“There is a war going on in AI, and Nvidia is the only weapon vendor,” said Raymond James managing director Srini Pajjuri in their market analysis. It is a concise and irrefutable statement of fact. Nvidia's unparalleled AI hardware is the engine that powers all players in the AI race.
This has been for a while, but 2023 was a significant year for Nvidia, characterized by meteoric growth and industry dominance, AI turned the company into a cultural phenomenon.
Investors have certainly noticed. Shares of Nvidia ( NASDAQ:NVDA ) are up more than 200% since the end of 2022, and today they are flirting near an all-time high north of $499 a share.
Amid this storm, the company's CEO, Jensen Huang, is becoming a fashion icon of sorts with his signature black leather jacket reminiscent of Steve Jobs' iconic turtleneck. Huang's appearance has become a symbol of tech culture, and his skills and hardware chips are so powerful that the Biden administration has declared their products a matter of national security. Suddenly, Nvidia's chips were a key chess piece in the multi-war technology war against China.
A company worthy of attracting the attention of world leaders, tech giants and fashionistas is the clear choice for 2023 Company of the Year.
Table of Contents
ToggleGrace Hopper Supercomputer
While there are many reasons behind Nvidia's meteoric rise, and while the company offers a variety of computer hardware, we've given top billing to one particular product in Nvidia's lineup: the DGX GH200 Grace Hopper Supercomputer.
The latest jewel in Nvidia's crown, the GH200 stands as the most powerful engine in the AI competition, and thanks to the combination of two of Nvidia's most powerful technologies: Grace CPUs and Hopper GPUs, it's in a league of its own. Ushering in a new era in AI infrastructure, the DGX GH200 packs unrivaled processing power through 256 integrated GH200 Superchips. This dwarfs the 8 GPUs that could be connected per system in previous offerings. The DGX GH200 is the only AI supercomputer that offers unified resource management.
It is also the first supercomputer to combine the Grace Hopper Superchip with Nvidia NVLink Switch, effectively eliminating limitations around distributed memory, complexity and parallelism, allowing all GPUs in the DGX GH200 system to work as one.
We are talking about a model so expensive and powerful that only a few in the industry have the right to use it. “Google Cloud, Meta and Microsoft will be the first companies to have access to (the GH200 supercomputer),” Huang said in a keynote address at Nvidia's Computex 2023. “This is a DGX GH200, a giant GPU.”
For reference, Nvidia's A100 chips used to train ChatGPT can combine to share 320GB of vRAM. These new chips take that limit to 144TB – 460 times that.
Already, key players in industries are eager to harness the power of the GH200 in real-world applications.
“Uber has achieved great results with OCI Compute and NVIDIA GPUs for Michelangelo, our machine learning platform,” said Kamran Zarghi, Uber's senior director of technology strategy. “Based on NVIDIA's Grace Hopper, we aim to achieve greater value for similar workloads in future instances.”
Meanwhile, AI service provider Kohere is looking forward to serving and serving LLMs quickly and economically in these new opportunities based on the NVadi GH200.
Continuous innovation
In artificial intelligence and high-performance computing (HPC), few names resonate as deeply as NVIDIA. In the year Founded in 1993 by Jensen Huang, Chris Malachowski and Curtis Prim, the company began its journey with a focus on 3D graphics for gaming and multimedia. In the year In 1999, Nvidia's creation of what many argue was the first consumer GPU for personal computers set the stage for reshaping the computing industry.
In the year In 2006, NVIDIA unveiled the CUDA architecture, opening up the parallel processing capabilities of GPUs to science and research. This development has paved the way for advances in AI and HPC, allowing GPUs to be used beyond traditional graphics processing.
The company's impact was further amplified in 2012 when it ushered in the modern era of AI by developing the AlexNet neural network, a cornerstone in deep learning and AI research.
Nvidia Research has made it a leader in the gaming world with its GeForce and RTX series of GPUs, but its AI has been incorporated into the company's values since the success of its A100 and H100 chips.
The latest iteration of Nvidia's AI products comes in the form of the GH200 Grace Hopper Superchip, a breakthrough CPU-plus-GPU hardware designed from the ground up for massive AI and HPC applications, combining the traditionally discrete workloads.
Named after the famous computer scientist and US Navy Rear Admiral Grace Hopper, the GH200 Grace Hopper Superchip is not only a technological achievement but also a catalyst for global scientific innovation.
Grace Hopper. Image: Wikimedia Commons
Powered by more than 40 AI supercomputers globally, the GH200 is at the heart of the latest hardware from giants such as Dell Technologies, Eviden, Hewlett Packard Enterprise (HPE), Lenovo and QCT.
In fact, just last week, HIV secured two of these supercomputers to further enhance its cloud computing services. The company seems to be following the money, from Bitcoin mining to AI as a safe bet for a profitable future.
“Simply put, we've been building the HIV GPU and AI business unit to be the ultimate AI infrastructure pick and shovel game,” said Hive Executive Chairman Frank Holmes. “As leaders in the technology sector, we believe we are undervalued as a company with both a very efficient Bitcoin mining operation and a thriving GPU computing business unit.”
Beyond the commercial field, several top supercomputing centers around the world, including Germany's Jülich Supercomputing Center and Japan's Joint High Performance Computing, have chosen the GH200 for their next-generation supercomputers.
These centers are demonstrating the potential of GH200s to facilitate important research addressing pressing scientific challenges, from climate change mitigation to pandemic response.
Award winning hardware
The Nvidia DGX GH200 Grace Hopper AI supercomputer stands out for its innovative CPU and GPU integration, an integrated integration that pushes beyond traditional hardware limitations. This combination is essential in this day and age where the demand for more powerful hardware to develop advanced software and services is increasing exponentially.
Beyond muscle, the GH200 is a testament to Nvidia's foresight in the industry's increasingly complex AI and computational challenges. It's not just about handling current workloads, it's about providing a gateway to future growth and helping drive generative AI and large-scale scientific research forward. The global adoption of high-end supercomputing centers indicates a shift towards more integrated and powerful AI and high-performance computing solutions.
And yes, beyond all that, Huang assured us, “it can even run Crysis.”
Edited by Ryan Ozawa.
