A must-read for Nvidia investors＂The Thinking Machine＂

The Thinking Machine

Introduction to this book

＂The Thinking Machine＂ is about the founding history of Nvidia, Jensen Huang’s personal biography, and how Nvidia has evolved and grown into the world’s most important chip manufacturer.

The author’s writing style is fluent and suitable for most people to read. There is no difficulty, and there is no topic that will hinder the general reader. If the average person concentrates, he can finish reading it in one day. In my case, I finished reading this book in three hours over the weekend.

If you are an investor in Nvidia stock or are interested in this company, it is recommended that you take the time to read this book, because it is a very basic and complete introduction to Nvidia and will definitely be helpful to you.

Reasons for recommending this book

I have read three biographies about Nvidia or Jensen Huang from the past to now, and this is one of the better ones. The reasons are as follows:

  • This book has a rating of 4.6 on Amazon and 4.41 on GoodReads; it is not a bad book.
  • The author of this book actually interviewed hundreds of people who were directly related to the founding of Nvidia or Jensen Huang at the time, and all the interviews are first-hand records, not second-hand transmission.
  • It does not only report good news but not bad news. It does not have the common biographies of Taiwanese entrepreneurs, company founders, celebrities, or politicians that only flatter, use official documents, internal propaganda, sing praises, and conceal bad and promote good; as if everyone is a good person or the reincarnation of Buddha. This book restores many fierce conflicts in business directions and fierce disputes within the company. And Jensen Huang’s repeated angry personality, habitually deliberately humiliating subordinates in public and embarrassing them, without any mercy.
  • Personally, because of my work, I know Nvidia very well. Basically, it is not far from my cognition. This book also helped me to gain a deeper understanding of some of Nvidia’s business strategies, the mental ties among the company’s founders, and the details of why the company failed to survive back then.

My Summary of this book

Biography of Huang before he went to silicon valley

Birth and immigration

Readers who care about Jensen Huang or Nvidia are familiar with this part, including that he was born in Taiwan and still speaks fluent Minnan dialect but not Mandarin. Because of his father’s work as a chemical engineer, he lived in Thailand for a short time when he was a child and witnessed the military coup in Thailand.

School days

He immigrated to the United States with his older brother in his teens and was sent to an elementary school in the southern United States by his parents. He was bullied along the way because he was Chinese.

Jensen Huang had excellent grades in high school. After graduating from high school, he did not apply to the Ivy League schools on the East Coast because he wanted to go to a university close to home. Instead, he joined the Department of Electrical Engineering at Oregon State University, where he met his future wife.

Career

Jensen Huang joined AMD after graduation, and his wife also joined Silicon Graphics, a well-known company at the time (now acquired by HPE Technology). After working at AMD for two years, Jensen Huang joined LSI Logic as the director of core hardware design.

Jensen Huang met two engineers from Sun Microsystems (acquired by Oracle) who were customers at LSI Logic, Chris Malakowski and Curtis Prim.

Founding Nvidia

Consumer graphics chips were a red ocean

Since Sun Microsystems only wanted to stick to the high-end enterprise graphics chip market with higher profit margin and had no interest in investing in consumer-grade graphics chips, Jensen Huang, Chris Malakowski and Curtis Prim decided to start a company to invest in this field, which was already quite crowded at the time. Jensen Huang consulted consultants and industry insiders at the time, and almost all of them advised him to give up the idea because there were 30 to 50 companies competing in the consumer-grade graphics chip field at the time, and the chance of success was very low.

Regarding the graphics processing unit (GPU), please click to read my previous article “Top vendors and uses of GPU” and “Comparison of AI chips GPU and ASIC“

Three founders

Jensen Huang served as CEO in charge of business, Chris Malakowski as vice president of engineering, and Curtis Prim as chief technology officer; the three had the same shares, but only Jensen Huang joined the board of directors, which laid the fuse for Curtis Prim later. “Don Valentine, founder of Sequoia Capital, father of Silicon Valley Venture Capital” Because of the endorsement of Jensen Huang’s former boss at LSI Logic, he invested millions of dollars in seed funds in Nvidia and took a seat on the board.

The company’s early days were not smooth

Launched NV1

Because Curtis Prim insisted on using the quadrilateral texture architecture that no one in the industry used in the design architecture of the company’s first-generation graphics display chip NV1, abandoning the triangular texture architecture, it was incompatible with the industry, and Microsoft’s DirectX also abandoned it and did not support it. In the end, it was crushed by a flood of bad reviews and returns.

Due to the defects in the design architecture of NV1 products, a large number of returns almost caused the company to go bankrupt, causing the company to be laid off from more than 100 people to about 30 people. Cash was almost exhausted, which is why Jensen Huang later emphasized that the company was once 30 days away from closing.

