Jensen Huang, founder and CEO of Nvidiaspoke during the Nvidia GTC, i.e. the big conference organized by the company, and it did so in a anything but banal way. In front of a packed audience of listeners, who flocked to the Sap Forum in San Jose, California to listen to his words, Huang addressed the young students of Stanford University, wishing them “pain and suffering“, and then focusing on the most interesting innovations produced by the company.
What received particular attention was the announcement of the “Blackwell“, i.e. new and powerful chips that will power future technologies based on artificial intelligence.
Why did Nvidia’s CEO wish “pain and suffering” on young people?
Before going on stage to talk about technology and innovation, Huang took time to chat with the students at Business School Of Stanford, a prestigious American university that trains some of the best managers in the world. To them, the CEO of Nvidia wished “pain and suffering“, explaining that “greatness comes from character and character is not formed by brilliant people, but by people who have suffered“. The manager’s wish, therefore, is not to be understood as an invitation to suffering as an end in itself, but rather as a good opportunity to overcome it and learn from it.
Huang then went on to explain that, along with good managerial acumen, it is also essential to be realistic and have contained hopes: “people with very high expectations have very low resilience and this is fundamental to success. One of my strengths is that mine are very low“.
What Nvidia’s new chips will be like
Once the moment of discussion with the students was over, the CEO of Nvidia concentrated on presenting the innovations produced by the company. In particular, the main protagonists of his keynote speech were the new chips based onBlackwell architecture. Born thanks to Nvidia’s avant-garde research, the new chips B100 And B200 they have a lot inside them 208 billion transistorscompared to 80 previously H100. These also offer up to 4x and 30x better performance in training AI models compared to the previous Hopper architecture, as well as lower power consumption up to 25 times less.
Together with them, a “superchips” or GB200, which will combine two Blackwell GPUs with Grace CPUs. These innovations will bring great benefits to those who use them for their technologies. Previous chips consumed much more water and were much more energy intensive. Chips based on the Blackwell architecture, on the other hand, will allow training generative artificial intelligence models much more efficiently.
These news, for the moment, have been received quite indifferently by investors, who perhaps expected something more. Nvidia’s stock, protagonist of an unprecedented rally over the past year, fell slightly after the stock market in New York, around 3%, before rising again.