CEO Jensen Huang reveals why the company’s enterprise partnerships matter far more than consumer chatbots for long-term growth
Nvidia CEO Jensen Huang sat down with CNBC on December 1 to discuss a partnership that he believes will revolutionize how the world approaches design and engineering. The conversation revealed a strategic vision that extends far beyond the chatbots and consumer applications currently dominating headlines about artificial intelligence.
Huang joined Synopsys CEO Sassine Ghazi to announce an expanded collaboration that transforms one of the most compute-intensive industries in existence. Synopsys is pivoting its entire company to build GPU-accelerated tools on Nvidia‘s CUDA platform, physical AI and Omniverse technologies. This shift affects everything from electronic design automation to system design automation to computer-aided engineering and drug discovery.
The decade-long preparation
The partnership represents the culmination of work that began years ago when Huang first started building the software stack necessary for this transformation. For over three decades, the design and engineering industry has relied on tools that ran exclusively on CPUs. Nvidia’s technology now enables simulations at speeds and scales previously unimaginable, allowing companies to complete entire engineering projects inside digital twins before building physical prototypes.
Huang emphasized that while consumer AI applications like ChatGPT capture public attention, the real opportunity lies in enterprise and industrial sectors. He explained that getting answers 90% correct works fine for recommending movies or advertisements, but industries designing cars, planes and factories require perfection because so much hangs in the balance.
The platform shift nobody sees
A striking statistic illustrates the transformation already underway. In 2016, the world’s scientific supercomputers were 90% CPU-based and 10% GPU-based. This year, those numbers have completely flipped, with 90% running on GPUs and only 10% on CPUs. This platform shift happened in scientific computing over the past decade, and Huang believes the engineering industry is now reaching that same tipping point.
The Synopsys partnership expands significantly with the company’s recent acquisition of Ansys, which brings sophisticated engineering solutions for automotive, robotics, drones and industrial machinery into the fold. Ghazi noted that designing these increasingly sophisticated, AI-driven systems requires advanced engineering solutions to manage complexity while keeping costs affordable and timelines reasonable.
Understanding the market size
Huang provided context that helps explain why he considers industrial AI more significant than consumer applications. Industrial companies typically spend hundreds of millions or low billions of dollars on engineering software tools. However, those same companies spend 10 to 20 times more on physically prototyping products. Nvidia itself spends hundreds of millions on design tools but billions on prototyping.
The opportunity emerges when companies can prototype everything digitally through simulations and digital twins, eliminating waste from physical prototyping. This shift increases the market opportunity by a factor of 10 to 100, affecting an industrial sector measured in trillions of dollars rather than billions.
Why this matters now
When asked about concerns over AI spending by hyperscalers and data center investments, Huang reframed the conversation. He explained that the world’s major cloud providers would have made this platform shift from general-purpose computing to accelerated computing regardless of chatbots or agentic AI. Moore’s law has slowed tremendously, and the industry needs more powerful, efficient computing methods going forward.
The technology reaching industrial applications represents a critical milestone because these sectors cannot adopt solutions that work 90% of the time. They require extraordinary precision and reliability, which takes longer to develop but serves vastly larger markets once achieved.
Ghazi added that workloads previously requiring two to three weeks can now be reduced to hours through GPU acceleration, delivering transformative value to customers. Companies across automotive, aerospace and industrial sectors will need to increase their R&D spending on automation and software exponentially to remain competitive as products grow more sophisticated.
Huang also addressed competition from custom chips, noting that Nvidia’s CUDA platform offers versatility and fungibility that application-specific integrated circuits cannot match. The Synopsys partnership demonstrates opportunities only available through Nvidia’s broader platform, which operates across every cloud provider, original equipment manufacturer, on-premise installation and edge deployment.
Source: CNBC