Nvidia CEO Jensen Huang on Ruben Chip, AI Demand, and Market Outlook
Jensen Huang who serves as Nvidia CEO shares his insights on the rising demand for artificial intelligence and the development of the new Ruben chip. Huang confirmed that the company faces ongoing demand for its AI chips which resulted in creation of a $500 billion backlog that exceeds the previous figure from a few months ago. Huang provided details about the company's upcoming developments through his interview which detailed how AI technology will grow and about current market conditions.
AI Expanding Beyond Language Models
Huang explained that growing demand for artificial intelligence is becoming visible because AI is now being used in different industrial sectors. He explained that AI technology has progressed beyond its original language model applications which now extend through healthcare and drug discovery and manufacturing operations. He pointed out that scientists have achieved major strides in physical AI because artificial intelligence now comprehends both causality and physical laws which enables scientists to develop:
- Self-driving cars
- All types of robotic systems
The 'Ruben' Chip A Leap in Efficiency and Cost Reduction
Huang presented information about Nvidia's upcoming chip design which engineers refer to as 'Ruben'. He asserted that Ruben delivers an exceptional upgrade over the existing Blackwell architecture:
- The system achieves 10 times lower energy consumption compared to Blackwell.
- The system enables token production at 10 times lower expense. He showed that the Blackwell system provided a 10 times improvement over Hopper through his explanation of how this system worked.
Huang stated that the company achieves yearly increases in energy efficiency while decreasing token production expenses. "This way AI can continue to be affordable."
On Selling Chips to China and National Security
Huang presented his organization through multiple strategies which he considered necessary to evaluate the dangers stemming from selling chips to China. He asserted that China's domestic AI chip sector functions effectively because local enterprises including Huawei and many emerging companies create their own advanced chips. He explained that the H200 serves no military purpose for China since they develop their defense systems with local chip technology just as the U.S. military does.
Huang stated that the United States needs to achieve victory in all AI technologies through its entire development chain which includes energy production and chip creation and infrastructure development and model building and application design. "We can't concede any market," he said because permitting American companies to compete across international markets creates revenue streams which maintain U.S. technological dominance.
Dismissing Talk of an AI Bubble
Huang denied that the artificial intelligence sector functions as an economic bubble. He claimed that present spending constitutes part of an enormous infrastructure project which historians estimate will reach $85 trillion within the next 15 years.
He provided the following evidence to support his claim that no bubble exists:
- Nvidia GPUs are completely rented out across all major cloud platforms which include Azure, AWS, GCP and CoreWeave.
- The market value of Nvidia GPUs which were sold in past years has increased because demand for them has reached such extreme levels.
- Blackwell and Ruben systems currently experience "incredible" demand according to enterprise sources.
"We're nowhere near a bubble right now," Huang concluded.
