NVIDIA and AI: Opportunities and Threats
The advent of AI has been an influential defining feature for the trajectory of many companies in 2023. Much of the momentum that AI acquired in 2023 has transferred over to 2024, and almost exactly one month ago, we witnessed a symptom of this rise of AI: Microsoft topping Apple in total valuation.
Companies investing in AI technologies are currently driving up the stock markets. The S&P500 rose to record heights in the past few months, and this is almost fully attributed to the growth of only seven stocks, which outperformed the ETF. One of these seven companies is NVIDIA, which lately became the 6th most valuable company in the world with a market cap of $1.781 trillion. This analysis will cover opportunities and risks associated with an investment in the corporation.
NVIDIA is an American technology company that designs graphics processing units (GPUs). These GPUs are specialized and powerful computational units that can perform highly complex graphical computations. NVIDIA is known for creating gaming hardware, and while many of the applications for their chips have lied in gaming, they also have professional purposes in fields such as architecture, engineering and construction, media and entertainment, automotive, scientific research, and manufacturing design. In 2020, with the rise of crypto technology, these GPUs have also proven to be suitable tools to mine cryptocurrencies. However, NVIDIA hasn't achieved its latest growth through any of the aforementioned. Instead, it was AI that enabled NVIDIA to attain these levels of success.
GPU Chip (NVIDIA)
To understand NVIDIA's association with AI and neural networking, we need to look back to 2012. That year, AlexNet, a convolutional neural network (CNN) architecture, was created by Alex Krizhevsky in collaboration with Ilya Sutskever and Geoffrey Hinton. The algorithm was designed to recognize the content of an image, and even though it was not the first of its kind, it played a crucial role in the development and research of AI and deep learning. There are two reasons for that.
Visualisation of a Neural Network
Firstly, the algorithm significantly improved the accuracy of its predictions when compared to its predecessors. What set AlexNet apart was its sheer size and depth, consisting of over 500.000 neurons. The dataset consisted of over 15 million pictures from 22.000 categories; to process each one of these pictures, 700 million math operations were needed.
Secondly, AlexNet pioneered the use of GPUs to train neural networks. It turns out that rendering the virtual scenery of a videogame would require a lot of graphical processing power. Gaming has always been one of the most demanding computational applications for personal computers, and GPUs have always been the core of gaming computers. In the 2010s, Jesen Huang, the founder and CEO of NVIDIA, would uncover that these GPUs, initially designed for gaming, were extremely efficient in training deep learning algorithms.
The paper published about AlexNet can be found through this link.
For almost ten years, NVIDIA has been one of the driving forces in research, development, and investment into AI while also providing the most advanced gaming ware on the planet. Unlike competitors like Intel or AMD, NVIDIA has invested significant amounts into developing hardware for AI systems. This strategic decision has set the company up for future dominance in an emerging industry rather than solely positioning it in a market with existing demand. In 2023, NVIDIA released the H100 GPU, the most powerful chip on the market, specifically made to accelerate AI development and training. These chips have become very sought after, and most noticeably, they have contributed to training large language models used by companies such as OpenAI.
In addition, NVIDIA has invested in more than two dozen AI companies in 2023 alone, ranging from established incumbents in the AI industry to niche start-ups. To a certain degree, the investees aim to enter a strategic partnership with the chip manufacturer. Prioritized access to the newest chips can help these companies maintain a competitive edge. NVIDIA has denied that there was a formal agreement for the investees to buy NVIDIA chips in the future. Still, the company has ensured consistent demand for its AI GPUs for the foreseeable future. The investments, therefore, will not only allow for stable returns but will allow NVIDIA to become one of the greatest players in the development of AI.
It seems undeniable that NVIDIA now plays a central role in the emerging sector. And even though there are many reasons to be optimistic about the company's prospects, the potential risks also need consideration.
To do that, we need to look back to 2020-2022. The pandemic led to the collapse of supply chains and exposed the dangers of undiversified sourcing. NVIDIA mainly procures its semiconductors in Taiwan, a region threatening to become the stage for yet another geopolitical conflict. As dependence on NVIDIA’s products becomes more intense, a collapse of the supply chains today will have much more wide-ranging effects on the economy and a broader array of industries. At the same time, China accounts for 20%-25% of NVIDIA’s sales as the country aims to advance AI research. Considering the dangers of an escalation of the China-Taiwan conflict and the possibility of intensification of the US-Chinese trade war (due to a potential second US-presidential term under Donald Trump), strategic decisions are needed to mitigate the risks. Outcomes are hard to predict, and the future will show whether NVIDIA will succeed.
TSMC building in Taiwan (Bloomberg)
NVIDIA’s earnings report is scheduled for February 21st, and the stock has an earnings ESP (Expected Surprise Prediction) of 3.68%. Investors are looking forward to that date, as NVIDIA tended to surpass expectations in the past.
*This essay is informative - not financial advice or recommendation. I do not provide personal investment advice and I am not a qualified licensed investment advisor!
Ferdinando Angeloni