Nvidia announces launch of GB200 Blackwell AI chip later this year | Wilnesh News

Nvidia announced on Monday the launch of a new generation of artificial intelligence chips and software for running artificial intelligence models. The announcement was made at Nvidia’s developer conference in San Jose, as the company seeks to solidify its position as a go-to supplier for artificial intelligence companies.

Since OpenAI’s ChatGPT kicked off the artificial intelligence craze in late 2022, Nvidia’s stock price has increased fivefold and total sales have more than tripled. Nvidia’s high-end server GPUs are critical for training and deploying large AI models.company likes Microsoft and Yuan Billions of dollars were spent buying wafers.

The new generation of AI graphics processor is named Blackwell. The first Blackwell chip, called the GB200, will ship later this year. Nvidia is luring customers with more powerful chips to spur new orders. For example, companies and software makers are still scrambling to get their hands on current-generation “Hopper” H100 and similar chips.

“Hopper is great, but we need bigger GPUs,” Nvidia CEO Jen-Hsun Huang said Monday at the company’s developer conference in San Jose, California.

The company also launched revenue-generating software called NIM that will make it easier to deploy artificial intelligence, giving customers another reason to stick with Nvidia chips amid growing competition.

Nvidia executives say the company is becoming less of a profit-seeking chip supplier and more of a platform provider, like Microsoft or Apple, on which other companies can build software.

Manuvir Das, Nvidia’s corporate vice president, said in an interview: “The commercial product that can be sold is the GPU, and the software is all about helping people use the GPU in different ways.” “Of course, we still do that. But what has really changed is, We now truly have a business software business.”

Das said Nvidia’s new software will make it easier to run programs on any Nvidia GPU, even older GPUs that may be more suitable for deployment but not suitable for building artificial intelligence.

“If you’re a developer and you have an interesting model that you want people to adopt, if you put it into NIM we’ll make sure it runs on all of our GPUs so you can reach a lot of people. ,” Das said.

Nvidia updates its GPU architecture every two years to achieve significant performance improvements. Many of the AI ​​models released last year were trained on the company’s Hopper architecture, used by chips like the H100, which is due to be released in 2022.

Nvidia says Blackwell-based processors like the GB200 offer AI companies a huge performance upgrade, with 20 petaflops of AI performance compared to the H100’s 4 petaflops. Nvidia said the additional processing power will allow AI companies to train larger, more complex models.

The chip includes what Nvidia calls the “Transformer Engine,” designed to run transformer-based artificial intelligence, one of the core technologies underpinning ChatGPT.

Blackwell GPUs are large, combining two separately manufactured dies into a single die made from British Semiconductor. It will also be available as a full server called GB200 NVLink 2, combining 72 Blackwell GPUs and other Nvidia components designed to train AI models.

Nvidia says the system can deploy 27-megapixel models. This is much larger than even the largest models such as GPT-4, which reportedly has 1.7 trillion parameters.Many artificial intelligence researchers consider larger models with more parameters and data Can unlock new abilities.

Nvidia didn’t provide a cost for the new GB200 or the system it’s powered by. Nvidia’s Hopper-based H100 costs between $25,000 and $40,000 per chip, with the entire system costing as much as $200,000, according to analyst estimates.