Nvidia is the most valuable company in the world. With a capitalization of around 3.4 billion euros (twelve zeros), the company has shattered all expectations by riding the wave of artificial intelligence (AI) like no other. The market expects this technology to cause a revolution in multiple sectors, but at the moment the real big revolution is the one led by Nvidia, which has gone from fighting to enter the 20 most valuable companies to to look in the rearview mirror at everyone else in less than two years.
This is a boom which, beyond stock market fluctuations, manifests itself in each income statement: almost 50 billion euros in net profit during the third fiscal quarter, presented this Thursday. That’s 190% more than the same period in 2023. Jensen Huang, its co-founder and CEO since 1993, explained on a call with investors what the two ways of looking at things are. Nvidia is considered the artificial intelligence chip vendor that everyone wants. The other is to be the lever of “the third industrial revolution”.
“The tremendous growth in our business is driven by two fundamental trends that are driving the global adoption of Nvidia computing,” began the businessman, known for appearing in public wearing a black leather jacket.
A different brain model
“First, computing is undergoing a reinvention, a platform shift, from coding to machine learning. From running code on CPUs to processing neural networks on GPUs,” describes Huang, who is 61 and has been running his company for over 30 years.
CPUs and GPUs are the two types of brains that a computer can have. The design of central processing units (CPUs) allows them to perform many different and complex tasks, such as running all types of programs, managing an operating system, and coordinating all business processes. ‘a computer. The problem is that their performance drops when they have to multitask.
Graphics processing units (GPUs), on the other hand, are designed to perform many simple tasks at the same time. They are not as versatile as processors, but they can handle immense volumes of data at the same time. Nvidia is a specialist in this technology, both at the hardware and software levels, and each of its new chips further increases the volume of data that a system that equips them can manage.
For decades, the primary commercial application for GPUs and the business through which Nvidia made money was video games. GPUs are ideal for processing your graphics and they get their name from them. With the explosion of artificial intelligence, GPUs have proven more useful than CPUs for the large volume of parallel calculations required to train and execute them. Last quarter, the video games division left around $3.2 billion in Nvidia’s coffers. That of data centers and chips, around 30.7 billion.
“The coding runs on the CPU, but the machine learning that creates the neural networks runs on the GPU. This fundamental shift, from coding to machine learning, is currently widespread. There is no company that will not do machine learning. Machine learning is also what enables generative AI, which is why computer systems worth trillions of dollars are currently being retrofitted for machine learning,” Huang summarizes.
In this global re-digitization, your business has no rival in the market. Google uses Nvidia products. Its maximum competition, Microsoft too. If Elon Musk one day manages to present a fully autonomous Tesla, it will be thanks to Nvidia chips. Without them, ChatGPT would be dozens of times less powerful. Situations like this are repeating themselves in thousands of small companies moving into the AI sector, as well as state supercomputers.
AI factories
The second trend Huang highlights is that “generative AI is not just a new software capability, but a new industry” that will change the design of today’s data centers.
“Data centers will truly become AI factories. We will produce AI just like we produce electricity today. And if the number of customers is large, as is the number of electricity consumers, these generators will operate 24 hours a day, seven days a week,” Huang describes: “We will see this new type of system. , I call it an AI factory, because that’s the closest thing to it. “It’s different from data centers of the past.”
I call it an AI factory because that’s the closest thing to it. It’s different from a data center of the past
Jensen Huang
— NVIDIA CEO
This is the “creation of a new industry”, a “third industrial revolution” as the CEO of Nvidia describes it. Not only for today’s generative AI, but for the next step, “agent” AI, a technology with more capacity than the current one that allows machines to move between humans. This new era of humanoid robotics is on the company’s roadmap and “it’s not science fiction”, as several robotics experts explain in this report from elDiario.es.
“AI is transforming every industry, business and country. Businesses are adopting agentic AI to revolutionize workflows. Over time, AI colleagues will help employees do their jobs faster and better. Investments in industrial robotics are skyrocketing as advances in physical AI drive demand for new infrastructure,” Huang continued.
AI colleagues will help employees do their jobs faster and better
Jensen Huang
— NVIDIA CEO
“Countries around the world recognize the fundamental trends we are seeing in AI and have realized the importance of developing their national AI and infrastructure. The age of AI is here and it is vast and diverse. Nvidia’s experience, scale and ability to deliver a full stack and comprehensive infrastructure enable us to address all of the multi-billion dollar AI and robotics opportunities that lie ahead,” the entrepreneur concluded during his call. to investors.
The hidden face of the revolution
Although he directly refers to electricity, Huang has not entered into one of the major debates that accompany the revolution he defends: energy. The thousands of GPUs like those manufactured by Nvidia, which, as its CEO indicates, can operate 24/7, consume a lot of electricity by working without rest and having very intensive calculation procedures.
New generations of powerful Nvidia GPUs entering data centers have caused Google and Microsoft to skyrocket their polluting emissions and move further and further away from their goal of being carbon neutral by 2030. Amazon employees, for example, for their part, denounced the fact that the multinational cheats on its emissions figures and that these are increasingly higher.
“Its energy-intensive data centers operate in the heart of coal-dependent regions, and the company’s expansion is increasing demand for oil and gas,” they noted after the publication of the report. environmental impact of Amazon, which ensures that its polluting emissions decrease.
It’s a debate that’s also taking place in Spain, one of the countries attracting new data center deployments. Amazon and Microsoft have invested billions in Aragon to build this type of infrastructure, Meta will do it in Talavera de la Reina and others like IBM or Google have also chosen this country to be the center of their IT processes in the south of Europe.
Although one of Spain’s competitive advantages in this sense lies precisely in the possibility of powering data centers with renewable energy, the concerns of some activist groups extend to the use of water for cool these large infrastructures. The employers’ association assures that the figures have been exaggerated and that new generation data centers do not involve water costs, but multinationals like Amazon and Microsoft have refused to reveal their consumption figures when asked by elDiario.es.
