Shulou(Shulou.com)11/24 Report--

This article is from the official account of Wechat: knowledge Automation (ID:zhishipai), author: Miao Qihao

Generally speaking, the English written by the British is often not as straightforward as that written by the Americans, such as "Chip War". And the title of this book "Blueprint of all things" has been pondering for a long time.

However, after reading the first chapter Everything,everywhere, it is easy to understand, which means that the design of ARM provides a design blueprint (instruction set) for chips across everything.

A serious production written by people in the media. It is interesting to compare this book with "Chip Wars" in appearance. "Chip Wars", written by an associate professor at the university, is described as a "non-fictional thriller" with a "cast list" with multiple photos in the middle, which is highly readable. At the end, there is no need for the norms of academic works, while the references are listed in the Notes.

The author of Blueprint of everything is an out-and-out media reporter with neither relevant professional background nor academic work. In his preface, the author wrote that his aim was to write the "biggest Business Story" (biggest business story) and suddenly thought of the chip.

But the book is much like an academic work, listing 32 pages of references (Reference) and 16 background monographs (Bibliography).

In recent years, most of the monographs on chip wars published in China emphasize their technical background, which are written by only one reporter, but they look like a collection of newspaper clippings, which is really unflattering. It feels that technical background and media background should each have their own strengths, and investigative journalists have their own unique perspectives and skills of mining, interview and analysis, which can not be replaced by technical experts.

The 2006 speech by Robin Saxby, honorary chairman of Amou, with the title "Chips with Everything", may be the basis for the author's selection of the title of the Blueprint for everything.

The first two chapters of this book are the background, one is about the science popularization of chip history, and the other is about the two completely tit-for-tat personalities of "Odd couple", Jack Kilby and Robert Noyce. Two engineers, who had never met at that time, invented the integrated circuit almost at the same time. The third chapter of "Chip War" is about the two of them. The author of the other 13 chapters is divided into three parts. The first part describes the source and commercialization process of ARM technology, as well as its key role in mobile communication. The second part describes how ARM leapt from old-fashioned phones to smartphones. In this process, there is not only a fierce competition with Intel, but also a love-hate relationship with Softbank Corp.. The third part describes its development in the market of low-energy sensors and high-end data centers, and how to avoid the trap caused by its market generalization. In addition to these three parts, there is a fourth part, which tells the story after 2022. I feel that the first draft may be completed early, and when some publishers like it, I will fill in the updated content again.

The legend of BBC nurturing talks about the tough start of chip bosses. Silicon Valley, garage innovation, venture capital and defense orders are familiar, but that's the American story. Amou ARM (Advanced RISC Machines), its predecessor is Acorn Computer. It was not the financial tycoon or the Ministry of Defense that helped the rise of the unknown Acorn, but BBC, the state-owned television station. In the late 1970s, the continuing education department of BBC saw that computers were leading the future, so they sent reporters to visit many countries in Europe, the United States and Japan to learn about the development and produce a number of TV series. BBC even acted as a think-tank, sending a 50-page report titled "Microelectronics" directly to every congressman's desk in December 1979. The report defines the emerging areas that multimedia education needs to address and believes that this will have an impact on the coming decades.

The report had a good persuasive effect, with the UK government setting up a project to equip computers in primary and secondary schools, while Acorn's low-cost brand was lucky enough to be included in one of the two product lines with 50 per cent government subsidies. For this reason, it produces a series of computers that are popular in primary and secondary schools, and even uses the brand of BBC Micro (BBC Software has been established to support it). Some people call Acorn "the Apple of Britain". In fact, it's hard to say how similar they are. This unique experience shows that without BBC, we might not be where we are today.

Acorn barely gained a foothold under the support of BBC, but after all, it was only one of many companies competing in the computer market, and later suffered twists and turns, and was even rated as one of the "top ten zombie IT companies" by the important IT media.

However, ARM has seized on an important key to chip development, namely reduced instruction set computing (RISC). The chip uses only 20% of its capacity most of the time, and if there is a design architecture that can reduce instruction set redundancy, the chip performance will be greatly improved. The concept has long been studied in IBM, Berkeley, Calif., and Stanford.