Jensen Huang and Chris Malakowski still think that this product is a disaster for NVIDIA, but Curtis Prim still refuses to admit his mistake; this is another key factor that led to Curtis Prim and NVIDIA drifting apart.

Sega once saved NVIDIA

When NVIDIA was about to go bankrupt, Sega decided to continue to trust NVIDIA and let NVIDIA get the life-saving money to complete NV3. Sega saved NVIDIA with NV3, and the product was a great success.

Started to dominate the video game display card market

Launched GeForce

The launch of GeForce was the most important move of NVIDIA in the non-AI era. Because after the launch of GeForce, NVIDIA has monopolized the graphics card industry and led all technologies until today.

Eliminate all competitors

No matter the approximately 50 graphics card manufacturers of all sizes when the company was founded, or the once trend-setting S3 (acquired by VIA), and 3dfx, none of them were spared. The most wronged among them is 3dfx, because although 3dfx won the patent infringement lawsuit against Nvidia, it could not survive and was eventually acquired by Nvidia.

Only the Radient series SURVIVEd

Now only ATI’s Radient series survives in the market, but ATI also failed to survive and was acquired by AMD. However, the market share gap between the two is very large, about 1 to 4; AMD’s Radient series can be said to pose no threat to Nvidia at present, because the market share, technology, funds, and reputation of the two are not at the same level.

Several breakthroughs

Launched the Quadro series

With its leading technology and capabilities, Nvidia then launched the Quadro graphics display chip series for the professional market and workstations, further swallowing up the workstation, professional, and advanced graphics display market. So far, Nvidia can be said to have unified the graphics display market around the world, and there is no strong competitor.

Parallel Processing

Nvidia’s greatest breakthrough in chip computing is the first introduction of “multi-pipeline parallel processing” on its graphics chip. This epoch-making technology not only pioneered the technology of graphics display, but also brought game display into another new world.

More importantly, the “multi-pipeline parallel processing” capability also proved for the first time that computing chips with parallel processing capabilities can be popularized and promoted to ordinary consumers; sweeping away the high prices encountered by all previous parallel processing chip manufacturers, which made it difficult to promote and none of them could survive.

It is also worth mentioning that it was Nvidia’s foresight in the past that first introduced the “multi-pipeline parallel processing” capability on the graphics chip, which created the fundamental reason why Nvidia can now dominate the artificial intelligence chip market and redefine chip computing.

Cuda

In addition to Nvidia’s first introduction of “multi-pipeline parallel processing” on the graphics chip, Cuda is the company’s most visionary invention to date. Without Cuda, there would be no NVIDIA today, because NVIDIA uses the Cuda software interface to allow users of NVIDIA’s display chips to fully access all functions on NVIDIA’s display chips. Of course, the most important thing is that it allows users to use the Cuda interface to use the “multi-pipeline parallel processing” function on NVIDIA’s display chips.

Cuda defines the market differentiation of NVIDIA’s display chips. This is also the most fundamental reason why the developers of artificial intelligence programs around the world cannot get rid of NVIDIA chips to this day. There is no other reason.

Regarding the importance of Cuda, please click to see my previous article “How does CUDA strengthen the moat of Nvidia’s monopoly?“

Artificial intelligence changed NVIDIA’s fate

Thanks to the discovery of two ai experts

This book details how two artificial intelligence experts discovered that they could use NVIDIA’s GeForce graphics card to assist artificial intelligence neural computing, reducing the calculation that might have taken hours in the past to 30 seconds. This important discovery immediately swept the academic and technology circles and changed NVIDIA’s fate.

For this story, please click on the description of my two previous posts of “Geoffrey Hinton, 2024 Nobel Physics winner, inadvertently helped Nvida transform to AI overlord” and “2024 Nobel Chemistry Prize awarded to 3 AI experts, accurately predict the 3D structure of proteins“

Once-in-a-lifetime opportunity

The two artificial intelligence experts are now using Nvidia’s GeForce graphics card, which can greatly shorten the time of artificial intelligence neural computing, which has made a big step forward in the evolution of artificial intelligence. It is also the main reason why we see various artificial intelligence applications emerging like mushrooms after rain in recent years.

At the beginning, the two artificial intelligence experts were very excited to tell Nvidia about their discovery, but Nvidia did not respond immediately or make any statement at first. Until the matter gradually fermented in the academic circle, the technology community, and the media; through the tips of experts around him; Jensen Huang immediately realized the importance of this matter, immediately cleared all schedules, and spent a few days on weekends studying artificial intelligence-related materials. A few days later, he determined that this would be a once-in-a-lifetime, rare opportunity that would change Nvidia and the industry.

