Image: NVIDIA
After his opening speech and the flurry of announcements at its development and innovation conference GTC, Jensen Huang, the company’s CEO, took the time to answer questions from reporters on the various topics. However, of course above all the announcements hovered the phrase Generative AI, which has replaced with all its might the Metaverse trend of The last two years, and rightly so.
As you remember, NVIDIA is responsible for the bulk of the GPUs that the main generative AI systems like those of OpenAI want. However, despite Nvidia’s series of announcements on the subject, Huang said that today the field is only responsible for a single-digit percentage of the company’s total revenue in the last 12 months or a “tiny” part as he said. However, when Huang was asked by a reporter what the size of the field would be 12 months from now, Huang said it was difficult to accurately estimate but it would be “pretty big,” as he said the whole infrastructure-as-a-service field is expanding even further. the company’s business model.
The one who will assist Nvidia in its efforts in this field is the Israeli startup company Quantum Machines. As a reminder, yesterday (Tuesday) Nvidia announced a new system called NVIDIA DGX Quantum, which was developed with the Israeli startup, and combines classic GPUs (Grace Hopper) with quantum computing. But what exactly does this collaboration mean and will we see more Nvidia collaborations with Israeli companies? We took advantage of Huang’s conversation to ask him a few questions.
Gigtime: We see quite a few developments in the fields of quantum computing at GTC. How do the latest developments affect the field?
Huang: “Let’s talk about quantum computing for a moment. The quantum research and development community is really vibrant around the world, and there are many different quantum methods. Although we are a decade to two decades away from widespread use of quantum systems, in the near term, there are many interesting things that need to be solved. One, if You are trying to build a revolutionary computing system, you still need to use the most powerful computer available now to check the reference point (the “control set”), to know if the quantum computer is producing the correct answers.
The second reason is because quantum computers require a different set of algorithms. You can’t take algorithms developed for classical computing and think it will work for quantum. They solve different problems, and the algorithms used are different, so NVIDIA is the ideal platform for quantum algorithm research.
The third reason is that the quantum computer will definitely be connected to a classical computer, and this classical computer will probably be an accelerated computing system like NVIDIA’s. That’s why we created the Open CUDA Quantum programming model and it has been adopted by researchers and developers all over the world.
The last reason is that the quantum computer, in order to be efficient, must have many qubits so that we can perform error correction on the very large number of qubits, to recover the signal from the noise. And the error correction needed to perform quantum error correction is essentially a supercomputer, and NVIDIA’s accelerated computer is the ideal solution for that. So whether it’s for quantum research, algorithmic research, or a classical quantum programming model, hooking up an NVIDIA DGX system with a very fast link between that computer and the quantum computer allows us to do the work I’ve just described.
We are really passionate about quantum research, and there is a lot of important work around the world – labs, companies, universities, and NVIDIA Quantum is the platform that almost everyone uses.”
Gigtime: It was precisely for these needs that you collaborated with the Israeli startup Quantum Machines. Do you have collaborations with other Israeli startups in the pipeline?
Huang: “Definitely. Israel is still the highest concentration of startups on the planet in such a small place. And it’s amazing, the vitality, the imagination, the enthusiasm to do innovative things. We have a very large development center there, with about 3,000 engineers, who came to us as part of the merger With Mellanox, one of the great technology companies in the world. I look forward to coming to Israel quite soon. We are really happy to have a large site there and so many NVIDIA families there, and great partnerships with Israeli startups.”