Acorn launched the highly secret "Project A" in 1981 to study the design and implementation of this concept. Acorn RISC Machine was born in October 1983 and the three-letter acronym ARM appeared for the first time. The first chips based on ARM were made by its partner VLSI in April 1985. The ARM1 integrates 25000 transistors, but its performance surpasses Motorola's latest product, which has 10 times the number of transistors, which is the predecessor of Apple's mobile phone processor.

However, this amazing news was not reported by the media until six months later, not because the reporters who received the information in the first place did not believe it at all. Not only journalists, but also industry giants Ruyi French Semiconductor and Philips of the Netherlands, which have worked with Acorn in Europe, have also missed business opportunities and stood idly by in the face of Acorn, which is in urgent need of funds to expand. The company had to set its sights on the other side of the ocean and found Apple and VLSI in Silicon Valley and set up a joint venture. Apple contributed 1.5m, 43 per cent of the shares, 250000 of the shares, plus 43 per cent of intellectual property and manpower, and the rest belonged to the producer VLSI. Acorn RISC Machine was renamed Advanced RISC Machine at Apple's request, and ARM was founded in November 1990. Millions of pounds is nothing in the IT industry, and few would have thought that ARM would become a global hegemon that provides a "blueprint" for the underlying chip structure.

Many people will think that Amou is the leader in the chip design world, but its name is not seen in the ranking of global chip design companies. It is also a term for the book Chip Wars, which is explained more clearly. There are subdivisions in the chip design, and Amou belongs to a more upstream instruction set (which is why the title of the book uses Blueprint). It's like building a house, having a blueprint and then designing it. ARM sells the architecture design to the design company, which then designs the chip according to the customer's needs.

The strange model makes money from intellectual property rights. An Asmaier lithography machine costs $120 million, but why can you set up a security company for less than £3 million? The reason is that the latter "produces" intellectual property rights (IP for short). When it comes to a joint venture with Apple, Acorn relies on IP and human resources for £1.25m. IP looks like a good deal, and customers have to pay Anmou an entry fee (millions to tens of millions of pounds) for its "blueprint". If Fabless, a chip designer, uses Anmou's underlying architecture and makes finished chips, it will also have a royalty of 1-2 per cent of the price. According to Softbank Corp. 's 2022 annual report, the revenue of Anmou in 2021 was as high as 2.6 billion US dollars, a growth rate of 34.6 percent. On the other hand, some of the underlying architectures have strong vitality and can be extracted for a long time. For example, ARM7 TDMI, which was sold to Nokia and Texas Instruments in 1997, still produced 200 million films until 2020, which seems to have been lying down for more than 20 years to make money. It can be said that Amou has entered a blue sea, but why don't others grab a bite of such a good thing that pie falls in the sky? The fact is that the blue sea is not calm, and entering no man's land is bound to face great risks.

Figure: is it easy to make money from IP for ARM chip shipments? Shortly after its formal establishment, an internal company meeting on December 18, 1990 revealed an analysis of strengths, weaknesses, opportunities and traps (SWOT). Among the various weaknesses listed, the most striking one was that the company had zero patents at that time, so it was not an easy start. As early as 1981, the team led by Stanford professor John Hennessy has taken the lead in RISC technology. the MIPS it developed has been adopted by commercial workstations, and the company has already been listed. What Amou is trying to overcome is the low-cost, more widely used RISC architecture. Before partnering with Apple, there was a shortage of R & D funds. The three main designers could only share a limited Apollo computer and had to go to Munich to do the test. when the team was most desperate, it felt that it might disband at any time, and one of the important founding members failed to survive the pre-dawn darkness and resigned to become a professor at the university at the age of 37. Making a chip can be sold by testing and mass production, while it usually takes 18 months for an invention patent to be approved, while downstream design companies take about the same time to finalize, and it often takes 4-5 years for chips to be mass-produced and sold.

