I had already understood that "MOS" was the future. The bosses hadn't understood anything about the microprocessor. I've always fought against these forces, until I said enough, now I'll start my own company and that's it.
We were 11, three years later we were 1100. And when I left we had 40-50 million in revenue. I came home richer, but also a bit disappointed. It was clear that the "business" as I imagined it wouldn't be possible.
We invented "touchpads", "touchscreens", which changed the world of how these things are done. Simply while trying to fall asleep, in a moment, a crazy energy comes out of my chest.
He is Federico Faggin and in 1971, exactly at 30 years old, he designed and built the first microprocessor in history, the Intel 4004. Without him perhaps computers would still be as big as cabinets. We wouldn't have "personal computers", "smartphones" or "smartwatches". Not only that, but years later he would also invent the "capacitive touchpad". And his company, now worth billions, would also commercialize the "touchscreen". During my last trip to the United States I had the opportunity to interview the Italian legends in America, who lived through the history of "Silicon Valley". Including Faggin, who made the microprocessor possible, and Alberto Sangiovanni Vincentelli, who made its mass production possible. So let's start from the origins.
Federico was born in 1941, right in the middle of World War II. And during the war period he grew up in "Isola Vicentina". Do you have memories of the war?
I remember when someone entered the house, later they told me it was a German soldier fleeing to Germany, right? He entered the house with a rifle and asked for food. I remember noticing fear in my parents. Fortunately he just ate and left.
Good! Then I remember when the Americans arrived sweeping away the remnants of the Germans, going towards Austria and then Germany, they stopped in "Isola" and I remember I was on my uncles' balcony overlooking the "Isola Vicentina" square, I remember I had a little flag that I waved like this and there was a soldier watching me from the tank turret and smiling at me.
First contact with America. A good omen. Yes, exactly.
And do you remember what you wanted to do in life when you were little?
As a child I had a passion for airplanes. The idea of a plane flying was inconceivable to me. The first model airplane I saw was a revelation for me. I was 11 years old, 10-11 years old. I had never realized that even a toy could fly, right? I thought only big planes fly, right? And when I saw a toy fly it was... Wow what joy! And I ran after them, wanted to understand how it was made, etc. And then I immediately wanted to make one.
The guy who was an 18-20 year old who was flying this model, gave me some tips, right? He told me look you need to buy a book. And the third one flew. So I learned bit by bit to really make a product. Because when I was 12 years old I imagined what I wanted, I drew it. When I liked it I made the plans, then bought the material, made the parts, the pieces, assembled them. And then I made it fly. So the whole process from conception to product success.
In high school Faggin became a radio technician. And after, before going to university, he gained experience at "Olivetti".
I became a radio technician in the last two years. I was interested in computers, so I studied computers, those books I found. And they were mostly in French at that time. And then I became interested in "transistors" that had just come out, because we're talking about '55. From '55 to '60 I did five years as a technician. I had shifted my interest from airplanes. I wanted to become an aeronautical technician, but there was no longer that specialization, so I had to fall back. I became a radio technician, so I could make a radio-controlled model airplane.
When I graduated I had some job offers. I chose "Olivetti", I learned "transistors" well, I learned computers well, how they work, etc. And then I was assigned a project. Lucky for me my boss got hurt. At a certain point the project had to be finished and two thirds of the project was missing. And he told me, well, read these books, when I come back we'll see together how to design the rest of the computer. And when he came back I had already designed the rest of the computer. I gave it to him, after a week he says: "yes yes, fine, build it". I built it, a small experimental computer. And I built it, I had four technicians working for me and it worked. This experience really directed me for the rest of my life, I must say. Without that experience I couldn't have done what I did later. For example I couldn't have made the microprocessor.
After the experience at "Olivetti" though, Faggin went to university. And he was curious to understand how "transistors" work. Not from an engineering point of view, but from a physics point of view.
And so I wanted to study physics instead of engineering. Because in engineering I understood how things are made. So I didn't need engineering which is more application. I wanted to get to the bottom of the theory. So I studied physics. Then I worked at "SGS Fairchild", I developed their first "MOS" process. And they sent me here for what was supposed to be six months and I'm still here.
