By Emil Bjerg, journalist and editor
Nvidia’s growth has skyrocketed in the current AI boom, but the company also finds itself in the middle of geopolitical tech-tensions between the US and China. And now Big Tech wants to make their own chips.
“Who’s getting how many H100s and when is top gossip of the valley rn,” OpenAI’s Andrej Karpathy tweeted in August. H100 is an Nvidia chip powering the AI boom.
Nvidia, originally a niche graphics card maker for video games, has become a giant that Big Tech relies heavily on to build AI software.
The demand and dependence translate into their skyrocketing market value: Last week, Nvidia became the world’s fourth most valuable company, only surpassed by the likes of Microsoft and Apple. As of this writing, Nvidia has a market cap of $1.98 trillion.
To anyone wondering how Nvidia could reach that valuation: a single of their in-demand H100 chips costs between $15,000 and $40,000, depending on packaging and other factors.
However, several factors, from in-house production of AI chips to geopolitical competition, threaten Nvidia’s impressive market cap. We will return to their challenges after a dive into Nvidia’s ascent.
The semiconductors supporting the crypto- and AI booms
Those who had the foresight to invest in Nvidia early have seen values soaring in the last five years, from $36.90 in March 2019 to $792.38 as of this writing.
This graph from Macrotrends illustrates the meteoric rise of Nvidia over the past years.
When Nvidia passed Intel as the biggest chip maker in 2020, it was their bet on powering cryptocurrencies that paid off: Nvidia’s GPUs, initially designed for gaming, proved exceptionally capable of performing the complex calculations required for cryptocurrency mining, driving up demand among miners worldwide. (The subsequent fall in Nvidia’s stock occurred as the cryptocurrency market experienced volatility and decline.)
ChatGPT’s public release in November 2022 sparked the latest meteoric rise, and Nvidia has grown exponentially – and dramatically so – ever since. In the AI gold rush, Nvidia has been the shovel, the enabler that has allowed software companies to participate in the AI race. And that has paid off: While AI companies generally struggle to monetize their AI-related products, Nvidia’s AI enablers are one of the few AI-related products that have been profitable at scale.
Today, Nvidia accounts for 70 percent of AI chip sales – their share is even higher when it comes to generative AI.
How did Nvidia beat Intel?
Nvidia has been better than Intel in crypto and AI for several reasons. Firstly, utilizing AI requires handling massive data sets. Nvidia’s chips feature a highly specialized architecture optimized for efficient parallel processing, a critical requirement for handling the massive datasets associated with AI workloads and crypto mining. A quicker recognition and adaption to the market shifts towards AI and machine learning has helped them position themselves as a leader in this space.
Nvidia has invested heavily in AI, developing hardware for AI and software and platforms like CUDA for AI development. That has created a community of programmers who consistently invent using the company’s technology. Today, Nvidia is a company that produces computers, software, cloud services, trained A.I. models, and processors.
Naveen Rao, a founder of an AI chip start-up bought by Intel, says that everybody “builds on Nvidia first[…] If you come out with a new piece of hardware, you’re racing to catch up.”
Jensen Huang: A new power player in tech
Nvidia’s ascent has made CEO and co-founder Jen-Hsun Huang, also known as Jensen Huang, a power player in tech. Often seen in a trademark leather jacket and black t-shirt, Huang has become someone tech leaders want to rub shoulders with to secure more H100s before their competitors.
On Nvidia’s position in the AI boom, Huang recently said: “The thing that we understood is that this is a reinvention of how computing is done. And we built everything from the ground up, from the processor all the way up to the end.”
This understanding has brought Nvidia close to the biggest companies in tech.
Big Tech’s biggest supplier
Nvidia is the biggest supplier for Big Tech, currently competing to dominate generative AI.
In 2023, Meta and Microsoft spent $4.5 billion on Nvidia’s H100 chips, with Amazon and Alphabet (Google’s parent company) following, spending 1.5 billion dollars each.
Big Tech’s reliance on Nvidia is a blessing and a curse for the chip maker, who’s grown equally dependent on those few companies. And now, several Big Tech companies are looking to end their Nvidia dependence.
In-house chip production
As The New York Times writes, the “boom in generative A.I. over the last year exposed just how dependent big tech companies had become on Nvidia. They cannot build chatbots and other A.I. systems without a special kind of chip that Nvidia has mastered over the past several years”.
The surge in demands has seen Nvidia struggle to keep up. And now Nvidia’s largest customers – Amazon, Microsoft, Google, and Meta – are all designing their own chips. To reduce costs and dependency on a company that also supplies their competitors.
Sam Altman, CEO of OpenAI, is also pursuing chip independence. He is currently seeking a whopping five to seven trillion in an attempt to reestablish competition and a steady supply chain of chips: “the world needs more ai infrastructure[…] than people are currently planning to build[…] building massive-scale AI infrastructure, and a resilient supply chain, is crucial to economic competitiveness,” Altman recently tweeted.
In 2023, Google, the leading chip developer among big tech companies, spent two to three billion dollars producing their own chips. According to the New York Times, Nvidia sells its chips from $15,000 and up. The search giant can build relatively similar chips for 2-3000 USD each.
However, after ten years of independent chip development, Google still relies on Nvidia’s chips for certain products.
Independent chip development could pose a future challenge for Nvidia. A more immediate threat is the tensions between the US and China.
Geopolitical showstoppers
Few resources are as vital to the supply chains in the modern world as semiconductors and chips. As a prerequisite for most technological products – from gaming consoles to modern cars to AI software – they also create substantial tensions.
Taiwan is home to Taiwan Semiconductor Manufacturing Company (TSMC), the world’s largest contract chipmaker, which produces the majority of Nvidia’s chips. The semiconductor factory has become a focal point of global technology and geopolitical rivalry between the United States and China. As such, Nvidia, a company that many companies are deeply dependent on, is deeply dependent on TSMC and relative political stability in Taiwan.
And that’s not the only way the rivalry between China and the US affects Nvidia. Also the US curbs on AI chip sales to China have significantly impacted the chip giant. Aiming to undercut China’s AI development, the US has tightened regulations, blocking access to advanced chips from companies like Nvidia.
Nvidia has faced a direct hit from these restrictions, with expectations of a significant decline in sales to China, which previously constituted a substantial portion of its revenue. To limit the losses, the company has been developing new processors tailored to Chinese customers to comply with US regulations.
Nvidia has stated that while it has managed to comply with regulations, the long-term impact could be detrimental, as Chinese companies may seek alternatives, potentially eroding Nvidia’s market share and influence in one of the world’s largest markets.
What does the future hold for the chip giant?
Amid its massive success, Nvidia is challenged by Big Tech chip independence and geopolitical strides. On the other hand, their chips are among the fastest on the market, and the company has a headstart with its community of developers and offerings combining hardware and software.
Nasdaq writes that Wall Street analysts expect Nvidia’s rally to continue and that it has a consensus rating of “Strong Buy” from analysts.
Nasdaq even goes as far as speculating that Nvidia could become the world’s biggest company: “While you might have been ridiculed just a few years back for talking about the possibility of Nvidia becoming the world’s biggest company, it looks like a real possibility now.“
Only time will tell if Big Tech and geopolitical tensions will slow down Nvidia’s ascent or if a hardware company fuelling the AI boom can, in fact, become the largest company in the world.
One thing is certain. Nvidia is no longer an underdog niche graphics card maker for video games.
