Nvidia's CEO on AI warnings: "I am a responsible optimist"

Nvidia's CEO on AI warnings: "I am a responsible optimist" OpenAI, Anthropic, and Google may be driving the AI revolution with well-polished chatbots. But behind the scenes, most of it is driven by chips from a single company: Nvidia, now the world's most valuable stock market company. They fill the new AI data centers – but have also been found in Russian drones in Ukraine. – It is impossible for us to know how they are being used, says Nvidia's CEO Jensen Huang to the DN. The AI boom began in some ways in a Canadian bedroom, in 2012. Alex Krizhevsky, a promising doctoral student at the University of Toronto, was working on an experiment that many saw as fruitless. It involved the programming technique of "artificial neural networks," which can be simplified as a way to mimic the brain's nerve cells. At that time, it was ice cold. Few believed it would lead to any sensible AI development. But Alex Krizhevsky wanted to give it a chance. He built a system for image recognition – software that can tell if an image depicts a cycle or a cat – and began training it by analyzing tens of thousands of images. The computer that did it sat in his bedroom, humming and drawing power. To manage this, he did something unusual: He bought hardware from a company most known for selling graphics cards to computer gamers: Nvidia. The graphics card chips could also be used for that kind of calculations, he reasoned. The result was an explosion. When AlexNet, as the system was called, participated in an annual competition for enthusiasts in machine learning, it completely outclassed all others. No one came close to its ability to guess what an image depicts. The tech world lost its breath and drew two conclusions: Partly artificial neural networks were the way to artificial intelligence. Partly it was the chips from Nvidia that should be used, not processors of the type found in regular computers. 14 years later, Nvidia's founder Jensen Huang stands outside the München brewery in Södermalm in Stockholm. As always dressed in a leather jacket and exclusive sneakers, he is now beside Volvo Cars CEO Håkan Samuelsson and a well-polished Volvo EX60. Jensen Huang is talking about exactly the same thing that AlexNet was about: Image recognition using Nvidia chips. The difference is that he is no longer talking about experiments to recognize cat photos, but about how self-driving cars can use it to understand their surroundings. Another difference is that Nvidia, after the success with AlexNet, has become the world's most valuable stock market company, ahead of Apple, Alphabet's parent company Google, and Microsoft. Worth over five thousand trillion dollars. And Nvidia has achieved this by betting everything on exactly what Alex Krizhevsky tested: computer chips, with the same technical roots as graphics cards for computer games, for AI calculations. It is said that those who made the most money in the gold rush in Klondike were not those who looked for gold, but those who sold the shovels. Nvidia sells the shovels to the AI gold rush. – All companies will become AI companies, says Jensen Huang. Earlier in the day, he celebrated with Marcus Wallenberg that Wallenberg's family is building a Swedish AI supercomputer with Nvidia chips for Saab, AstraZeneca, and Ericsson's computing needs. He spoke in particular about how he sees the future for pharmaceutical companies like AstraZeneca. – All these companies will transform into companies where robots discover new medicines. The demand is enormous as AI companies are launching their gigantic data centers around the world and want to fill them with Nvidia's machines. In some periods, it has not been a matter of wanting to buy the products, but of whether Nvidia can deliver fast enough. This means that Jensen Huang and Nvidia have extreme power: Those who get to buy their chips can be part of the AI journey. Those who are left behind risk being in the rear. The question is bigger than competition between companies – for the USA it is a hot potato which AI chips China will buy. The goal is that the most powerful do not get Peking. That is why the world stock markets and the AI world were frightened when Chinese Deepseek appeared recently with impressive AI performance, despite having been trained with simpler chips. In other cases, access becomes literally a matter of life and death. On July 6th this year, a Russian drone flew over the Ukrainian city of Zaporizhzhia. It was alleged to have been aimed at an gas station, possibly a gas tank that was there, according to the New York Times. When it smashed into a wall and exploded, three people were killed. The drone was controlled by an Nvidia chip, which according to Ukraine likely meant that it functioned independently, could navigate based on what it saw on the ground, and make decisions about what to attack. A literal killing robot. To Nvidia's defense: The computer on the drone was a consumer variant that can be bought almost anywhere. It is not exported to Russia, but it is not difficult to get through indirect routes. How does he view what happened? First, he pulls back. – We sell these computer systems for general purposes. They are used for everything from shopping to logistics systems and self-driving cars. And sometimes drones. It is impossible for us to know how they are distributed and used, he replies. Then he sharpens his tone, leaning more towards the idea that an arms race is the right way. – The technology will be used in autonomous systems like drones. We must ensure that the good side has all autonomous systems and all drone technology that others have. I cannot know all the details of the case you mention, but I am there because Nvidia's technology is used for autonomous systems for people's safety and to create a safer world. Jensen Huang Born in Taiwan in 1963 (with the name Jen-Hsun), but his parents sent him to the USA when he was just nine years old. He was placed there in what the family believed to be a prestigious boarding school, but which turned out to be a religious upbringing institution for young boys. He studied engineering in electrical engineering, learned to design computer chips, and worked with them for several Silicon Valley companies before founding Nvidia in 1993. The target audience was computer gamers who wanted the best graphics in the games possible and were willing to pay for special hardware for it. When the AI boom started, largely built on Nvidia's hardware, the company's stock surged – and with it Jensen Huang's fortune. Today, he is worth about 190 billion dollars, making him the world's seventh richest person. It is another of Jensen Huang's characteristics: He strongly opposes any hint of braking in AI development and dismisses warnings about the dangers of AI development. This is while many others have begun to advocate for a slowdown and warn about the consequences of artificial superintelligence: A row of dissenters from the big AI companies. In fact, the leaders of some of them, although one can sense a PR strategy in the background, have also all to gain from rapid AI development. Nevertheless, they are opening up to more regulation. You are known for dismissing doomsday scenarios and seem to be much more optimistic than others in the industry. Why did you come to such a different conclusion? – I am a responsible optimist, says Jensen Huang. – If we are not optimistic about the future, we should just stop. We continue because we are optimistic, but we must do it responsibly. It is not as different as we are talking about now, he says, first towards the Volvo CEO, then towards the electric car. One can argue that the automotive industry is highly regulated, with a number of requirements for safety and environmental considerations. Jensen Huang relies more on the companies' self-imposed responsibilities. – As CEOs, we must say: If the product we create is not safe, then do not deliver it. It is very simple. Challenged from all sides Nvidia's dominance in AI chips is not unchallenged. On the contrary, many of the big AI companies are trying to become less dependent on Jensen Huang's company. Google relies largely on its own chips. Amazon is developing its own variants. Anthropic is supposed to be considering starting to design AI chips according to media reports, and OpenAI claimed last summer that a chip they developed themselves, called Jalapeño, surpasses Nvidia.