Shulou(Shulou.com)11/24 Report--

According to news in the morning of March 8, Beijing time, Nvidia has been leading the revolution in computer graphics technology for about 25 years, and has become a popular brand among gamers in the process.

Nvidia dominates the graphics processing unit (GPU) market, which it entered in 1999 with a product called GeForce 256. Despite the recent decline in its business, the gaming industry generated more than $9 billion in revenue for Nvidia last year.

But Nvidia's latest earnings show a new phenomenon in the GPU business. The technology is now at the center of the vigorous development of artificial intelligence.

Jensen Huang, Nvidia's chief executive, has said: "We discovered a long time ago, about a decade ago, that this way of making software can change everything. We changed the company from the bottom to the top. Every chip we produce is focused on artificial intelligence technology."

As the power engine behind large language models (LLMs) such as ChatGPT, Nvidia's early investment in artificial intelligence has finally paid off. This will help ease the blow to the semiconductor industry caused by trade tensions and global chip shortages.

From March 20 to 23, Nvidia will hold its annual GTC developer conference, when artificial intelligence will be one of the main focuses of the event. Prior to the event, the media had a conversation with Huang Renxun at Nvidia headquarters in Santa Clara, Calif., and the two sides discussed the company's central role in the generative artificial intelligence explosion.

When asked whether Nvidia's return was luck or foresight, Huang Renxun said: "We just believe that one day something new will happen and the rest needs some chance." this is not prescience. the foresight is to speed up the calculation. "

GPU is Nvidia's main business, accounting for more than 80 per cent of its revenue. They are usually sold as cards plugged into the PC motherboard, adding computing power to the central processing unit (CPU) made by companies such as AMD and Intel.

Today, technology companies competing with ChatGPT are openly bragging about how many Nvidia A100s they own for about $10,000. Microsoft says the supercomputer developed for OpenAI uses 10000 of the device.

"it's very easy to use their products and add more computing power, which is now the hard currency of Silicon Valley," said Vivek Arya, a semiconductor analyst at Bank of America Securities.

Huang Renxun showed off the company's next-generation system, called H100, which is now being shipped.

Huang Renxun, holding a server motherboard, said: "what is really amazing about the new system is this new processing method called the converter engine. The engine is the T in GPT, the generative pre-training converter. This is the first computer in the world designed to process converters on a huge scale. As a result, large language models will be much faster and much more cost-effective."

Huang Renxun revealed that he personally delivered the new system to OpenAI, the developer of ChatGPT. ChatGPT is the world's first artificial intelligence supercomputer.

Today, Nvidia is one of the 10 most valuable technology companies in the world, with a market capitalization of nearly $600 billion. It has 26000 employees and a new polygon-themed headquarters. It is also one of the few Silicon Valley giants whose founders have been at the helm for 30 years.

Huang Renxun, 60, emigrated to the United States from Taiwan Province in his childhood and studied engineering at Oregon State University and Stanford University. In the early 1990s, Huang Renxun met with engineers Chris Malachovsky (Chris Malachowsky) and Curtis Prim (Curtis Priem) in a restaurant to talk about realizing the dream of 3D graphics on personal computers.

In 1993, three people founded Nvidia in a mobile apartment in Fremont, Calif., inspired by the name NV in "next version" (Next Version) and the jealous Invidia in Latin. The name of the company reflects that they want to greatly speed up the computing speed, which is the envy of everyone, so they chose the envious green eyes as the logo of the company.

"they were one of dozens of GPU manufacturers at the time. They were the only ones that really survived, basically only Nvidia and AMD stayed because Nvidia worked very well with the software community and developers," Aya said.

Huang Renxun's ambition and seemingly impossible investment preference have pushed the company to the brink of bankruptcy, and this has happened more than once.

"every company makes mistakes, and I make a lot of mistakes," said Huang Renxun, who was named one of time magazine's most influential people in 2021. Some of these mistakes put the company at risk, especially at the beginning, because we were very small and our competitors were very large companies, and we were trying to invent a whole new technology. "

In the early 1910s, for example, Nvidia made a foray into smartphones with its Tegra processors, but the attempt ended in failure. The company subsequently withdrew from the field.

