
The Blackwell platform is Nvidia's upcoming GPU architecture, set to power the RTX 50-series graphics cards3. It is expected to bring significant improvements in performance and efficiency, as well as support for new memory formats and AI inference. The platform is named after David Harold Blackwell, a mathematician who specialized in game theory and statistics, and the first Black scholar inducted into the National Academy of Sciences.
The Blackwell platform is significant for Nvidia's future for several reasons:
Generational Improvements: Each new generation of Nvidia GPUs has brought substantial improvements in performance and efficiency. Blackwell is expected to continue this trend, with potential increases in raw compute for both graphics and general workloads.
AI and Inference: The Blackwell architecture is expected to feature native support for FP4 and FP6 numerical formats, primarily useful for AI inference. This could make Nvidia's GPUs even more attractive for AI workloads, an area where they already dominate.
Memory Advancements: The Blackwell platform is likely to introduce GDDR7 memory, providing a much-needed boost to memory bandwidth at all levels. This could lead to significant performance improvements.
Competitive Edge: Nvidia's dominance in the GPU landscape is well-established, with the company currently holding a significant market share. The Blackwell platform, with its expected improvements, could help Nvidia maintain this lead and stay ahead of competitors like AMD and Intel.
Market Expansion: The Blackwell platform could help Nvidia expand its reach in the AI and data center markets. With the increasing demand for AI-powered solutions, Nvidia's GPUs are well-positioned to benefit from this trend.
In summary, the Blackwell platform represents Nvidia's continued commitment to innovation and performance improvement. It is expected to bring notable advancements in various areas, from AI inference to memory bandwidth, which could help Nvidia maintain its competitive edge and expand its presence in key markets.

Nvidia's upcoming stock split has a ratio of 10-for-1, meaning that for every one share of Nvidia stock owned, investors will receive ten additional shares2. The stock split will take effect after the close of trading on Friday, June 7, 2024, and the shares will begin trading at the lower split-adjusted price on Monday, June 10, 2024.

The primary reason for the high demand for Nvidia's AI chips is their superior performance in handling complex computing tasks required for artificial intelligence applications. Nvidia's AI chips, also known as graphics processor units (GPUs), use parallel processing, breaking each computation into smaller chunks and distributing them among multiple cores in the chip. This allows for faster calculations and makes Nvidia's chips more energy-efficient and better at handling sophisticated computing demands compared to regular central processing units (CPUs).
Nvidia's GPUs were initially designed for video games, but their ability to perform calculations concurrently made them ideal for AI applications as well. The company has since expanded its offerings and now dominates the market for specialist AI chips, controlling more than 95% of the market. Nvidia's chips are now used in various industries, including automotive (for driver-assistance software), data centers, and healthcare, making them highly sought after by companies looking to leverage the power of artificial intelligence.