Bloomberg Línea — Nvidia (NVDA) recently exceeded Wall Street’s income expectations by a whopping 28%. The company reported an 88% increase in second-quarter revenue compared to the same period of 2022, posting a remarkable $13.51 billion. Notably, the data center segment, which accounts for sales of semiconductors used in Artificial Intelligence (AI) applications, has become four times larger than the gaming segment, with revenues reaching $10.32 billion, marking a remarkable 171% increase year-on-year.
Marcio Aguiar, Director of Enterprise Sales for Nvidia in Latin America, shared his insights about the company’s rapid growth in an interview with Bloomberg Línea. While technology certainly plays a pivotal role, Aguiar emphasized that Nvidia’s success goes beyond its technological prowess.
Nvidia’s growth can also be attributed to its strategic approach, which includes partnerships with both major corporations and startups that amplify demand, he said.
Additionally, a commitment to nurturing a robust ecosystem, comprising thousands of developers and potential entrepreneurs, has been a significant driving force behind its success, Aguiar continued, also noting that many of these developers are surprisingly young, with some still in their teenage years.
This unique blend of products and services is not limited to a specific country; it is a strategy that is actively applied in Brazil and Latin America. According to Aguiar, the Latin American region accounts for approximately 8% of Nvidia’s global revenue within the Enterprise division, with the majority of this revenue coming from Brazil. To put this in perspective, based on the most recent quarter’s figures, this would equate to roughly $800 million in revenue.
However, Aguiar emphasized that these numbers represent just a fraction of the region’s potential for Nvidia. One of the main challenges, however, will be Latin America’s bureaucracy around the acquisition of technological innovation equipment, such as the servers required to meet the computational demands of artificial intelligence.
These purchases are typically made through public tenders, which, in and of themselves, are not problematic. The issue lies in the prolonged decision-making process associated with equipment procurement. Aguiar provided the example that “major nations around the world start the year with an estimated budget dedicated to technological innovation, and we don’t see that in the countries in our region yet.”
In Brazil, the majority of Nvidia’s sales are facilitated through partnerships with cloud solution providers like Amazon (AMZN), Microsoft (MSFT), and Google (GOOG). Aguiar highlighted that there is currently a race among technology giants to adopt generative AI techniques, which are transforming industries across the board.
According to Aguiar, “Brazil has been a very important player in relation to this demand. This is work that we also do on the other side: we work with large corporations that are looking to develop their innovation concepts using artificial intelligence techniques.”
Aguiar acknowledged that there are headwinds to implementing large natural language models (LLMs) in Brazil due to a shortage of qualified labor and inadequate infrastructure. Access to appropriate hardware for researchers is also limited. LLMs possess the capability to transform programming languages into new information.
Demand that Is Hard to Measure
“We have many purchase orders lined up and we do not have full capacity, because today the big companies in this space are finally seeing the movement that we made a couple of years ago”, said Aguiar, referring to Nvidia’s “investment in ‘smaller’ data center companies in the past, teaching them how to set up their data centers with millions of servers”. Large companies are starting to purchase or partner up with data centers to meet end customer demand. Therefore, there a race for chips has started, according to Aguiar.
The executive stated that Nvidia’s sales are also influenced by competition among major nations such as the United States, China, Japan, and England.
Richard Camargo, an analyst at Empiricus Research, said that Nvidia has been the biggest beneficiary of the chip competition and even helped “stoke the fire” when it recently announced a $50 million investment in the biotechnology startup Recursion.
This company will use almost all of the capital to purchase Nvidia’s H100 GPUs and develop its artificial intelligence models, according to the analyst.
The H100 GPU line “is a huge commercial success,” and its demand is very difficult to measure, with “industry reports indicating that, given the shortage and uncertainty about how quickly applications will come to market, customers are building up inventories,” according to Camargo.
Ecosystem Fuels Demand
In Brazil, Nvidia has seen demand from large companies to use software techniques that limit certain information from being exposed in public clouds.
“Many companies today are beginning to be a bit concerned and are looking to create their own ChatGPT so that their data is not used as public data,” said the company’s leader in Latin America, referring to OpenAI’s solution.
“But this also requires an investment in terms of infrastructure and new talent,” Aguiar noted. “We have seen more and more of the top Brazilian companies, which have a lot of funding for technology, becoming stronger in adopting new computing concepts.”