Obviously, Amou's business model is slow burn. According to a 2019 article in the Economist, it will take 300 man-years to design a new CPU architecture. Obviously, this is not something that all colleagues can handle. The persistence of putting 42% of your income into Redd is even more expensive. This comes from the correct judgment of the direction of research and development, and it is necessary to have a long-term vision to ensure that the technology does not lag behind after 4-5 years. Moreover, to be able to have a strong adaptability to the application prospect, it should be easy to expand. Need to see the future, but also to withstand loneliness, which is a kind of suffering for many enterprises.

Silicon Valley's high-tech companies seem to give the impression of being rich and burning money. And Amou has been short of money for a long time, not only because of the aforementioned lack of funds for research and development, but also in an old barn. The company also planned to move to a visionary industrial park full of high-end companies near Cambridge to show the "image" of high-tech start-ups.

CEO, known as a "salesman", was determined to settle down in a cheap place, so he found the old building that used to raise turkeys. In order to save money, Amou made an exchange with the furniture company. Amou put a picture of the company's desk in a 17th-century barn and printed it in a marketing manual. In return, Amou can buy solid wood at the price of plastic. When it is difficult for employees to stop getting a raise, they can only expect options to bring some compensation at some point in the future.

However, when it comes to spending money, there is no discount. Amou paid 10,000 pounds to hire someone to write a 600-word application paper. Poverty does not limit imagination. Limited financial resources force Anmou's business model to be highly focused, focusing only on the industrial ecology. It is this factor that has prompted Amou to design a very simple structure, reduce costs and improve energy efficiency, thus struggling to develop a unique low-cost energy-saving solution in the highly competitive RISC architecture. It turns out that this is precisely the magic weapon for the huge target market it will face, which is the "poor" low cost and high energy efficiency.

Anmou and Apple complement each other and even its birth benefits from the discerning eye of Apple, but cooperation is not an one-way favor. When it was still in Acorn, Apple sensed the company's extraordinary capabilities. Amou had a unique RISC architecture that surpassed that of AT&T, the largest chip company at the time. The first successful collaboration between Apple and Amou was a personal digital assistant (PDA) named after Newton, and the chip of the ARM610 architecture became Newton's trump card.

SWOT analysis done at the beginning of the establishment of Amou identified "portable applications" as "business opportunities", and it is expected that its huge market in the future will need large reserves and cheap and energy-saving chip architecture. Although Amou developed the 600 series ahead of time, this product was born at an untimely time. It is considered to be the embryonic form of iPhone's idea, but it is too advanced to keep up with the relevant supporting technology. As a result, the day when Apple and Amou worked together was the end of Newton's PDA, which led to the collapse of Apple betting on Newton (which was an important factor in Jobs' return). Fortunately, Amou's equity gains are not enough to shut down Apple. Therefore, Apple also admitted that the greatest achievement of the Newton project was to find a solution. Just since then, the field of mobile communications has ushered in an explosive period, and this cooperation has also opened a great victory for both sides in this field. Apple has become the leader of the IT industry because of iPhone, and the behind-the-scenes mou has also become the overlord of chip architecture.

Low-profile executives and high-profile capital-all the most famous technology companies in the world seem to have famous leaders. Amou has no stars. Naturally, the name with the most information from Anmou is Japanese investor Masayoshi son. He can't be regarded as a planner at all, but he bought Anmou for 24 billion pounds in 2016. The capitalist is neither on the board nor involved in specific business. Amou yearns for independent and promising designs, but there is no leader. It always does things first and then goes to the leader, which seems a little strange. Of course, the book also describes many senior executives and technical backbones of Anmou, one of which is the first CEO Robin Saxby. He has a technical background, but has been described as a "strategic salesman" several times in the book, making a significant contribution to Anmou's intellectual property-based business model. He realised presciently that "standards with broad appeal must be established and users can be found in different industries and regions". Amou's success strategy almost follows this line of thinking. Amou's chip architecture is widely used in the Internet of things. Because of its high performance and low price, South Korean fishermen are said to have installed Amou chips in fish farms to monitor water quality and nutrients. The other is Simon Segars, who wrote the best-selling ARM7 TDMI architecture for more than 20 years. He served as Amou CEO from 2013 to 2022, and it was during this period that he really established the global dominance of chip architecture.

