We are often confronted with a legitimate question: "Is artificial intelligence showing the characteristics of a stock market bubble? This month, we take a look at Nvidia 's remarkable stock market performance over the past year and analyse the fundamental and practical factors that we believe have underpinned the rise in its share price.

A SPECTACULAR YEAR FOR THE STOCK

Nvidia is an American developer and manufacturer of high-performance graphics microprocessors (often referred to as GPUs, short for Graphic Processor Units). On 31st December 2022, its shares were trading at $136 on the US stock exchange. On 26th February 2024, just over a year later, its share price was $788, an increase of around 480%.

Nvidia ​ IS ​ NOW ​ ONE ​ OF ​ THE ​ WORLD'S BIGGEST STOCK MARKET GIANTS

This rare, if not unique, stock market performance for a large company has taken Nvidia 's market capitalisation to $1,971 billion, from an already respectable level of $340 billion at the end of 2022. Nvidia is now the 4th largest listed company in the world, surpassed only by Microsoft ($3,042 bn), Apple ($2,803 bn) and the Saudi Aramco oil company ($2052 bn). Nvidia is not only ahead of Amazon ($1,817 bn) and Alphabet , Alphabet 's parent company ($1,749 bn), but also well ahead of Meta ($1,229 bn for the former Meta ) and Tesla ($637 bn).

As an aside, it is important to note that the composition of this club of "mega market caps" reflects both the importance of technology in our lives and the ultra-dominance of the United States in the technology ecosystem.

Nvidia 'S STOCK MARKET PERFORMANCE STEMS FROM THE UNIQUE COMPUTER PROCESSORS IT PRODUCES

A brief historical review is in order. Nvidia 's processors were initially sought after by video game 'addicts’, who were eager for speed because the visual effects required a lot or processing power from the computers' electronic components. Nvidia ’s special architecture (known as 'parallel') enables this computing power, making the group one of the most efficient in the semi-conductor industry. Subsequently, the 'miners' of bitcoins (and other digital currencies) incorporated Nvidia chips into their computers, for the same performance reasons.

THE POWER OF Nvidia 'S PROCESSORS HAS MADE THE GROUP A KEY PLAYER IN THE FIELD OF ARTIFICIAL INTELLIGENCE

In order to train and infer, AI’s large language models (of which ChatGPT is the most famous, and which allow non-computer scientists to interact with AI) require an increasing amount of data, because the more data there is to analyse, the better the quality of the results generated by the AI. However, the need for computing power increases almost exponentially with the amount of data to be processed. Nvidia has therefore attracted the interest of AI players, thanks to the power of its processors. This appeal has even grown with the performance of Nvidia 's GPUs: already at the top end of the market, they have improved spectacularly in recent years. According to research firm Bernstein Research, the computing power of Nvidia 's GPUs has increased 1,000-fold in 10 years: from 4 billion operations per second in 2012 for the K20X chip to 4,000 billion in 2022 for the H100 chip.

DEMAND FOR Nvidia PRODUCTS IS BOOMING

The unique computing power of its chips and their special technical architecture have therefore placed Nvidia in a position of virtual monopoly for the GPUs that are crucial and indispensable to artificial intelligence. For AI, Nvidia has become the equivalent of the shovel and pick-axe dealers during the gold rush: Selling tools that were indispensable to all prospectors, they ended up making more money than the rare diggers of nuggets of yellow metal. To sum up (simplistically): today, without Nvidia 's high-speed GPUs, there is no AI.

These GPUs are used in very large quantities in the servers that make up the 'clouds’ that store the data that is essential for training and running AI models. These clouds can be both public (e.g. the huge hosting centres that Microsoft , Amazon and Alphabet in particular have set up to store their customers' data) and private (e.g. within companies to host their own data). And as the major AI players mentioned above, who are Nvidia customers, are competing fiercely to get their language models adopted, they are willing to invest massively by ordering Nvidia chips in very large quantities, to meet the sudden high demand that they themselves are receiving from their customers.