After a few years you entered this house which is the same for almost 50 years, right?
This one yes, this one is 48 years old, yes.
You entered "SGS Fairchild" in Italy, which was an Italian semiconductor company, but in "sisterhood" with "Fairchild". And then the original "Fairchild" called you to America. So, after physics Faggin goes to "SGS", "Società Generale dei Semiconduttori", an Italian company at the time partially owned by "Fairchild Semiconductors". "Fairchild" isn't a random company, but in '57 it was born from the best eight engineers of William Shockley, the inventor of the "transistor". Who by the way had his office at Stanford, right where the office of an Italian professor I interviewed is now located. Anyway, these eight engineers were called the eight traitors, because in reality Shockley was notoriously an asshole and they more or less escaped. Did you have the opportunity to meet the founders at "Fairchild", the famous traitors, the eight traitors?
Not all. I knew, I met Noyce, Moore. Moore, naturally, because he was the head of the laboratory where I worked in '68. Then he and Noyce started "Intel". Taking the technology I had developed, I knew Kleiner, I knew Jean Hoerni, I knew, well, 4-5 of the eight traitors.
Exactly. As he said Noyce and Moore then founded "Intel", where Noyce was the first "CEO", and Gordon Moore went down in history giving his name to Moore's law, that the number of "transistors" per chip doubles every 18 months. In fact, as I also said in the video with Vincentelli, if once chips had few hundred "transistors", today they have hundreds of billions. And if this is possible it's precisely because of the invention of the microprocessor. The first step in that direction Faggin took at "Fairchild", studying and working on chips with "MOSFET" transistors, different from the more popular chips based on bipolar "BJT" "transistors", which were the standard for computers in the '60s, as they were easier to build, but they were also larger and consumed more energy than "MOSFET". However, chips based on "MOSFET" at the time did not become popular, because they were too slow and too difficult to produce. In fact they required a layer of ultra-pure silicon oxide, and in the '60s the technology wasn't yet advanced enough to do it reliably. Despite this, Faggin had the intuition that perhaps certain barriers could be overcome and that this type of chip was the future.
I had already understood that "MOS" was the future, even though most of those working in semiconductors considered "MOS" as a little thing. I mean, "MOS" had a delay time, I don't know, of microsecond and a half, two microseconds, while "transistors" had ten nanoseconds. I mean, how can you compete with such a fast technology? Instead "MOS" made it. For me it was instinctive, it was a simpler thing. So, being simpler, it was understood that you could reduce the dimensions, so for me that was the future, even though I never really bothered to plot the future.
To understand better, let's explain better how a chip is built today. Everything starts from a "wafer", that is a disc of pure and very flat silicon. Through a process called photolithography, the chip design is practically drawn several times on the silicon wafer through UV rays that pass through a mask, which is nothing but a "stencil". Various layers are composed on the wafer that are impressed through different masks. This happens with very expensive machines with nanometric precision. At the time, instead, they were much more rudimentary. And the first obstacle for these chips was the alignment of the masks. Because, between one layer and another, you have to be very precise to have the masks aligned. Faggin at the time found a way to align them automatically.
When I had the opportunity to improve the technology, because there was this self-aligning process that was intuitively much better, but nobody understood how it could really be done, and I had the winning idea, that is, precisely the geometric game of the masks, how to make them, so as to do it. Nobody had figured it out. I did everything from start to finish, starting only from the idea that I can use silicon as a gate, which had already been verified by "Bell Labs" a year before. And this technology is what really revolutionized semiconductors, because it allowed to make all the missing pieces, which allow to make a computer with the same technology, in a monolithic piece of silicon. All the things that made the "MOS" process difficult, impossible in certain ways, were solved. It took a while because, well, I found several difficulties, for example silicon would crack, which before was amorphous, instead with polycrystalline silicon it was possible to solve this problem, and things like that, but one by one we solved everything. And so the process, at the end of '68, was five times faster than the "MOS" before. You could integrate twice the number of transistors of the "MOS" before, the leakage current, the "leakage", was from one hundred to a thousand times lower, you could make more than one "gate", one on top of the other.
