"Please run vigorously towards your ideals; not by walking, but by running. Become the predator, not the prey."

NVIDIA CEO Jensen Huang delivered a speech as a guest at the graduation ceremony of National Taiwan University yesterday. He first greeted everyone in Taiwanese and encouraged the graduating students to learn from the predecessors who effectively utilized new technologies throughout history, and to seize the opportunities brought by the AI era.

As he often mentioned at NVIDIA GTC, he believes that AI will bring revolutionary changes and opportunities to industries such as manufacturing, transportation, healthcare, and finance. It will also be an important opportunity for Taiwan. In this era, every individual or company must make good use of AI. Those who can flexibly utilize AI will progress, while those who cannot master AI will wither away.

In addition to industrial upgrading, Huang also mentioned that, just like in the past when new technologies emerged, there will be new forms of industries and jobs, such as prompt engineering, AI factory management, AI security engineers, and data engineers.

He encouraged the graduates of National Taiwan University, "Please accomplish unimaginable and incredible things with the help of AI." The graduates present here will become the important foundation for Taiwan's AI development in the next few decades.

Huang then shared three stories of NVIDIA's journey since its inception, encouraging the graduates to understand mistakes, persevere in their will, and know when to change strategies. The first story was about NVIDIA's development of cost-effective forward texture mapping technology, which led to a contract with SEGA, a game console manufacturer. It also attracted game developers to use NVIDIA and SEGA's platform for game development.

However, Microsoft developed Windows 95 Direct3D, which made the NVIDIA and SEGA joint development incompatible with Windows. Huang had to contact the CEO of SEGA to explain the mismatch in development directions, but it required SEGA to fully pay the contract fee; otherwise, NVIDIA would face bankruptcy. At the critical moment, SEGA agreed to pay, giving NVIDIA a three-month breathing space and the opportunity to create NVIDIA RIVA 128, turning the tide of their fate.

The second point is about sticking to investment in CUDA development. CUDA is essentially a technology that allows developers to use NVIDIA graphics cards for computations other than image processing. At that time, people believed that NVIDIA graphics cards were only meant for gaming or 3D graphics, without any other purposes. However, Huang Renxun persisted in the belief that NVIDIA graphics cards would eventually surpass CPUs in dominating the computational market. In 2012, when it was discovered that some people were using the GTX580 for deep learning (Geoffrey Hinton's two students), NVIDIA realized that they would become the engine required by AI developers, and their persistence paid off.

The third point is the story of NVIDIA's painful decision to give up on the mobile phone market. Many people should know that NVIDIA had the opportunity to enter the smartphone chip market (with the Tegra series processors). However, at that time, the smartphone chip market had become highly competitive, and although their stock price performed well for a while, NVIDIA decided to withdraw from the mobile processor market and instead focus on the robotics and computer market, creating a new blue ocean. "Retreating is not easy for intelligent people like you (National Taiwan University graduates), but success also requires genuine needs and strategic retreat."

Here is the full speech text:

"Hello, graduates of National Taiwan University in 2023.

Today is a special day for all of you, and it marks the realization of your parents' dreams. You are about to leave home soon, and today is a day of honor. Your parents have sacrificed themselves to make you successful, and my parents are here too, as well as my brother. Right here, we express our gratitude to our parents.

Ten years ago, I came to National Taiwan University for the first time when I was invited to visit a laboratory. When I arrived, he showed me what they were working on—an entire room filled with NVIDIA gaming graphics cards, and there were large fans from Tatung along the hallway for cooling. He had built a supercomputer himself, using gaming graphics cards, in the Taiwanese way.

They were here, starting the earliest NVIDIA JOURNEY. He proudly said that, thanks to my support, he was able to achieve his goals. Our company's value lies in helping these individuals achieve their ambitions.

When I graduated, the world was relatively simple; there were no LCD TVs or smartphones yet. That was in 1994 when the IBM PC and MAC PC began the personal computer revolution. Your world now is even more complex, with cars starting to drive autonomously. Forty years ago, PCs emerged, and later, we invented AI, enabling computers to perform tasks automatically.