In contrast, Softbank Corp. is much more high-profile. The story of Softbank Corp. and Anmou can be read a lot in the media. Son has called Amou the "center of the universe" for its business. In the 2017 annual report, he even said, "one day in the future, when I look back on the acquisitions and investments I have made in my life as an entrepreneur, I believe that Amou is one of the most important examples." However, Softbank Corp. 's acquisition of Anmou was interpreted as a roller coaster. Acquired in 2016 and delisted at the same time. It plans to go public again in 2019 and change its mind in early 2020. Finally, he wanted to change hands of Nvidia and was blocked by the anti-monopoly law. Even Qualcomm, an American chip behemoth, opposes it. It is widely rumored that Amou will still go public in the autumn of 2023. Despite the constant reversal of the plot, the ambition of capital predators to make huge profits may objectively become an unusual driving force for technological development. Softbank Corp. sees the rising tide of the Internet of things and is keenly aware of the key role that Amou will play in it. After Softbank Corp. acquired Anmou, he immediately asked for more R & D and even reduced its profit margin to zero, in the words of this book, "even to achieve the break-even point." If it is a listed company, there is no doubt that it will be severely punished by the collapse in stock prices, and Softbank Corp. dares to spend most of his profits on recruiting talents into research and development in order to seize the technological highland. But it can not be said that Amou is a plaything of capital, it is more like a hot potato in Softbank Corp. 's hand, knowing it is a good thing, but dare not toss about it easily. Son originally had a five-year listing plan, but recognizing that Amou's business cycle was as long as 10 years, he chose to abandon the listing and find an acquirer, Nvidia, who was equally visionary and heart-eating. Some quality of Amou played a key role.

Based on accurate positioning, ambition Anmou insists on being based on the front end of the design, but it is also ambitious. In the PC market monopolized by Intel, ARM is determined to take a different approach and set its low-power chip architecture as the world standard in embedded devices. "at a time when too many peer companies in the UK are afraid of preached modesty, Anmou has set a bold goal." In this already refined link, Anmou will deal with refined downstream users separately by creating or holding subsidiaries. For a specific device, such as memory, Anmou sells a series of copyrights in blocks. This is called "physical" intellectual property. In doing so, it is not only better to meet the needs of users, but also once a technological progress leads to product changes, the ecosystem is also easy to absorb local losses. With the rapid development of manufacturing technology, it is very important to produce new ideas and processes in time.

Anmou and TSMC have some similarities, both of them define their roles as enablers and partners. And the starting point is to find and firmly build their own small ecology in the large IT ecosystem. It has become an indispensable existence in the large ecosystem by serving the large ecosystem to become bigger and stronger. To some extent, Asmael also falls into this category. When people talk about legends such as Internet giants such as FAANG, (Facebook, Apple, Amazon, Netflix and Google), an alternative technology giant like Amou is also worth paying attention to.

Summary: it is not easy to read the original book, but this one is more tiring to read. I'm afraid there are both my proficiency and the author's writing problems. I feel that the text of this book is rather difficult, and the layout of the whole book seems to be a little messy. So far, apart from some promotional words posted by publishers, only the British hometown Financial Times has published a book review and rated it as the technical Best Summer Books of 2023. Major media such as the New York Times, the Wall Street Journal and even McKinsey's recent dozens of summer recommended books have not mentioned the book. The more well-known commentator seems to be Miller, the author of "Chip Wars" (at the invitation of the publisher).

The best part of the chip craze has been going on for several years, and the masterpiece seems to be too early, while the new "Chip Wars" and "Blueprint of everything" are still worth reading. Finally, the book is summed up in two words: should be read and difficult to read.

A brief introduction to the author

Miao Qihao: former researcher of Shanghai Institute of Science and Technology Information, personal WeChat account "Laoshugen2023"

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.