Nvidia’s market capitalisation soared to $3.01 trillion surpassing Apple ($3 trillion) on Wednesday, making it the second most valuable company in the world after Microsoft ($3.15 trillion). Nvidia's stock also hit a new high of $1,200 at the close.
Fueled by surging demand for its AI chips, particularly the Hopper H100, Nvidia has seen an unprecedented growth in the last two years, especially after the release of ChatGPT in November 2022.
Last month, Nvidia reported a 628% Year-on-year (YoY) increase in profit, while its revenue tripled to $26.04 billion from $7.19 billion a year ago.
Nvidia's market cap rose to $1 trillion in May 2023 and in February, it crossed $2 trillion, making Nvidia more valuable than even Amazon and Alphabet. Nvidia's AI chips account for nearly 70% to 95% of the AI chip market, according to Mizuho Securities, a Japanese financial services firm.
The chipmaker has played an instrumental role in the success of ChatGPT. Its DGX AI supercomputer was used to train the large language model (LLM) GPT-3 on which ChatGPT was initially based. Now AI chips like the Hopper H100 power several supercomputers that train large language models including ChatGPT and Google's Gemini.
Nvidia’s foray into AI is not limited to chips. The firm has also developed several generative AI platforms for enterprises. At the GTC developer conference in March 2023, Nvidia released a cloud-based service called AI Foundations, which can be used by enterprises to build, refine and operate custom LLMs. The chipmaker also announced new generative AI models such as NeMo (text-to-image generation model), and Picasso (a multi-modal LLM that can generate images, video, and 3D applications).
The chipmaker is hoping to cement its dominant position in the AI chip market with the new Blackwell chips. Unveiled in March at GTC 2024, Blackwell chips boast 208 billion transistors and can power LLMs with trillion parameters at 25x less cost and energy consumption, according to Nvidia.
“For three decades we’ve pursued accelerated computing, with the goal of enabling transformative breakthroughs like deep learning and AI,” said Jensen Huang, founder and CEO, Nvidia at the time of the launch of Blackwell chips.
Huang also called generative AI as the “defining technology” of the current times. He has a point, as several companies including Nvidia owe their recent success to it. For instance, Microsoft's market cap also surged by $1 trillion in a year, allowing it to surpass Apple as the world's most valuable company in January. All this happened after it became a major investor in ChatGPT creator OpenAI and integrated generative AI across most of its products, from Azure Cloud to Bing Search.
That said, rival chipmakers such as Intel are also rushing new AI chips to the market to tap into the massive surge in demand for them. In April, Intel unveiled the Gaudi 3 lineup of AI chips, which it claims is faster and more power efficient than Hopper H100. Intel has tested the Gaudi 3 chips on Meta’s open source LLM Llama.
To slash the dependence on Nvidia and the rising cost of training LLMs, companies such as Open AI are also reportedly considering in-house chip development. They are also looking to diversify their chip suppliers beyond Nvidia.
In March, three authors sued Nvidia in a US court for using copyrighted books without permission to train NeMo. Authors Brian Keene, Abdi Nazemian and Stewart O'Nan claimed that their books were part of a dataset of about 196,640 books used to train NeMo to simulate written language.
Nvidia is also one of the leading suppliers of graphics processing units (GPUs) for personal computers. However, its market share in PC GPUs is only 18% as compared to Intel’s 67% as of Q4 2023, according to a Statista report.
Image credit: Nvidia