ITRI TODAY Fall issue 2024

Jensen Huang: Front-Runner for AI Innovation

Known for his trademark black leather jacket and company logo tattooed on his left arm, Jensen Huang, the unconventional leader, has gained his popularity almost as rapidly as AI has grown in recent years. Few tech CEOs command as much attention in Taiwan as he does. During his frequent visits to the island, he shows his affection for his birthplace by appearing at night markets, endorsing local delicacies, and speaking in the Taiwanese dialect. With a charismatic charm akin to that of a rock star, he captivates both crowds and media alike.

Pioneering GPU and Ushering in a New Era of 3D Graphics

Huang, a native of Tainan, moved to the U.S. at the age of nine. By 16, he had graduated early from high school and enrolled in electrical engineering at Oregon State University. After earning his degree, Huang stayed in Silicon Valley, working as a chip engineer for AMD and LSI Logic while pursuing a master’s in electrical engineering from Stanford University.

In the early 1990s, personal computers were primarily used for academic and office purposes. Huang envisioned realistic 3D graphics on personal computers, a concept so novel at the time that it lacked a specific term. Recognizing the budding potential of computer gaming, Huang co-founded NVIDIA in 1993 with two other engineers, starting with just US$40,000. At the time, he was only 30 years old.

NVIDIA’s development of the GPU, or graphics processing unit, revolutionized graphics rendering on computers, game consoles, and mobile devices. Traditional CPUs couldn’t meet gamers’ demands for real-time, clear, and smooth graphics, which spurred NVIDIA’s initial growth. The 1999 release of the GeForce 256, the world’s first GPU, marked the beginning of a new era and established NVIDIA’s dominance in the 3D graphics chip market. That same year, NVIDIA went public on Nasdaq.

Over the past 30 years, Huang has relentlessly pursued groundbreaking GPU technology. For nearly a decade, NVIDIA invested in “non-existent markets,” seeking future success signs before the market materialized. NVIDIA navigated through waves of gaming consoles, mobile devices, and cryptocurrency mining, growing with each industry shift.

Leveraging GPUs for AI Computation

Graphics processing alone didn’t build NVIDIA’s empire. The GPU’s thousands of cores can handle extensive parallel processing, accelerating AI and machine learning processes. This has been NVIDIA’s primary growth driver in recent years. Huang actively explored the diverse applications of GPUs, launching the CUDA architecture in 2006. CUDA enabled customers to use GPU parallel processing for general computation, such as deep learning, machine vision, and scientific research.

The pivotal moment came in 2012 when a team led by renowned AI deep learning expert Geoffrey Hinton used CUDA and NVIDIA GPUs to train neural networks, winning the ImageNet competition and catapulting NVIDIA into the AI spotlight. Today, the NVIDIA GPU and CUDA platform are essential for AI development. Many companies and research institutions now use GPUs to build supercomputers, autonomous systems, and robotics, fostering the growth of Gen AI. NVIDIA now holds a 90% market share in AI chips, earning Huang the title of “AI Godfather.”

In May 2023, NVIDIA’s market value surpassed US$1 trillion, making it the first chip company to reach this milestone. As AI’s demand for GPUs soared, NVIDIA’s stock price surged. By June 2024, it surpassed Microsoft to become the world’s most valuable company, with a market capitalization of US$3.34 trillion.

Overcoming Obstacles with Unwavering Belief

Despite NVIDIA’s current success, Huang has faced numerous failures—product flaws, strategic errors, and market changes. Like many companies, NVIDIA was on the verge of bankruptcy multiple times. However, Huang’s ability to quickly adapt and persist, driven by unwavering belief, was crucial.

Developing CUDA involved high costs and significantly impacted NVIDIA’s profits for years, with its market value once dipping below US$1 billion. Yet Huang persisted, confident that the era of accelerated computing was imminent.

In 2014, NVIDIA conducted large-scale AI training on GPUs, igniting the deep learning explosion. Recognizing deep learning’s potential, they took significant risks, which paid off as the AI revolution unfolded. “We invented CUDA for AI, but the process also forged our character. Achieving our vision required enduring pain and hardship,” Huang reflected.

NVIDIA also ventured into the smartphone chip market but eventually exited to focus on robotics. Huang believes that the company’s mission is to build computing machines that solve problems, ultimately creating another new industry. “For smart people, strategic retreats are challenging, but knowing what to give up is key to success,” he noted.

Taiwan: The Epicenter of the AI Revolution

Although rooted in the U.S., Huang has deep ties with Taiwan’s tech industry. NVIDIA’s first successful chip was produced by TSMC. In 1995, Huang wrote to TSMC’s founder, Morris Chang, requesting chip production. Chang’s personal call in response marked the beginning of a 30-year partnership, with TSMC producing every generation of NVIDIA products since. Despite a 31-year age difference, Huang and Chang share a strong mentor-mentee relationship. Huang even declared, “Without TSMC, there would be no NVIDIA!”

NVIDIA’s connections in Taiwan extend beyond TSMC, including partnerships with Foxconn, Quanta, Wistron, Wiwynn, Inventec, ASUS, ASRock, and Pegatron, forming a critical supply chain. “The word ‘AI’ is in ‘TAIWAN,’” Huang pointed out, asserting Taiwan’s role in building global AI infrastructure. NVIDIA’s journey began in Taiwan, and it will take Taiwan to the world.

Huang has ramped up investments in Taiwan, where he has established a hub warehouse logistics center and NVIDIA’s first AI R&D center in Asia. The R&D project has also built Taiwan’s largest AI supercomputer “Taipei-1.” He regards Taiwan as the “epicenter of the AI revolution” and a crucial element of NVIDIA’s strategic development, enhancing Taiwan’s AI chip supply chain and advancing its growth in the AI era.

“Run! Don’t walk!” Huang told graduates at National Taiwan University’s commencement. His words resonated with many and reminded us of his dedicated pursuit in seizing opportunities and embracing the challenges presented by AI. As he frequently emphasizes, his company does not measure itself against the competition, but against the speed of light—perfection. It is his swift execution and high standards that have consistently propelled NVIDIA to the forefront of trends and towards an unlimited AI future.