Photo: Nvidia CEO Jensen Huang in Las Vegas. Credit: Flickr/LG Electronics (CC BY 2.0)
Nvidia’s brief streak as the world’s highest valued company in the last several weeks has catapulted the chipset manufacturer into the headlines all over the world. A year ago, Nvidia was valued at just under $1 trillion. Nine months later, it blew past the $2 trillion mark, and just 30 days after that, for a moment in June 2024 Nvidia peaked at $3.3 trillion. Apple and Microsoft are household names in tech, but what is Nvidia? And what does its massive surge in valuation tell us about the direction of the tech industry and the economy?
For most of its 30-plus years, Nvidia has situated itself as the prime supplier of consumer graphics processing units, or GPUs. For decades, Nvidia has been among the top manufacturers of GPUs worldwide, but recent years have seen the company become one of the biggest beneficiaries of financial bubbles around emergent technologies — first with cryptocurrencies and now with so-called “Artificial Intelligence” applications.
From video games to AI powerhouse
Until the early 2010s, Nvidia carved out a niche within the computer hardware industry producing consumer GPUs on a large scale. Initially designed for video gaming purposes, GPUs are effective at performing many simple tasks simultaneously and have become a fundamental component of computers worldwide. First quarter reports for 2024 show the company holding a staggering 88% market share for discrete GPUs.
The first major wave of growth for Nvidia began with the cryptocurrency bubble in the late 2010s. In the rush to generate as much cryptocurrency as possible, so-called “miners” began stringing dozens of consumer GPUs together in data centers to mine cryptocurrencies like BitCoin. This caused a surge in prices — the average selling price of GPUs increased from around $275 in late 2019 to nearly $1,100 in mid 2021 — and an opportunity for Nvidia to greatly expand its market share.
Still, the cryptocurrency bubble did not see Nvidia stock surge into bubble territory itself — when crypto burst and its value plummeted, Nvidia found itself the center of the next big thing in tech, the bundle of emergent technologies currently branded as “Artificial Intelligence.” This moment opened up the space for Nvidia to expand into GPUs purpose built for large scale AI data centers, especially important for generative AI applications like ChatGPT.
Some of Nvidia’s largest customers now include OpenAI, the firm responsible for ChatGPT, Meta, Microsoft, Tesla and Amazon — all of which have stated that they are stockpiling Nvidia’s flagship H100 generative AI optimized chip. In this way, Nvidia has positioned itself, as many commentators have noted, as the main mover of “picks and shovels” for this “AI gold rush.”
Consequences of the AI ‘gold rush’
There are many differing interpretations and debates of what constitutes artificial intelligence and even if simulating consciousness — the ostensible goal of many of the big players in AI — is even possible, much less desirable. While philosophically interesting and the basis of much science fiction, these debates take center stage and more often than not serve to obscure what is materially happening in the tech industry around what is currently being sold as AI.
The direction of technological development under capitalism is directed by capital for the purposes of capital. Sometimes that is to the ends of automating or surveilling or otherwise repressing labor. Sometimes it is simply to maximize return on investments in the speculative arena of the stock market. Rarely is technology developed for the purpose of improving living and working conditions for the vast majority of humanity despite decades of utopian promises from Silicon Valley.
The recent hype around AI applications exemplify many of these trends. On one hand, Nvidia maintains an undeniable monopoly on the hardware powering the AI boom. On the other, its explosive stock prices and brief stint as the world’s most valuable company is fueled largely by hype, marketing and the very familiar dynamics of capitalist markets and emergent technology going back to the dot com bubble of the late 90s and early 2000s.
Nvidia is one node in a network of monopolies, startups and venture capital firms that are pushing the AI craze to its extreme. Now, nearly every digital service has a half-baked “AI” component, whether that is a chatbot and image generator embedded into Meta’s messaging platforms, corporate enterprise adoption of ChatGPT powered programs, or a comically problematic “AI summary” of your Google search request. These use cases are not only largely seen by the public as unnecessary and unwelcome intrusions, but also present a massive environmental concern.
Generative AI processes require more energy to power, and water to cool components, than even crypto mining ever did as GenAI data centers are being deployed at an industrial scale. The International Energy Agency recently estimated that energy demand from data centers, currently about 2% of global energy consumption, could double by 2026, representing additional demand equivalent to “at least one Sweden and at most one Germany.” For example, an AI-powered Google search result requires ten times as much energy than a traditional keyword search.
The future of Nvidia and AI
The U.S. government has identified the AI industry as a key battleground in the new Cold War it is waging against China. Maintaining monopoly power and diversifying production sites of high end microchips has been repeatedly cited by subsequent U.S. administrations as a key national security issue. For example, both Nvidia and legacy giant in the semiconductor industry Intel rely on the Taiwan-based TSMC for manufacturing, long seen as an unsustainable choke point in the global economy.
Intel holds a 76.4% market share in CPUs — central processing units. CPUs are also an integral component in computers and data centers of all sorts, but Intel has struggled to compete with Nvidia in the GPU market for a variety of reasons, especially concerning AI applications. The company was recently targeted by the Palestinian BDS movement for its extensive operations in Occupied Palestine and also bailed out by the U.S. government as the company continues to see declining — although still massive — profits and executives focus on stock buy-backs.
In November 2023, the Biden administration expanded the export regulations restricting the export of high-end Nvidia AI chips to China, Russia, Iran and numerous other countries deemed official enemies of the U.S. government. Meanwhile, China has turned extensive investment, over $47.5 billion, through state-owned banks and enterprises into domestic semiconductor development.
It is easy to get wrapped up in overblown hype — or fear mongering — about AI. Ultimately, technological development and how it is deployed in the world under capitalism is shaped by the realities of the global economic system, not the volatile trends of the stock market week to week.
Whether Nvidia crashes imminently or not, it does not change the systems that created the tech industry as it exists today. There are numerous use cases for all sorts of emergent technology, including AI, that could be leveraged to the benefit of all people, but only if we fight for true popular, democratic control of technological research, development and deployment.
