This week a single company has concentrated all the focus, and all the anxiety, of the market. It has been written in the economic press that nothing like it has ever been seen on Wall Street. Keep this name: Nvidia.
After the publication of its accounts on Thursday, its profits – of 12,285 million dollars – skyrocketed by 769% compared to the 225% of its market revaluation. Its stock market value alone has exceeded 2 trillion dollars, which is equivalent to almost twice the entire Spanish economy. It is already the sixth most valuable company in the world.
This company has become the example of the bonanza of a technology company with a quasi-monopoly on microchips to train AI models. It has practically eaten Intel, it has an almost irreplaceable graphics card for AI, and it already captures more than 70% of its market. Nvidia does not generate Artificial Intelligence, but it builds all the routes so that third parties can achieve it so the machines can do what they do. All this without prejudice to political, regulatory and social concerns about the strong impact of this new technology on our world as we know it.
It is a great paradox that at the epicenter of the AI ??boom is a hardware company, and not a software company: the physical prevails. In a single year, the company has doubled sales. Its founder and CEO, Taiwanese American Jensen Huang, 61, estimates they will grow 20% more this quarter. The AI ??opportunity is so great, he assures, that in three years the investment in chips across the planet can reach a trillion dollars.
I have read one of the best descriptions of Nvidia from an economic correspondent of a competing newspaper. This analyst compared the AI ??fever to the Sierra Nevada gold rush of almost 180 years ago. So those who made money were not the nugget hunters, but the blade manufacturers. According to this, Nvidia is selling the ‘shovels’ to all companies interested in doing business with AI, whether to improve productivity or for generative applications. Chips and more chips for language models (ChatGPT, Claude), recommendation engines (TikTok, Google), advertising optimization (Facebook, Instagram), SaaS companies (Adobe), biomedical research, AI services,… The potential business is colossal.
The market shakes. There is fear of another bubble like the one that burst 25 years ago with American dot.coms. But judging by Nvidia’s numbers, there will be no such bubble: the use cases are real, the product is solid, and Nvidia’s profits justify the meteoric rise in its share price. Half of the chips are sold to cloud companies and other hosting companies specialized in GPUs, the data processing units. The other half goes to big companies… and governments! India, France, Canada or rich Arab countries are already creating publicly owned language models.
If this is not a revolution, what is?
