Geopolitics of artificial intelligence: Nvidia's dominance, and Italy?

In recent times, there has been exponential growth in the field of artificial intelligence, with particular emphasis placed on the geopolitical and technological dynamics that influence its development.
In the current technological context, the success of Nvidia, a company that mainly deals with hardware components, emerges in a surprising way.
Over the course of 2023, Nvidia saw its market value triple, a remarkable achievement that places the company in a leading position in the industry.
Its computing systems are critical to the operation of ChatGPT and the evolution of large language models (LLMs).
This means that companies like Meta, in their attempt to remain competitive, are forced to invest billions in purchasing additional hardware from the same vendor.
In fact, in the last quarter alone, Nvidia's data center division generated more than $14 billion in sales, a significant increase from the previous year.
Jensen Huang, CEO and co-founder of Nvidia, has received considerable media attention for this success.
His presence was especially relevant a few hours before Elon Musk's speech at the New York Times DealBook summit, where he discussed the company's prospects and his expectations regarding self-rationalizing artificial intelligence.
Nvidia was compared, in an article in The New Yorker, to the companies that supplied equipment during the gold rush, thus underlining the crucial role that the company plays in the current AI "race".
However, despite this positive outlook, Nvidia faces significant challenges, particularly around AI development in China.
Recently, the company revealed that it expects problems in sales of its data centers in China, following new export restrictions imposed by the US Department of Commerce.
These restrictions particularly affect Nvidia GPUs, which are critical to the operation of advanced AI systems and difficult to replicate.
Such measures could dramatically reduce Nvidia's business in China, which accounts for about a quarter of its total data center sales, equivalent to about $3.6 billion.
In response to these restrictions, Huang expressed cautious optimism, underlining the need to develop new chips that comply with current regulations.
This situation has created a bitter confrontation between Nvidia and the US government, with the company trying to adapt its products to new regulations and the government imposing additional restrictions.
This dynamism reflects US concern about the overall development of AI in China, rather than specific technical issues related to GPUs.
US Commerce Secretary Gina Raimondo has taken a firm stance on this front, declaring that any attempts by Nvidia to redesign chips to circumvent the restrictions would be promptly monitored and regulated.
This presents Nvidia with a complex challenge: finding a balance between creating chips powerful enough to be useful, but at the same time limited enough to meet export regulations.
As Nvidia navigates these troubled waters, China is looking to bolster Huawei as an alternative in the chip and AI space.
However, despite its efforts, Huawei seems far from being able to compete with Nvidia in both scale and quality.
The separation of markets between the United States and China makes the term “competition” almost inappropriate, as Chinese LLMs lag significantly behind those in the United States.
Despite this, Huawei may be able to gain a prominent position in the Chinese market, but this would result in a further delay in the overall development of AI in China.
This power dynamic between the United States and China in the field of AI is seen by many as a strategic geopolitical competition, where every move has global repercussions.
For the United States, the imposed restrictions can be considered a victory, but they leave Nvidia in a vulnerable position.
Intense demand for Nvidia products is fueled by the current enthusiasm for AI, but without continued improvements and innovations, this demand may decline.
The history of AI research is dotted with periods of stagnation, known as “AI winters,” during which interest in and funding for AI declines dramatically.
A new “winter” could be upon us if the initial promises of companies like OpenAI, producer of ChatGPT, do not materialize.
And this scenario represents the biggest threat to Nvidia's future.
And Europe? The European Chips Act, officially known as the European Chips Act, is a legislative proposal announced by the European Commission with the aim of strengthening the European Union's capacity in semiconductor manufacturing and technology.
This initiative was started as a response to global chip shortages and growing demand for semiconductors, which have become critical components in numerous industries, from automotive to consumer electronics and telecommunications.
Unfortunately, investments in this direction are very low for now.
The program should mobilize 43 billion euros of public and private investments, with the aim of doubling the global market share in semiconductors held by the EU from the current 10% to at least 20% by 2030.
Figures much lower than the Silicon Valley investments.
What is certain is that the sector is also strategic for Italy and Europe; an increase in investment, in the order of tens of billions of euros, will be fundamental to bridge the gap with the United States, China and Taiwan which currently holds the record as the largest microchip producer in the world.
read also From electric cars to garlic: where the US-China economic clash comes from