AI will create new jobs that never existed before, such as prompt engineering. These jobs were nonexistent in the past. Automated tasks will render some positions obsolete. AI will transform every job, empowering software engineers, designers, and manufacturing planners to play bigger roles. Every company and individual must learn to harness the power of AI and collaborate with it to create amazing things.

We are at the beginning of a new era, just like the personal computer and the internet. But AI is even more fundamental, and we need to learn how to write software to harness its processing power. This is a renaissance for the computer industry and an opportunity for Taiwanese enterprises.

My journey began 40 years before yours, and 1984 was a perfect year for graduation. I predict that 2023 will be the same. Let me tell you, today is the most successful day of your lives so far because you are graduating from here.

I have also experienced success in the past, but it was not until I founded NVIDIA that I encountered failures—some significant failures that were extremely embarrassing and shameful. There were many instances where we found ourselves in difficult situations. Let me share three of those stories with you.

I have encountered many failures, and let me share three of those stories with you."

Our first application was a 3D graphics application developed for PC games. We utilized a low-cost forward texture processing technique and aimed to collaborate with SEGA to establish their gaming console. However, during a year of development, we discovered that the architecture was flawed and performed poorly. Meanwhile, Microsoft was introducing their own gaming console using the opposite approach. If we had completed that console, it would have marked a significant technological catch-up with the Windows system.

But if we fulfilled the contract, we would run out of money. At that time, I contacted the CEO of SEGA and informed him that our approach was incorrect. They should find another contractor, and we had to terminate the collaboration. However, I required SEGA to fully pay our fees, or else we wouldn't be able to continue operating.

To my surprise, the CEO of SEGA generously granted us three months' time to develop RIVA 28, which saved the company. It was because we confronted our mistakes and sought assistance with humility that NVIDIA was rescued.

In 2007, we announced CUDA computing technology with the hope of enhancing scientific and physics computation capabilities. This was an extremely challenging task that had rarely been accomplished in history.

We used GEFORCE as our gaming display card, but it came with a significantly high cost. Our profits were affected for years, leading to our market value falling below $1 billion. Despite this prolonged downturn, we persevered, believing that the era of accelerated computing was on the horizon. Then applications started emerging, such as CT reconstruction, molecular modeling, and image processing, one field after another. Our developers wrote algorithms and accelerated the speed of chips.

In 2014, we trained on GPUs, triggering the explosion of AI. Fortunately, we recognized the potential of deep learning and pursued it at all risks. Years later, the AI revolution began, and NVIDIA became the driving force. We invented CUDA for AI, and this journey has sharpened our character. Enduring pain and hardship are necessary for realizing our vision.

Let me share another story. In 2010, Google developed Android with impressive graphics, and NVIDIA's computing power made us a great partner for Android. We achieved success and saw our stock price soar.

The smartphone market was enormous, and we could have competed in that market. However, we chose to step back. Our mission was to solve problems with computers, and we should dedicate ourselves to our vision. This sacrifice paid off. By leaving the smartphone market, we created a new market—a new type of computer designed specifically for robots, running AI with neural networks and secure architecture. It is a market worth billions of dollars.

Creating a robot market from a vast smartphone market, we have now opened a new industry. Retreating is not easy, especially for brilliant individuals like yourselves. Strategic retreat, deciding what to abandon, and determining the core of success are crucial.

As graduating students in 2023, just like me, you are standing at the starting line of the personal computer and chip revolution, while also entering the early stages of AI. Every industry will undergo revolutionary transformations. Be prepared to embrace new ideas, your ideas.

In the past 40 years, we have created personal computers, the internet, mobile devices, and the cloud. Now, it is the era of AI. What will you create? Whatever it is, run after your creation, just as we have done. Don't slow down; keep moving at a steady pace.

Remember, run! Either run to hunt or else you will be considered prey and run away.

When you can't distinguish the difference, whichever situation arises, run. Draw from my experiences during your journey, make mistakes, seek help, endure pain and hardship to realize your dreams, and dedicate your life to your career.