Nvidia (NVDA), which currently dominates the market for A.I. chips, yesterday (June 6) saw its market cap exceed $3 trillion for the first time, surpassing Apple (AAPL) to become the world’s second most valuable public company behind Microsoft (MSFT). The milestone comes ahead of the company’s upcoming 10-for-1 stock split tomorrow (June 7), which will artificially discount the price of individual Nvidia shares by 90 percent in a bid to make stock ownership more accessible. After the split, one $1,200 Nvidia share will become ten $120 shares. A stock split doesn’t change a company’s market cap or any of its business fundamentals, but the move often boosts a company’s stock price in the short term.
The split will cap a meteoric rise for Nvidia, helmed by Jensen Huang, which in the first quarter reported profit and revenue that rose by 628 percent and 268 percent respectively from a year prior. Nvidia’s booming success is largely driven by its H100 graphics processing units (GPUs) that underpin generative A.I. models and have become highly sought-after amongst the world’s leading tech companies.
Who is buying up Nvidia’s A.I. chips?
Microsoft and Meta (META) are two of the largest buyers of Nvidia’s H100 chips. The tech companies spent a combined $9 billion on the accelerators in 2023 alone, according to analysts at DA Davidson, while Omdia Research estimates the two companies acquired 150,000 chips each.
Microsoft reportedly plans to amass some 1.8 million GPUs by the end of 2024, much of which will likely come from Nvidia. Meanwhile, Mark Zuckerberg in January announced Meta’s plans to purchase some 350,000 H100 GPUs from Nvidia during the same time frame. During Nvidia’s first-quarter earnings call, the chipmaker revealed Meta’s most recent large language model, Llama 3 was trained on a cluster of 24,000 H100 GPUs.
Other major clients include Google (GOOGL) and Amazon (AMZN), which each purchased around 50,000 chips from Nvidia last year. Alongside Microsoft and Meta, the four Big Tech companies account for nearly 40 percent of Nvidia’s revenue, as reported by Bloomberg. Amazon in a recent earnings report announced plans to make its cloud service AWS “the best place to run Nvidia GPUs, helping customers unlock new generative A.I. capabilities.”
The other 60 percent of Nvidia’s business goes to cloud service providers including Oracle, CoreWeave and Lambda, as well as Chinese tech giants Tencent, Baidu, ALIBABA (BABA) and ByteDance (the parent of TikTok). Tesla (TSLA) is also a large customer, estimated to have bought 15,000 A.I. chips from Nvidia in 2023. The automaker’s CEO Elon Musk announced during the company’s first-quarter call that 35,000 H100s are currently active, with plans to increase that number to 85,000 by the end of the year. Musk earlier this week also revealed that logistical issues forced 12,000 chips originally intended for Tesla to be redirected to his social platform X, formerly known as Twitter, instead.
In March, Huang launched Nvidia’s next-generation A.I. chip Blackwell, which will be available later this year. Microsoft, Meta, Google, Amazon, Oracle, Tesla, xAI and OpenAI were listed as time-to-market customers for the new chip during the company’s quarterly earning call. Huang also recently announced plans to follow the Blackwell with the Blackwell Ultra in 2025 and a new A.I. chip platform called Rubin in 2026, with plans to upgrade its A.I. accelerators annually.