All the Nvidia news announced by Jensen Huang at Computex
Jensen Huang wants to bring generative artificial intelligence to every data center, Nvidia’s co-founder and CEO said during Computex in Taipei today. During the speech, Huang’s first public speech in nearly four years, he said he made a number of announcements, including chip release dates, the DGX GH200 supercomputer and partnerships with major companies. Here is all the news from the two-hour keynote.
1. Nvidia’s GForce RTX 4080 Ti GPU for gamers is now in full production and is being produced in “large quantities” with partners in Taiwan.
2. Huang announced the Nvidia Avatar Cloud Engine (ACE) for Games, a customizable AI model foundry service with pre-trained models for game developers. It will give NPCs more character through AI-powered language interactions.
3. The Nvidia Cuda computing model now serves four million developers and more than 3,000 applications. Cuda saw 40 million downloads, including 25 million last year alone.
4. Full volume production of the GPU server HGX H1[ads1]00 has begun and is being produced by “companies all over Taiwan,” Huang said. He added that it is the world’s first computer to have a transformer motor in it.
5. Huang called Nvidia’s 2019 acquisition of supercomputer chipmaker Mellanox for $6.9 billion “one of the biggest strategic decisions” it has ever made.
6. Production of the next generation of Hopper GPUs will start in August 2024, exactly two years after the first generation started production.
7. Nvidia’s GH200 Grace Hopper is now in full production. The superchip reinforces 4 PetaFIOPS TE, 72 Arm CPUs connected with chip-to-chip link, 96 GB HBM3 and 576 GPU memory. Huang described as the world’s first accelerated computer processor that also has a giant memory: “this is a computer, not a chip.” It is designed for highly resilient data center applications.
8. If Grace Hopper’s memory is not enough, Nvidia has the solution – DGX GH200. It is made by first connecting eight Grace Hoppers together with three NVLINK switches, then connecting the 900GB pods together. Finally, 32 are assembled, with another layer of switches, to connect a total of 256 Grace Hopper chips. The resulting ExaFLOPS Transformer Engine has 144 TB of GPU memory and acts like a giant GPU. Huang said Grace Hopper is so fast that it can run the 5G stack in software. Google Cloud, Meta and Microsoft will be the first companies to access the DGX GH200 and will research its capabilities.
9. Nvidia and SoftBank have partnered to introduce the Grace Hopper superchip in SoftBank’s new distributed data centers in Japan. They will be able to host generative AI and wireless applications in a common multi-tenant server platform, reducing costs and energy.
10. The SoftBank-Nvidia partnership will be based on the Nvidia MGX reference architecture, which is currently used in collaboration with companies in Taiwan. It provides system manufacturers with a modular reference architecture to help them build more than 100 server variants for AI, accelerated computing and omniverse use. Companies in the partnership include ASRock Rack, Asus, Gigabyte, Pegatron, QCT and Supermicro.
11. Huang announced the accelerated network platform Spectrum-X to increase the speed of Ethernet-based clouds. It includes the Spectrum 4 switch, which has 128 ports at 400GB per second and 51.2T per second. Designed to enable a new type of Ethernet, Huang said, the switch was designed end-to-end to do adaptive routing, isolate performance and do in-fabric computing. It also includes the Bluefield 3 Smart Nic, which connects to the Spectrum 4 switch to perform congestion control.
12. WPP, the largest advertising agency in the world, has partnered with Nvidia to develop a content engine based on Nvidia Omniverse. It will be able to produce images and video content to be used in advertising.
13. Robot platform Nvidia Isaac ARM is now available to anyone who wants to build robots, and is full-stack, from chips to sensors. Isaac ARM starts with a chip called Nova Orin and is robotics’ first full-reference stack, Huang said.
Largely thanks to its prominence in AI computing, Nvidia’s stock has soared over the past year, and it currently has a market capitalization of around $960 billion, making it one of the most valuable companies in the world (only Apple, Microsoft , Saudi Aramco, Alphabet and Amazon are ranked higher).
China business in limbo
China’s AI firms are no doubt paying close attention to the state-of-the-art silicon Nvidia is bringing to the table. Meanwhile, they likely fear another round of US chip bans that threaten to undermine their progress in generative AI, which requires significantly more computing power and data than previous generations of AI
The US government last year restricted Nvidia from selling its A100 and H100 graphics processing units to China. Both chips are used to train large language models such as OpenAI’s GPT-4. The H100, the latest generation chip based on the Nvidia Hopper GPU computing architecture with its built-in Transformer Engine, is seeing particularly strong demand. Compared to the A100, the H100 is capable of providing 9x faster AI training and up to 30x faster AI inference on LLMs.
China is obviously too big a market to miss. The ban on chip exports will cost Nvidia an estimated $400 million in potential sales in the third quarter of last year alone. Nvidia thus resorted to selling China a slower chip that meets US export control rules. But in the long term, China is likely to look for more robust alternatives, and the ban serves as a poignant reminder for China to achieve self-reliance in key technology sectors.
As Huang recently said in an interview with the Financial Times: “If [China] can’t buy from … USA, they just want to build it themselves. So the US needs to be careful. China is a very important market for the technology industry.”