Simplifying, Faggin solved production problems, replacing metal with polycrystalline silicon in the "gates" of "transistors", and then introduced self-aligning masks. These things made it possible to reduce dimensions and integrate all components, that is "transistors", memory and logic, in a single chip, while before each function required a separate chip. Moreover he improved speed, introducing a small capacitor that increased the voltage of the "transistor", the "Bootstrap Load".
The "Bootstrap Load" was necessary and indispensable to make fast logic circuits. I also fell for it thinking it couldn't be done, and instead I figured out that it could be done. And just a few months before going to "Intel", I managed to solve this problem. I arrived at "Intel" with the ace up my sleeve.
So with the solution in hand Faggin went to "Intel" to give life to the 4004 project, the first microprocessor.
Well, first of all, the idea of making a microprocessor on a chip wasn't such an extraordinary idea, I mean, however, you had to be able to do it. And it wasn't extraordinary because CPUs were already being made with 6, 7, 8 "chips", so that they could at some point be condensed into a chip wasn't the issue. The issue instead of using software instead of hardware to solve a problem was more sophisticated because few knew, understood well what a computer can do. They hadn't understood that a computer is a universal simulator, right?
Exactly. Remember that before this moment every company commissioned calculators programmed according to needs. If you were a bank, you commissioned a calculator to process transactions. If you were "NASA", for space launches. Only with the microprocessor did the idea of having a powerful and standard machine programmable through software spread. So that on a single computer you could have various different software. That's why today, on a single computer, we can have both "Photoshop" and "Minecraft". And we don't have a single computer that does "Minecraft" and another computer that does "Photoshop".
So, if one can simulate a system with a computer, and if they can simulate it in real time, if the computer costs almost nothing, it can be convenient to use the computer instead of making a dedicated system to solve the problem. It's the concept of digital.
Yes, and the company that first commissioned the 4004 from "Intel" was called "Busicom". A Japanese company that made calculators. And had immediately understood the concept of programmable machine.
This concept was inherent in the "Busicom" project. In fact "Busicom" had understood that they could make a series of calculators, "cash registers", so they had designed a computer themselves that they thought could be realized with seven different chips of which three were a CPU. And the others were ROM memory, RAM memory didn't exist yet. And with this structure they made a much bigger investment the first time, but after these same chips could be used to make all the other products. The other products were made much faster, simply by changing the software.
As Faggin says, actually the idea of making chips programmable had been in the air for some time, but the real difficulty was in being able to build them. And even once built, this idea took time to catch on. Also because those who already produced customized circuits for companies, had all the interests to continue doing so.
Well, this is bad weather, because at that time, especially the companies that made "MOS", made custom integrated circuits and so they would have lost their business and said that was the direction. So I still remember that the head of "American Microsystem", the CEO, when he found out that we were making a microprocessor to replace custom devices, he said: "whoever thinks they can make something like that is crazy." I put the "quote" in my office because it was clearly the idea that was somewhat there. Already with the 4004 but it was successful with the "Busicom" business who was initially the exclusive client, then once put on the market towards the end of '71 within nine months to a year there were several clients. The grip was strong immediately then the 8008 came out but it was really the 8080 which being much better as a product also from a usage point of view, the idea that "Intel" had at the beginning wasn't so "bastardized" by the idea that it had to be 16 pins or 18 pins which is crazy.
The 4004 has 16 feet or pins which are the contact points of the chip. The 8008 has 18 while the 8080 has 40. Building chips with more pins was more expensive though and the "Intel" bosses didn't immediately accept Faggin's requests.
This wasn't "off", this was Vadász, Grove, the bosses hadn't understood anything about the microprocessor and so they imposed "constraints" that were absurd. We wasted a lot of time, performance because the same pins were used for "multiplexing" the address, data instructions, so losing time instead of making separate buses.
The "multiplexing" he refers to consists of using the same pins to transport different data to the chip, like instructions and addresses, this allows you to work with fewer pins but makes the chip slower and less efficient. Faggin instead wanted more pins, that is more free roads to be able to transmit data simultaneously and increase the chip's efficiency.
