  208 billion

• Die Configuration

  Combines two large dies into a unified GPU

• NVLink Generation

  Fifth-generation

• NVLink GPU Support

  Up to 576 GPUs

• Other Applications

  High-performance embedded computing (defense applications), ray tracing, neural graphics

• Primary Application

  Generative AI, Large Language Models (LLMs), AI training, real-time inference

• Inter-die Link Speed

  10 terabytes per second

• Manufacturing Process

  TSMC 4NP

• Peak Performance (single chip)

  20 petaFLOPS

• Performance vs H100 (GB200 for LLM inference)

  30x increase

• Cost/Energy Reduction vs H100 (GB200 for LLM inference)

  25x

Timeline

• The name of the Blackwell architecture was leaked. (Source: Wikipedia)

  2022

• B40 and B100 accelerators were confirmed in an official Nvidia roadmap shown during an investors presentation. (Source: Wikipedia)

  2023-10-01

• Blackwell architecture was officially announced at Nvidia's GTC 2024 keynote. (Source: Wikipedia)

  2024-03-18

• Nvidia officially detailed the first four cards in the Blackwell RTX 50-series family at CES 2025 (expected). (Source: Web Search Results)

  2025-01-06

Blackwell (microarchitecture)

Blackwell is a graphics processing unit (GPU) microarchitecture developed by Nvidia as the successor to the Hopper and Ada Lovelace microarchitectures. Named after statistician and mathematician David Blackwell, the name of the Blackwell architecture was leaked in 2022 with the B40 and B100 accelerators being confirmed in October 2023 with an official Nvidia roadmap shown during an investors presentation. It was officially announced at Nvidia's GTC 2024 keynote on March 18, 2024.

Web Search Results

• NVIDIA Blackwell GPUs: Architecture, Features, Specs

  NVIDIA Blackwell is the largest GPU ever created, featuring 208 billion transistors—more than 2.5 times the transistors in NVIDIA's Hopper GPUs. It utilises TSMC’s 4NP process specifically developed for NVIDIA. Blackwell offers groundbreaking compute performance, reaching 20 petaFLOPS on a single chip. It achieves this by combining two large dies into a unified GPU, with each die being as large as possible within reticle size limits. These dies are linked via a high-speed 10 terabyte-per-second [...] NVIDIA Blackwell is a cutting-edge GPU architecture designed to meet the demands of generative AI, offering exceptional performance with features like the second-generation Transformer Engine, confidential computing, and advanced NVLink for scaling large AI models. ### Does the NVIDIA Blackwell use NVlink? Yes, the NVIDIA Blackwell uses fifth-generation NVLink with 50GB/sec per link bandwidth and supports up to 576 GPUs, providing seamless scaling for large AI workloads. [...] NVIDIA Blackwell introduces Confidential Computing to protect sensitive data, extending Trusted Execution Environments (TEE) to GPUs. As the first TEE-I/O-capable GPU, Blackwell offers robust, secure, and high-performing solutions for handling LLMs and confidential AI training. It ensures excellent throughput in encrypted modes and safeguards AI intellectual property. These capabilities allow secure AI training, inference, and federated learning, protecting sensitive models while maintaining

• Blackwell (microarchitecture) - Wikipedia

  Blackwell is a graphics processing unit (GPU) microarchitecture developed by Nvidia as the successor to the Hopper "Hopper (microarchitecture)") and Ada Lovelace "Ada Lovelace (microarchitecture)") microarchitectures. [...] At the Graphics Technology Conference (GTC) on March 18, 2024, Nvidia officially announced the Blackwell architecture with focus placed on its B100 and B200 datacenter accelerators and associated products, such as the eight-GPU HGX B200 board and the 72-GPU NVL72 rack-scale system.( Nvidia CEO Jensen Huang said that with Blackwell, "we created a processor for the generative AI era" and emphasized the overall Blackwell platform combining Blackwell accelerators with Nvidia's ARM-based Grace CPU.( [...] 21. ^#cite_ref-21)Prickett Morgan, Timothy (March 18, 2024). "With Blackwell GPUs, AI Gets Cheaper and Easier, Competing with Nvidia Gets Harder". _The Next Platform_. Retrieved March 24, 2024. 22. ^ _a_#cite_ref-Nvidia_2024-03-18_22-0)_b_#cite_ref-Nvidia_2024-03-18_22-1)"Nvidia Blackwell Platform Arrives to Power a New Era of Computing". _Nvidia Newsroom_. March 18, 2024. Retrieved March 24, 2024.

• NVIDIA Blackwell: Next-Gen AI-Accelerated Embedded Computing

  NVIDIA’s Blackwell architecture is a game-changer for high-performance embedded computing in defense applications. NVIDIA’s Blackwell architecture builds upon years of GPU innovation, bringing cutting-edge AI, ray tracing, and neural graphics to HPEC systems deployed in land, sea, air, and space operations. Designed to process vast amounts of real-time sensor data, power AI-enhanced decision-making, and enable next-generation visualization, Blackwell GPUs provide a significant leap in

• All About the NVIDIA Blackwell GPUs: Architecture, Features, Chip ...

  The announcement revealed NVIDIA's latest breakthrough called the “Blackwell”. It is the NVIDIA Blackwell architecture succeeding the NVIDIA Hopper architecture launched on September 20, 2022. The NVIDIA Blackwell GPU is named after David Harold Blackwell, a statistician and mathematician specialising in game theory and statistics.The NVIDIA Blackwell GPU will enable organisations across the globe to build and run real-time inference on trillion-parameter large language models at 25x less cost [...] Two B200 GPUs are combined in Blackwell’s flagship accelerator, the NVIDIA GB200 Grace Blackwell chip, which also utilises an NVIDIA Grace CPU. The GB200 provides a 30x performance increase compared to the NVIDIA H100 Tensor Core GPU for LLM inference workloads and reduces cost and energy consumption by 25x. [...] NVIDIA Blackwell Features to Power Generative AI ------------------------------------------------ NVIDIA Blackwell GPUs are powered by six revolutionary technologies that will enable AI training and real-time LLM inference for models scaling up to 10 trillion parametres. The ground-breaking GPU will include the following features:

• Nvidia Blackwell and GeForce RTX 50-Series GPUs - Tom's Hardware

  The next-generation Nvidia Blackwell GPU architecture and RTX 50-series GPUs are coming, basically on schedule. Nvidia officially detailed the first four cards in the Blackwell RTX 50-series family at CES 2025, during CEO Jensen Huang's keynote on January 6. We expect the various Blackwell GPUs will join the ranks of the best graphics cards, replacing their soon-to-be-prior-generation counterparts. [...] The consumer Blackwell GPUs are "late," based on historical precedent. The Ada Lovelace RTX 40-series GPUs first appeared in October 2022. The Ampere RTX 30-series GPUs first appeared in September 2020. Prior to that, RTX 20-series launched two years earlier in September 2018, and the GTX 10-series was in May/June 2016, with the GTX 900-series arriving in September 2014. That's a full decade of new Nvidia GPU architectures arriving approximately every two years. Even so, we're still only a few [...] widespread use of 16-pin, that means Blackwell will be the third generation of Nvidia GPUs to at least partially adopt the standard.