NVIDIA CEO Jensen Huang was recently invited to discuss the future of AI with Goldman Sachs CEO David Solomon. In addition to sharing NVIDIA's development history and future prospects, Huang also delved into the future challenges and opportunities of generative AI, data centers, and global supply chains.

From gaming graphics cards to AI cores, NVIDIA's road to growth.

Jensen Huang first reviewed the development process of NVIDIA since its establishment in 1993. The early vision was to solve complex problems that traditional CPUs could not solve through 'accelerated computing'. It started with designing graphics cards for personal computers and TV games, and later expanded to scientific computing, AI, and data processing. Huang Renxun emphasized that maintaining 'technical architecture consistency' and 'protecting developer investment' are important cornerstones of the company's success, and have made technologies like CUDA one of the largest development ecosystems in the world.

(Note: CUDA is a software-hardware integration technology that aims to use graphics processing units (GPUs) for general-purpose computing. Developers can use high-level programming languages such as C to write programs and transfer the computing tasks originally performed by the Central Processing Unit (CPU) to the GPU, thereby improving computational efficiency.)

Traditional data center efficiency is low, Spark improves data processing efficiency by 20 times

Huang Renxun pointed out that traditional data centers are inefficient due to factors such as hardware decentralization, complex connectivity, and energy consumption. He mentioned that NVIDIA's solution is to 'centralize' and use liquid cooling technology to drop operating costs while improving energy efficiency.

Huang Renxun cited an example, a single NVIDIA server rack can replace thousands of traditional Nodes, greatly reducing hardware and connectivity costs. In addition, GPU-accelerated solution technologies such as Spark can improve data processing performance by 20 times, enabling customers to achieve a return on investment of more than 10 times.

The generative AI revolution, digital assistants work 24/7.

Huang Renxun emphasized that generative AI is not only a tool, but also a skill. AI can learn from data to generate various content, such as images, text, and protein structures, allowing AI to participate in work like a 'digital employee.' Examples include applications such as autonomous driving, digital customer service, and robots, which demonstrate the concrete application of AI skills in reality. Huang Renxun also mentioned that almost all of NVIDIA's engineers now have a 'digital assistant' working 24/7, which is a milestone that will change the future of the industry.

The cross-disciplinary AI application continues to expand, combining drug innovation and research and development.

Huang Renxun also stated that the cross-disciplinary application ability of AI is also one of its core advantages. For example, AI can convert text into images, images into chemical molecules, and even process complex drug development and virtual screening. This provides possibilities for computer-aided drug discovery (CADD) and greatly improves the efficiency and precision of pharmaceutical innovation.

(Note: Computational drug development refers to the combination of chemistry, biology, pharmacy, and information technology to discover potential new drugs by simulating, analyzing, and predicting the interaction between chemical molecules and biological targets.)

The main advantage is the ability to integrate software and hardware, and the consistent technical architecture makes development more convenient.

Facing the competition in the AI ​​field, Huang Renxun emphasized that NVIDIA's key advantage lies in the ability to integrate hardware, software, and overall infrastructure, rather than simply hardware. He emphasized that the supercomputer designed by NVIDIA is composed of multiple chips, combined with customized software for optimization, and provides more efficient and energy-saving computing power.

In addition, NVIDIA's 'consistent technical architecture' design ensures that developers can seamlessly use it on different devices (such as computers, cloud servers, etc.), which is also the reason why NVIDIA stands out in the field of AI.

Flexibly addressing supply chain challenges, highly praised TSMC

In the face of growing geopolitical tensions, Huang said NVIDIA was "on file" at the time of design, that is, to increase the flexibility of the supply on-chain and avoid the influence of a single source. He also stressed that while NVIDIA relies on TSMC as the main chipmaker, the company has the ability to shift some of its manufacturing to others if necessary. At the same time, Huang Jenxun also praised TSMC's resilience, believing that TSMC is an important driver of NVIDIA's rapid rise.

Bearing the global demand for technology, moving towards a new era of computing

Huang Renxun admitted that NVIDIA is currently facing a lot of pressure from customers because its technology directly affects customers' revenue and competitiveness. However, he believes that this is also a rare opportunity because NVIDIA is shaping the next era of computing and promoting the integration of AI technology into various industries. Huang Renxun mentioned, 'Being able to participate in this global transformation and witness the birth of these breakthrough applications is an exciting challenge, despite the enormous responsibility.'

(Siggraph Forum | Nvidia CEO Huang Renxun and Meta CEO Zuckerberg discuss the future of AI)

This article NVIDIA CEO Huang Renxun: Will lead the next AI computer era, flexible response to supply chain challenges first appeared in Chain News ABMedia.