It took me a long time to convince my boss to let me make the 8080 because one of the difficulties that he absolutely didn't want was to use 40 pins, which by the way this was a standard "package" that was commonly used. We lost nine months. I already had the idea of the 8080. We had a year and a half advantage over the competition, that is a product cycle, because when they saw the 8008 it's clear they started working, right? And so if we had immediately made the 8080 as I had proposed, we would have maintained this advantage. Instead we lost nine months due to lack of judgment. I've always fought against these forces, until I said enough, now I'll start my own company and that's it.
Did you by chance also have shares in "Intel" thanks to your experience or were you completely out?
They had given me a "stock option" at that time.
And one of the first to enter "Intel"? Yes, yes, no... Do you know what number you were at "Intel" in terms of hiring?
I think my "badge" was 120, of people, let's say, at a professional level, there were 30, 40. The rest were workers.
I'm curious about these things because you know then anyway despite everything "Intel" continued to prosper for years and so one says maybe even just from that experience a small fortune just thanks to those "Intel" shares you would have made.
Certainly, if I had kept the "Intel" shares I certainly would have done very well even without having "ZiLog" or what I did later.
But you didn't keep them.
Little by little I sold them.
And that's how Faggin leaves "Intel" to make his own chip company. Clearly everything changes because before you remain 100% inventor, you have to add the business component. How was this transition?
Well, it was easy-difficult, that is enormous growth, right? Because making a "business plan", how do you make a "business plan"? Learning also finance, "accounting", how do you make an "income statement", how do you make a company balance sheet? Those things I had to learn. And then taking the right people, how do you hire a good finance chief or administration chief, right? A "marketing" or sales chief. I mean those things there...
I learned them by making mistakes. I made mistakes but not too many. "ZiLog" started in '74, towards the end of '74 essentially. The first product we took six months before finding the money. The money arrived in April.
"Exxon Enterprises" which at that time was the "venture capital" part of "Exxon", the Italian "Esso". On the other hand in '75 there was no money. The stock market had crashed, "venture capitals" had disappeared. In '75 there were only 10 million dollars of "venture capital" in the whole year for all the United States.
By the way it's precisely because of "Exxon" you leave "ZiLog" after a few years.
Yes, because "Exxon" essentially wanted us to become a publicly traded company, they simply wanted to incorporate it. Practically they deceived us. What happened was that "Exxon" forced "ZiLog" to enter inside "Exxon". Faggin instead wanted "ZiLog" to remain autonomous and so he left. Moreover "Exxon" wasn't good at managing "ZiLog", so much so that they later sold it to another company. There was the idea, for example, at "Exxon" that a good "manager" can "manage" anything, right? It's not true. Especially in the technology field, if this "manager" doesn't understand anything about the market, the technology, even if he knows how to "manage", he doesn't understand anything and therefore makes wrong decisions. Anyway "ZiLog" was a very important life school for me, also because in relations with my partner we had difficulties. But at the same time the company was very successful, and even now, it was bought, but still lives as a division.
Yes, it was taken over, but the Z80 became a microprocessor... more popular among the age groups... 8-bit, most important ever made, because especially it taught a generation of engineers how to use microprocessors, it was taught in all schools. I mean, it was really a microprocessor that had thousands of applications all over the world. In fact "ZiLog" stopped last year, said to make the last "buy", because it ceased to manufacture the Z80, but there are four other companies of second source that continue to make it, so still in production 49 years later.
And by the way the business wasn't going badly.
No, just think that in 1976, in March, when we had... The Z80 worked, we were 11, three years later we were 1100. In fact I started to get gray hair. You have to think though that at that time billing 10 million was like 80 million now. I think when I left we were billing 40-50 million, in '80 maybe 60 million.
Anyway, Faggin does his "exit" in '79 at 38 years old, taking home enough to be set for life. "Exxon" buys your shares, basically.
Yes yes, they bought my share and I came home richer, but also a bit disappointed because I expected differently.
