Nvidia has officially announced its next-generation microarchitecture, Blackwell, during the GTC 2024 keynote on March 18, 2024. The new design replaces the company’s existing Hopper and Ada Lovelace architectures, which currently power data centers and gaming systems. This move represents a foundational shift rather than an incremental update to the existing hardware lineup.
The architecture is named after David Blackwell, a statistician and mathematician, with the name first appearing in a leak in 2022. Industry observers have anticipated this release since that time.
Nvidia confirmed two specific accelerators under the Blackwell banner: the B40 and the B100. These products were previously listed on an official Nvidia roadmap during an investors presentation in October 2023. They are now confirmed as real products designed for artificial intelligence, machine learning, and high-performance computing workloads.
The timing of the announcement aligns with accelerating AI adoption across the technology sector. Major companies are deploying large language models and generative systems that require clusters of GPUs operating for weeks at a time.
Blackwell is engineered to address these demands by delivering higher performance per watt, which could reduce the significant costs associated with AI training and inference. Organizations that upgrade to the new hardware may train models faster and at lower expense, while those that do not risk falling behind competitively. The B40 and B100 accelerators are intended to integrate into existing data center infrastructure.
However, the new architecture will necessitate upgrades in cooling systems, power delivery, and software stacks, representing substantial investment and risk for buyers. The choice of the Blackwell name is deliberate, linking the architecture to rigorous mathematical thinking and distinguishing it from Nvidia’s consumer-focused GeForce branding.
The products are positioned for enterprise applications including hospitals, banks, and defense systems. Leaks and rumors about Blackwell have circulated for two years. The official confirmation provides clarity for customers who can now plan their hardware strategies with certainty.
Nvidia currently dominates the AI hardware market, with competitors such as AMD and Intel working to close the gap. Blackwell is expected to widen that gap by raising the performance bar and increasing the investment required for rivals to compete effectively.
Matching Blackwell’s efficiency and scale will require years of development and billions of dollars. The GTC 2024 keynote made clear that Nvidia is betting on continued growth in AI compute demand. The Blackwell microarchitecture and its B40 and B100 accelerators represent the company’s answer to that demand. For the broader industry, the message is that adaptation to this new hardware standard is necessary to remain competitive.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Nvidia Unveils Blackwell Microarchitecture in Major Computing Power Leap”, “description”: “Nvidia has officially announced its next-generation microarchitecture, Blackwell, during the GTC 2024 keynote on March 18, 2024. The new design replaces the com”, “datePublished”: “2026-06-22T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Sebastian Smith”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}, “image”: “https://infopulsetoday.com/wp-content/uploads/sites/3/rehomed/s24/2026/06/companion_3492.jpg”}
