Nvidia about to announce “supercomputing plan” for Brazil

While discussing the global AI chip race at the center stage of Web Summit Rio this week, Marcio Aguiar, director of enterprise sales at Nvidia Brazil, said that the company is not interested in fighting for market share, but in opening new markets.

“That is why we also work side by side with Google, AWS, Microsoft, etc.” he said, referring to companies that have been creating their own microprocessors for generative AI, while using Nvidia graphics processing units (GPUs) in their cloud computing services. “All our platforms are compatible with them. It is what we call ‘coopetition.’ The biggest winners in this process are the customers.”

The company is preparing to launch a new generation of chips in the second half of this year. It seeks to meet the growing expectations of AI developers and researchers, a segment in which it has an 80 percent market share.

Many of Nvidia’s investments are currently focused on the biopharmaceutical industries, robotics, and what the company calls the “omniverse” — a platform through which it is possible to create digital twins, that is, virtual representations of infrastructures that operate as in the real world, using AI. “We’re having a lot of investments in developing platforms for drug discovery,” he said.

But it is in the race that came before AI that the company is about to make a new announcement in Brazil. Earlier this week, Nvidia announced a new supercomputer called “Venado,” installed at Los Alamos National Laboratory in the U.S. in partnership with Hewlett Packard Enterprise (HPE). The machine can accelerate atomistic simulations for material research and high-resolution astrophysical simulations.

According to Mr. Aguiar, Brazil is on the path of high-capacity computing investments. “Very soon, in the next few months, we will announce a major supercomputer plan for Brazil,” said the executive.