In 2009, when Nvidia held his first developer conference, the event was something of a science grant. Dozens of academics filled a San Jose, California, hotel decorated with white poster boards of computer research. Jensen Huang, the chief executive of the chip maker, roamed the floor like a judge.
This year the developer conference of Nvidia is much different.
More than 25,000 people were expected to meet on Tuesday around the event, known as Nvidia GTC. The crowd filled a National Hockey League -Arena to hear a speech about the future of artificial intelligence of Mr Huang, who was nicknamed 'Ai Jesus'. NVIDIA, the world's leading developer of AI chips, has also packed San Jose in the neongroene and black colors of the company, closing the streets of the city and sending hotel prices that are as high as $ 1,800 per night.
A who's who of industrial leaders, including Michael Dell, the Chief Executive of Dell Technologies; Jeffrey Katzenberg, the co-founder of Dreamworks and Wnddrco, a risk capital company; and Bill McDermott, the Chief Executive of Serviceenow.
“GTC is jampacked,” Mr. Huang said, started the conference on Tuesday morning. “The only way to keep more people on GTC is that we have to grow San Jose.”
The transformation of the Nvidia conference from an academic event to the super bowl of AI – a week -long showcase of robots, large language models and autonomous cars – is symbolic for the metamorphosis of the company. Because AI Mainstream has become, customers have called for the graphic processing units of Nvidia, or GPUs, the powerful chips that help create technology. That has pushed the chip maker to a rating of almost $ 3 trillion, an increase of $ 8 billion in 2009.
Nevertheless, the climb of Nvidia has asked questions. Generative AI, who can answer questions, make images and write code, is celebrated because of the potential to improve companies and create trillions of dollars in economic value. Microsoft, Amazon, Google, Meta and others spend hundreds of billions of dollars to realize that idea.
But the expenditure has led to concern about Wall Street and Silicon Valley about whether AI will earn enough money to justify its amazing costs. And the process of technology can be increased by newcomers such as Deepseek, a small Chinese company that has made an advanced AI system with a small part of the NVIDIA chips that other companies used. (In January, when investors realized what Deepseek had done, Nvidia lost $ 600 billion in value in one day.)
At Nvidia GTC, Mr. Huang tried to reassure people that AI will deliver the potential. He spoke about how AI systems give more accurate answers by making thousands of more calculations for every request. The result is AI systems that are more useful and able to offer services that people want to pay for, such as AI agents, who can perform autonomously tasks, such as shopping for groceries.
But those calculations mean that the demand for more powerful computers will increase. Enter Nvidia's next line of chips.
At the end of 2026, Mr. Huang said, Nvidia will release his next generation of GPUs, called Rubin. The chips can be packed in a supercomputer with four times as many GPUs as today's model. The system will be 14 times more powerful than the supercomputer of today and uses less power.
Mr. Huang bets that those chips will explain a considerable share of the more than $ 1 trillion that analysts predict annually to data centers in 2028. “The more you buy, the more you save,” said Mr. Huang.
The Rubin chip is crucial for the in the forefront of Nvidia's who is at the forefront of AI, the company stands for challenges while its customers, including Amazon, Google and Meta, make their own AI chips. And Nvidia's chips should also change as AI companies try to get better performance from their AI models.
“The Jus Train comes to a screaming stop when cloud companies stop spending,” says Patrick Moorhead, founder of Moor Insights & Strategy, a technical research agency. He said that Mr. Huang's product card “more trust” that Nvidia's products demands a lot, as long as AI systems create business opportunities to justify their costs.
In addition to sketching the coming products from Nvidia, Mr. Huang said that General Motors has committed himself to use Nvidia's AI tools to design cars and plan the car factories. He said that Nvidia also works with Google DeepMind and Disney research into a software system to improve the precision of robots, which he demonstrated on stage with a waist-high robot in the form of Pixar's Wall-e.
Mr Huang's ability to order a crowd is reminiscent of Apple's Steve Jobs. In anticipation of large corporate events, the Apple-MEDE founder rehearsed his speeches about a new iPod, iPhone or iPad before he entered the stage on the thunderous applause and his comments seemed to make as if they were not being written.
Mr. Huang, 62, prepares a lot in the same way for Nvidia GTC. Two months prior to the event, he works with the product divisions of the company to identify what to announce, said Greg Estes, Vice -President of Corporate Marketing of Nvidia. Mr. Huang also works with the marketing team to develop slides and demonstrations to show on stage, creating list signs and checking facts that he can quote.
But Mr. Huang never writes a speech, Mr Estes said. When he enters the stage in his characteristic black leather jacket, he speaks extremely. A speech planned for 90 minutes can take more than two hours.
“Sometimes a mistake will happen and he will say,” You know, we are not rehearsing, “said Mr. Estes.” He's not joking.
Nvidia GTC used to be the GPU technology conference, named after the graphic processing units or GPUs. The event, which was designed to encourage developers to use the company's chips, included an research stop in which academics set up poster boards describing how they used the components for computer use. Mr. Huang spoke with those present about what they did with the chips and over the years often heard them used to develop AI
David Cox, who presented research during an early conference as a professor in Harvard, said that most people present treated the academics as 'this weird little footnote'. But he said that Mr. Huang and other Nvidia leaders took them seriously.
“They seemed to understand that we had something here,” said Mr. Cox, who is now the vice president of AI models at IBM Research.
In 2014, Mr Huang began to devote most of his speech to the conference to the way in which Nvidia chips could be used for machine learning and AI gaming developers, who used GPUs to make video game graph and had long been the heart of the company's case, were angry by the shift.
“They were from:” What is this shiny new thing? “, Said Naveen Rao, the Chief AI officer at Databricks, who offers software tools for storing and analyzing large amounts of data.” We had something like: “No. This is the sea change. “
Mr. Huang Wedde that AI would stimulate the next big boom from Tech and that GPUs would be essential. In 2016, Nvidia developed a supercomputer full of its chips and delivered it to OpenAi, an AI -Lab. A little more than six years later, OpenAi released the chatgpt -chatbot and unleashes an AI racing.
(The New York Times has sued OpenAi and his partner, Microsoft, for infringing the copyright of news content with regard to AI systems. OpenAi and Microsoft have refused the claims.)
Since then, the finances of Nvidia have risen. The company, which was founded in 1993, increased its annual profit by more than 1500 percent in a period of two years to $ 72.88 billion last year of $ 4.37 billion in tax 2023.
“Jensen has become the CEO of Celebrity that he always wanted to be,” said Mr. Rao. “It is one day to the next one day to make it into the making because he conquered AI”
