The company that began with a Denny’s meeting 30 years ago has evolved into the semiconductor corporation at the core of the artificial-intelligence revolution, placing it on the verge of being the first $1 trillion chip company.
Nvidia’s stock jumped more than 24% to an all-time high on Thursday as the company said the AI boom is translating into record sales, fuelling speculation that the new computer era is arriving sooner than previously imagined.
“When generative AI came along, it triggered a killer app for this computing platform that’s been in preparation for some time,” Nvidia CEO Jensen Huang said on Wednesday.
“This is going to be the largest change in technology that we’ve seen since the internet, there’s little doubt in my mind,” said Michael Sansoterra, chief investment officer of Atlanta’s Silvant Capital Management, which owns Nvidia. What excites investors today, he says, is that after months of speculation about AI, money is finally flowing.
This year, the stock has risen roughly 160%. The $183.8 billion in market valuation it added Thursday, the third highest for a U.S. business, brings its total to more than $938 billion, putting Nvidia on track to join Apple, Microsoft, Amazon, and Google parent Alphabet on the list of the world’s trillion-dollar companies.
Nvidia, on the other hand, is not a household name. It does not manufacture consumer electronics or internet services that the majority of the world’s population uses on a regular basis. However, its chips have grown vital behind the scenes. They are found in computers, automobiles, and robotics. With AI, its computers power new chatbots that generate coherent phrases, as well as a slew of other tools that the world’s largest corporations are racing to install.
For Huang, who often appears in public wearing his distinctive leather jacket, it is the culmination of a journey that began at a Denny’s in San Jose, Calif., when he debated with two colleague engineers how to improve computer graphics.
When Huang opened up Nvidia’s graphics-processing units to software developers to tamper with, allowing them to use their computer power for purposes other than making visuals appear better, there was little hoopla. On an earnings call, he told analysts that the move “will open up a whole new field called GPU computing.” None of the analysts sought additional information.
Developers quickly learned that the Nvidia processors were exceptionally good at the intricate calculations that underpin modern AI systems. They excel at performing multiple computations at the same time, which traditional computing engines (central-processing units) are less suitable for.
Nvidia’s first great success outside of videogames was bitcoin mining, where GPUs excelled. Nvidia surpassed semiconductor giant Intel in market value in 2020 as cryptocurrency values climbed, and its stock proceeded to rise to a new high of about $330 per share in late 2021.
When the crypto winter hit last year, Nvidia’s stock plummeted, only to be resurrected by the AI surge in recent months.
The AI mania also caused investors to forget about another Nvidia setback. In the midst of the pandemic, Huang undertook one of his most daring moves, attempting to acquire British chip-design expert Arm from SoftBank Group in a deal estimated at around $40 billion at the time. Last year, the two called off the merger when rivals objected to Nvidia acquiring a company that had built a reputation for functioning as a type of Switzerland for the chip industry, providing its ideas to everyone without favoring any one company.
According to analysts, the AI revolution offers far larger and longer-term promise for Nvidia than crypto. There is currently no competition that can equal it in terms of the variety of processors and software available for the computing-intensive demands of generative AI. UBS experts estimate that 10,000 Nvidia GPUs were required to construct OpenAI’s ChatGPT, the first significant generative AI system to be publically available.
What distinguishes generative AI is that there are concrete applications for the technology as organizations seek ways to leverage its capabilities, according to Stacy Rasgon, a chip-industry analyst at Bernstein Research. “It’s not crypto,” he pointed out. “I’d bet that in five or ten years, the overall opportunity will be significantly higher than it is today because we’ll have grown into it.”
The AI mania is also fueling investor interest in other chip businesses. Marvell Technology, which develops networking and data storage devices, had its stock rise 30% on Friday after management stated that AI-related revenue would more than double in the current fiscal year.
As the AI arms race between Amazon, Microsoft, and Alphabet’s Google heats up, Huang likened the current computer shift as a “iPhone moment,” referring to the rapid growth of smartphones after Apple debuted its hallmark phone in 2007. Data center operators are redesigning their facilities to make them more AI-friendly, he added, capitalizing on the benefits of Nvidia’s chips and software.
“We’re seeing incredible orders to retool the world’s data centers,” Huang said during an analyst conference call.
Nvidia’s chip demand has recently been so high that the company’s supply chain has struggled to keep up. According to Nvidia’s chief financial officer, the business has secured “substantially higher” supplies of chips for the second half of the year.
Nvidia designs but does not manufacture chips. From the start, the corporation adopted a business model that outsourced manufacture to contract chip producers, notably the world’s largest, Taiwan Semiconductor Manufacturing.
The changing computing landscape may pose issues for Intel, the major producer of data-center CPUs that serve as the backbone of business networks and the internet. Intel, whose stock sank more than 5% on Thursday, is making its own attempts to meet AI demand, including specialist AI chips and new CPUs that do AI calculations more efficiently.
“I think you’re seeing the beginning of, say, a 10-year transition to basically recycle or reclaim the world’s data centers and build it out as accelerated computing,” Huang said. Nvidia’s market capitalization is currently almost eight times that of Intel.
AI has become a battleground in the tech war between the United States and China. According to The Wall Street Journal, China’s main nuclear-weapons development institute has purchased Nvidia chips, among other things, despite being placed on a U.S. export blacklist in 1997.
The Biden administration enforced new license requirements on Nvidia’s most advanced semiconductor shipments to China last year, costing the business hundreds of millions of dollars in sales. Since then, Nvidia has begun to provide a new graphics-processing unit with requirements that allow it to be exported to China.