“We have an ecosystem of nearly 200,000 developers who know how to use our technology in Brazil, and 4 million who have access to our technologies worldwide. Many of them are being absorbed by the industry. For us, this is excellent because it leads us to enter more and more companies.”
When asked about Nvidia’s expansion and growth plans in Latin America, Aguiar emphasized the training and engagement of the partner ecosystem and mentioned education and startup acceleration programs being implemented by the company.
Currently, there are about 390 startups in Brazil that work with artificial intelligence techniques and are part of Nvidia’s ecosystem. Worldwide, there are more than 14,000 startups, and the executive explains that Nvidia guides them on how to use the company’s software and hardware platforms.
Consequently, “these startups are now much closer to corporate customers and are being acquired. This brings Nvidia into those organizations.”
“This strategy, according to Aguiar, comes from the stance of Nvidia’s founder and CEO, Jensen Huang, who says the company needs to continue with a ‘startup mentality’ to be more agile and more assertive, ‘because we have a very strong commitment to the entire ecosystem we have created.’”
Hiring in Mexico and Brazil
Aguiar spoke with Bloomberg Línea via teleconference while in Mexico, where he attended a Dell (DELL) event.
The executive highlighted Nvidia’s significant presence in the country and partnerships with software companies and manufacturers to empower the local ecosystem.
“I had meetings with major software companies. Many software developers are serving a huge demand, especially from manufacturing companies, especially in the automotive industry, which have to adapt to new car designs (see more below).
He added that many of these software development companies interact with Nvidia to learn how to develop on top of the company’s platforms in the so-called “metaverse,” a mix of artificial intelligence with augmented reality.
“This allows these companies to simulate a factory that is still to be built and, with that, greatly reduce costs and increase productivity.” According to Aguiar, the company has major consulting firms as partners in Brazil. “This allows us to be very agile internally within our own operation and reach many new partners.”
Nvidia, which usually operates with a leaner operation, will start hiring more people in Mexico and Brazil in October.
Aguiar stressed that the focus is on training and recruiting new partners with knowledge to offer services. According to him, “for these IT techniques, you need a team that knows how to develop the platform, not just a partner to quickly sell hardware.”
“When we sell, it is a consequence of the computational demand of that particular application or the applications the customer uses,” he explained.
Voice AI and Autonomous Vehicles
Among the main technology trends at the moment, Aguiar believes that voice AI “will change all the concepts we have now.”
“Companies are adopting large natural language processing techniques, but then there is the question of translation, from text to video, etc. All of this requires much more computational power,” he said.
“We are facing new challenges so that data processing can consume less energy. Not all nations have the energy efficiency to truly provide this kind of infrastructure to their ecosystem,” Aguiar said.
Another technology trend for him is autonomous vehicles and the entire ecosystem where vehicles will operate.
Nvidia has a suite of hardware and software products for autonomous vehicle automation. One of its main customers is Mercedes-Benz. However, the automotive segment, with sales of sensors and semiconductors for vehicles, was the only “negative” highlight for Nvidia in the last quarter, with sales of $253 million, a 15% year-on-year decline.
The leading company in the AI semiconductor market recently announced a $25 billion stock buyback program, a measure generally taken when the company’s leadership believes its shares are undervalued.
This year, Nvidia’s shares have risen by about 240%, reaching a record level and raising the company’s market value to $1.2 trillion—only the sixth company in the “trillion-dollar club” currently, behind Apple, Microsoft, Saudi Aramco, Alphabet (Google’s holding company), and Amazon.
Matheus Popst, a partner at Arbor Capital, pointed out that the market’s expectations for the quarterly results were already considered high. “The guidance [for the current quarter] came in at $16 billion, while the consensus was $12.5 billion. A year ago, the consensus was $9 billion,” the manager told Bloomberg Línea.
“Margin and revenue expansion came much faster than they could hire and invest in the business. The conversation on the buy-side was that if Nvidia had [guidance] above $12.5 billion, it would probably be around $14 billion. Guidance of $16 billion was a surprise,” he said.
According to Empiricus analyst Richard Camargo, the performance of the data center segment brought operational leverage to the company, with a gross margin that increased by 26 percentage points year-on-year and 5.5 percentage points versus the first quarter.
Correction: The article was corrected at 8:40 AM to clarify that the 8% share of Nvidia in Latin America refers to the Enterprise division, not the company’s total revenue.