Nvidia bets on robotics to drive future growth

Open the editor’s digest for free

FT editor Rula Khalaf picks her favorite stories in this weekly newsletter.

Navidi is eyeing robotics as the next big growth driver as the world’s most valuable semiconductor company faces increasing competition in its core artificial intelligence chipmaking business.

The US tech giant, known for its infrastructure underpinning the AI ​​boom, is set to launch its latest generation of compact computers for humanoid robots – the Jetson Thor – in the first half of 2025.

Nvidia is positioning its technology group to be at the forefront of what it believes is the next robotics revolution. The company sells a “full stack” solution, from the software layers that go into the chips to train AI-powered robots.

“For physical AI and robotics, it’s almost time for Chat GPT,” Deepu Talla, Nivea’s vice president of robotics, told the Financial Times, believing the market has reached a “tipping point.”

The robotics push comes as Nvidia seeks to reduce its reliance on the US semiconductor giant for powerful AI chips from rival chipmakers such as AMD, as well as cloud computing giants such as Amazon, Microsoft and Google.

Nvidia, whose value has soared to over $3tn on the back. Great demand He has positioned himself as an investor in AI Chips in the ‘physical AI’ space, to help grow the next generation of robotics companies.

It was in February One of many companiesTo invest in humanoid robotics company Image AI at a valuation of $2.6 billion, including Microsoft and OpenAI.

Robotics has been a relatively new field that has yet to yield significant profits. Many startups in the space are struggling with streamlining, reducing costs, and increasing accuracy of robotic products.

Nvidia doesn’t break out robotics product sales, but it currently represents a relatively small share of overall revenue. Data center revenue, which includes its on-demand AI GPU chips, accounted for 88 per cent of the group’s $35.1bn in total sales in the third quarter.

But Tala’s transformation in the robotics market is being driven by two technological breakthroughs: the explosion of generative AI models and the ability to train robots in simulated environments based on these base models.

The latter is a huge advance, he said, as it helps address what roboticists call the “sim-to-real gap” by allowing robots trained in virtual environments to function effectively in the real world.

“In the past 12 months . . . “(This gap) has grown enough that we can now run simulations in combination with generative AI, which we couldn’t do two years ago,” Tala said. “We provide a platform for all these companies to do whatever they want.”

Tala joined Nvidia in 2013 to work on the ‘Tegra’ chip, initially aimed at the smartphone market. However, the company quickly oversaw the deployment of nearly 3,000 engineers into “AI and autonomous learning (for example, for vehicles).” This was the genesis of the Jetsons, Nvidia’s robotic ‘brain’ modules that came out in 2014.

Nvidia offers three tools in robotics development: software for training basic models, from Nvidia’s ‘DGX’ system; simulation of real-world environments in its ‘universal’ platform; And the hardware that goes into the robots as the ‘brain’.

Aptronics, which uses NaviD technology to enhance humanoid robots, also announced in December a strategic partnership with Google DeepMind to improve its products.

According to American market researchers BCC, the global robotics market is currently valued at $78 billion and is expected to reach $165 billion by the end of 2029.

Amazon has already deployed NVIDIA’s robotics simulation technology to three of its warehouses in the US, and Toyota and Boston Dynamics are among the customers using NVIDIA’s training software.

David Rosen, who directs the Robust Autonomy Lab at Northeastern University, says the robotics market still has significant challenges, including training the models and ensuring their safety during deployment.

“Until now, we don’t have the most effective tools to verify the safety and reliability characteristics of machine learning systems, especially in robotics. This is a big open scientific question in the field,” Rosen said.