In 1999, after laying off most of its staff, Nvidia announced what it claimed to be the world's first official GPU--GeForce 256. This is the first programmable graphics card that allows users to customize shading and lighting effects. By 2000, Yingwei had become the exclusive graphics card supplier of Microsoft's first Xbox game console. In 2006, the company made another big bet by releasing a software toolkit called CUDA.

"for 10 years, Wall Street investors have been asking Nvidia, 'Why are you making this investment? no one is using this technology," said Brian Katanzaro, vice president of applied deep learning research at Nvidia. "in our market capitalization, investors value the technology at $0." Katanzaro joined Nvidia in 2008 and is one of the few employees of the company who work in artificial intelligence. Today, the company has thousands of employees working in this field.

Until 2016, 10 years after the birth of CUDA, people suddenly realized that this was a completely different way of writing computer programming. Katanzaro said: "its ability to speed up is revolutionary, and then it has produced a breakthrough in the field of artificial intelligence."

Although artificial intelligence is developing rapidly, games are still Nvidia's main business. In 2018, the company used its expertise in artificial intelligence to make another major leap forward in graphics. The company launched GeForce RTX based on its experience in artificial intelligence.

Huang Renxun said: "in order for us to take computer graphics and video games to a new level, we have to reinvent and subvert ourselves and completely change what we have invented. We invented this new way of dealing with computer graphics, ray tracing, which basically simulates the path of light and simulates everything with generative artificial intelligence. So, we calculate one pixel, r and then use artificial intelligence to imagine the other seven pixels. "

From the beginning, Huang Renxun has been committed to making Nvidia a chip company without a fab, or a company that designs products but outsources production to other companies that own chip factories. Nvidia reduces capital expenditure by outsourcing the special cost of making chips to Taiwan Semiconductor Manufacturing Corporation of China.

This reliance on third-party manufacturers worries investors.

C.J., an analyst at Evercore. "if I were a shareholder in Nvidia, it would be the only thing that keeps me awake at night," said C.J. Muse. "it's not just a risk for Nvidia, it's a risk for AMD, Qualcomm and even Intel."

TSMC said it would invest $40 billion to build two new chip manufacturing plants in Arizona. Huang Renxun said that Nvidia will definitely use TSMC's Arizona plant to make its chips.

Then there is the question of requirements and how many of GPU's new use cases will continue to grow. Nvidia saw a surge in demand for its products when the cryptocurrency mining industry took off, as GPU became the core of effective competition in the market. The company even created a simplified GPU specifically for cryptocurrencies. But with the collapse of the cryptocurrency, Nvidia experienced an imbalance between supply and demand.

Aya said: "this has caused problems because cryptocurrency mining has always been a boom or bust cycle. Game graphics cards are out of stock, prices are raised, and then when the cryptocurrency mining boom collapses, there will be a big collapse in the game."

Nvidia caused a major price shock among some game players last year because its new 40-series GPU was priced much higher than the previous generation. Now, due to too much supply, game revenue fell 46% in the most recent quarter compared with the same period last year.

In addition, as more and more technology giants begin to design their own custom chips, Nvidia is facing increasing competition. Tesla and Apple are doing the same, as are Amazon and Google.

"the biggest problem for them is how to stay ahead," Aya said. Their customers could also be their competitors. Microsoft can try to design these things internally. Amazon and Google are already designing chips internally. "

Huang Renxun said that this kind of competition is a good thing.

"the amount of energy the world needs in data centers will grow, which is a real problem for the world," he said. the first thing we should do is: every data center in the world, no matter what you decide to do, accelerate everything you can for the benefit of sustainable computing. "

In the auto market, Nvidia is developing self-driving technology for Mercedes-Benz and other companies. The company's systems are also used to power robots in Amazon warehouses and run simulations to optimize the flow of millions of packages every day.

Huang Renxun described the system as "the whole universe". "We now have more than 700 customers testing it, from the auto industry to logistics warehouses to wind turbine plants," he said. "it may represent one of the greatest containers of all Nvidia's technologies: computer graphics, artificial intelligence, robots and physical simulations. I have high hopes for it."

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.