The technology that gets the most attention is an unpretentious piece of silicon that goes hand in hand with the chips that drive video game graphics.
San Francisco — San Francisco (AP) —
The hottest thing in technology is the unpretentious chunks of silicon that are closely associated with the chips that power video game graphics. This is an artificial intelligence chip specifically designed to make building AI systems like ChatGPT faster and cheaper.
Chips like this have suddenly taken center stage in the AI revolution that some experts believe has the potential to reshape the tech sector, and possibly the world with it. Nvidia’s stock, a leading designer of AI chips, rose to around 7.5% last Thursday after the company predicted a significant increase in revenue as analysts said sales of its products would surge. Soared 25%. The company’s value briefly exceeded $1 trillion on Tuesday.
What is an AI chip in the first place?
It’s not an easy question to answer. “There is no fully agreed upon definition of an AI chip,” said Hannah Domen, research her analyst at the Center for Security and Emerging Technologies.
However, the term generally includes computing hardware specialized for processing AI workloads. For example, AI systems can be “trained” to tackle problems that are difficult for traditional computers.
Origin of video games
Three entrepreneurs founded Nvidia in 1993 to push the boundaries of computational graphics. Within a few years, the company developed a new chip called a graphics processing unit (GPU). It dramatically speeds up both video game development and play by performing multiple complex graphics calculations at once.
This technology, formally known as parallelism, will be key to the development of both games and AI. Two of his graduate students at the University of Toronto used GPU-based neural networks to identify photographic images with much lower error rates than their competitors, winning his prestigious 2012 AI competition called ImageNet. Did.
The win boosted interest in AI-related parallelism, opened up new business opportunities for Nvidia and its rivals, and gave researchers powerful tools to explore the frontiers of AI development. .
Latest AI Chip
Eleven years later, Nvidia has become a leading supplier of chips for building and updating AI systems. His one of the company’s most recent offerings, the H100 GPU, contains 80 billion transistors. That’s about 13 million more than the latest high-end processors for Apple’s laptop MacBook Pro. Naturally, this technology does not come cheap. At one online retailer, the H100 is priced at his $30,000.
Nvidia doesn’t manufacture these complex GPU chips themselves, which would require huge investments in new fabs. Instead, it relies on Asian chip foundries such as Taiwan Semiconductor Manufacturing Company and South Korea’s Samsung Electronics.
Some of the biggest customers for AI chips are cloud computing services operated by Amazon and Microsoft. By renting out AI computing power, these services enable small businesses and groups that could not afford to build their own AI systems from scratch to use cloud-based tools for everything from drug discovery to customer management. help with various tasks. .
Other uses and competition
Parallelism has many uses beyond AI. For example, a few years ago there was a shortage of Nvidia graphics cards because most of them were bought up by crypto miners who set up banks of computers to solve difficult math problems for Bitcoin rewards. is. This problem disappeared when the cryptocurrency market collapsed in early 2022.
Analysts say Nvidia will inevitably face tougher competition. One of his potential rivals is Advanced Micro Devices, a company that has already put him up against Nvidia in the market for computer graphics chips. AMD recently took steps to beef up its AI chip lineup.
Nvidia is based in Santa Clara, California. Co-founder Jensen Huang will continue to serve as president and CEO of the company.