Since then, Jensen Huang has put all his focus, company resources, and funds into this field; he has worked year-round, thus creating the world’s first true artificial intelligence empire.

Launching AI data center chips

Due to the strong market demand, general consumer-grade Nvidia display chips cannot meet the needs of professional artificial intelligence manufacturers. Please click on the explanation of my previous post “Data center, a rapidly growing semiconductor field“. Nvidia took the opportunity to launch the DXG and HGX series of server chips specifically for artificial intelligence data centers. This is the major event that changed Nvidia’s destiny, becoming the largest source of revenue for Nvidia that you and I see now, and also making it the world’s largest listed company in terms of world value.

Support for ai software

Jensen Huang invested resources very early, allowing Nvidia to support Cuda for artificial intelligence from the beginning, and launched NVIDIA CUDA deep neural network program library cuDNN software developers develop programs related to artificial intelligence. cuDNN is a GPU-accelerated primitive library for deep neural networks. cuDNN is also a major reason why Nvidia is loved by AI developers, because cuDNN can save them a lot of effort, and AI developers are of course happy to adopt it.

In addition to cuDNN, Nvidia has also launched many development kits that directly support AI developers, such as DNN/RNN compilers. Due to the rapid progress of AI, Jensen Huang immediately asked the company’s team to abandon the DNN/RNN compiler that was being developed and enhanced, and turned to the development of the Transformer compiler on the same day. And now the more important inference compiler (Inference Compiler). These examples have repeatedly proved that there is a reason why Nvidia has no competitors in the basic device chip computing of AI.

Two magic weapons to get rid of opponents

NVLink

NVLink is a bus and its communication protocol developed and launched by Nvidia. NVLink adopts a peer-to-peer structure and serial transmission, which is used for the connection between the central processing unit (CPU) and the graphics processing unit (GPU), and can also be used for the interconnection between multiple Nvidia graphics processors. This Nvidia-exclusive agreement can be said to be the driving force behind Nvidia’s victory over other competitors.

InfiniBand

InfiniBand is a technology that Nvidia acquired through supply. It is a high-speed, low-latency computer network communication standard, mainly used in high-performance computing (HPC) and data centers. It is designed to achieve efficient data interconnection within and between computers, providing high throughput and extremely low latency.

The combination of InfiniBand and NVLink has made Nvidia’s competitiveness in the moat of artificial intelligence even stronger, leaving potential competitors facing a situation where they are almost unable to fight back. For InfiniBand and NVLink, please click on the explanations of my previous two posts “The reasons for Nvidia’s monopoly and the challenges it faces” and “Nvidia’s new business set to grow its share price“

Other Fields

Cryptocurrency Mining

Due to the powerful performance and parallel processing capabilities of Nvidia graphics cards, early Bitcoin and other mining machine operators saw business opportunities and switched to Nvidia graphics cards for mining, which also created another important use for Nvidia graphics cards in addition to games, displays, and artificial intelligence.

Jetson

Jensen Huang promoted Nvidia’s Jetson platform, which is specifically used for embedded platforms, and based on it, promoted various applications that require power saving, miniaturization, and simplification, such as artificial intelligence, self-driving cars, and robots.

Omniverse

Omniverse is a platform for Nvidia to build and operate 3D pipelines and virtual worlds, which can achieve real-time, physically accurate simulation and collaborative workflows. In fact, Nvidia already had the core technology of the metaverse before META. Now Nvidia is pushing Omniverse to the actual application of digital twins in various industries. It can replicate warehouse center processes to test human-machine interaction and then activate specific robot functions in a real-time environment.

POACHed Nintendo from AMD

The three major game handheld companies have always used AMD. Nvidia uses its leading technology in display technology, embedded systems, miniaturization, and artificial intelligence; plus the powerful Tegra platform that Nvidia has developed on mobile devices before, to convince Nintendo, the largest game handheld company, to adopt its own Tegra platform.

The Thinking Machine

Related articles

Disclaimer

  • The content of this site is the author’s personal opinions and is for reference only. I am not responsible for the correctness, opinions, and immediacy of the content and information of the article. Readers must make their own judgments.
  • I shall not be liable for any damages or other legal liabilities for the direct or indirect losses caused by the readers’ direct or indirect reliance on and reference to the information on this site, or all the responsibilities arising therefrom, as a result of any investment behavior.

Author: Andy Lin

Stock Investment Books Published English Blog English Simple Bio English Website Facebook Twitter Account LinkedIn Account 股票投資著作書籍 中文部落格 中文簡歷 中文網站 臉書 推特帳號 領英帳號

error: Content is protected !!