Intel's Crumbling Dominance

Intel was facing the biggest crisis in its history. From what looked like a dominant position of strength, its rivals had surpassed it technologically. Its margins and market share had been eroded to almost nothing. And its leaders were struggling to turn things around. If you think I'm talking about 2022, I'm not.

This symbol outside, it's with intel, the computer inside.

This was the mid-80s and intel was on its knees. Yet by the turn of the century, it had not only survived, it had it became one of the most successful companies in history. Today, a deep dive into one of the most amazing comeback stories of all time, and the man who made it happen. To do justice to such an epic story, you need to go back to the beginning and a guy named Bill Shockley. Shockley is one of the most famous and notorious people in the history of technology. Trained as a physicist, he he worked at Bell Labs and was instrumental in early semiconductor research. During World War II, he conducted secret research into radar techniques to help the Allied war effort. Shockley became so integral that the American government asked him to prepare a report on the number of probable casualties from an Allied invasion of the Japanese mainland. His findings influenced the decision to drop nuclear bombs on Hiroshima and Nagasaki. After the war, Hokoshi Shockley returned to Bell Labs. There he was at the heart of the most important invention in the history of technology, the transistor.

Early electronic research in the field of transistors gave promise of new developments that would revolutionize communications equipment as well as control systems for rockets and guided missiles.

In 1948, Shockley, along with his colleagues John Bardeen and Walter Bratton, applied for a patent for a germanium transfer resistance unit. Shockley continued his research, often in secret. He considered his colleagues his fiercest competitors. In 1950, Shockley released a 558 page book titled Electrons and Holes in Semiconductors. It would become a bible for scientists working to develop new variants of the transistor and other devices based on semiconductors. By this stage, he was generally regarded as the sole inventor of the transistor, which put further strain on his already tense relationship with his colleagues Bardeen and Bratton. Believing he wasn't receiving enough recognition for his work and having fallen out with almost everyone he worked with, Bill Shockley decided he needed a change. So he left Bell Labs. He spent a semester at Caltech and then a year working for the government in Washington D.C. but nothing fully satisfied him. Shockley believed he was more capable than anyone else. So he figured, why not do things his way? Why not start a company? Shockley flew to la where where he met with Arnold Beckman, a chemist and businessman. The pair discussed the idea of a new lab that would build cutting edge semiconductor devices. At a lavish lunch In February of 1956, Shockley and Beckman announced that their semiconductor lab was open for business. Beckman wanted Shockley to establish his lab in Southern California near his own company. Shockley, as ever, trusted his own judgment. He thought Mountain View was a better location. For one thing, as the head of Stanford explained, it would be highly beneficial for Shockley Semiconductor and for Stanford to be close together. The other reason Shockley chose Mountain View was to be close to his mother in Palo Alto. It's kind of crazy to think that one of the reasons that Silicon Valley became this massive tech hub is all because Bill Shockley wanted his office to be close to his mom. Today, the original location of Shockley Semiconductor, 391 San Antonio Road is so famous that Mark Zuckerberg built a meta office there. The address has even lent its name to a documentary about the history of the semiconductor industry. That was February of 1956. In December, Bill Shockley would be awarded the Nobel Prize in physics for his work inventing the transistor. He would share it grudgingly with his former colleagues Walter Bratton and John Bardeen. But at the same time, the Nobel committee was deliberating about who to award the prize to. 6,000 miles away, Soviet tanks were rolling into Hungary's capital, Budapest. They had been called in by Hungary's puppet regime to crush a popular uprising in protest to the Soviet Union's heavy handed rule over Hungary. A delegation of students had broken into the main building of a local radio station where they broadcast their 16 demands for political and economic reform. They were detained by armed guards and when protesters demanded that the students be released, police responded with bullets. This was the spark that lit the Hungarian revolution. Thousands of normal people organized into militias to fight against the authorities. In an effort to quell the violence, the Hungarian government withdrew from the Soviet aligned Warsaw Pact and pledged to reinstate free elections. By the end of October 1956, violence had subsided. The protesters had won, or so they thought. Orders were received from Moscow. The crackdown was to be swift and brutal. Children joined their parents in the streets to fight the Soviets. It didn't make a difference though. By the middle of November, the authorities had killed more than 2,500 protesters and extinguished the flame of revolution. 200,000 Hungarians fled. Most escaped through holes in the Iron Curtain to Austria. One of the Hungarians who successfully made it across the border was a 20 year old chemistry student named Andras Istvan Grohlf. Andras didn't know it at the time, but he would go on to become the most important person in the history of Intel. By 1956, Andras had made it to New York, where he was taken in by extended family. He was safe, but he had already lived through enough tragedy and turmoil to last a hundred lifetimes. Eager to leave his past behind him, Andras decided to change his name so that it would be easier for his professors to pronounce Andras. Isvan Grof became Andrew Stephen Grove. Andy for short. He won a scholarship to study chemical engineering at the City College of New York. Andy's English wasn't great, so he had to repeat each day's work again every night with a dictionary by his side. Despite these challenges, he still graduated at the top of his class. This was such a big achievement that the New York Times even wrote a story about it. After graduating, he moved to the other side of the United States, where He obtained a PhD in chemical engineering from Berkeley. Let's leave Andy for a moment and go 500 miles down the road to Mountain View. Bill Shockley had just founded Shockley Semiconductor. Remember how earlier I said that despite his undeniable brilliance, Bill Shockley was one of the most notorious figures in the history of technology? Yeah, that's underselling it. Shockley's own biographer, someone you'd figure would be sympathetic to him, said that he may have been the worst manager in the history of electronics. Bill Shockley was a tyrant. Despite having some of the smartest people in the country at his disposal, his egomania meant he didn't trust them with anything. He was impossible to please and increasingly paranoid. Once, after a secretary suffered a small cut, Shockley demanded all his staff take lie detector tests so he could find the culprit. It didn't take long for people to begin looking for the exit door. In late 1957, less than a year after Shockley received his Nobel Prize, a group of eight of the lab's best researchers decided that they'd had enough. They walked out of the lab and straight into history. One of the young men, Eugene Kleiner, wrote to a connection at a New York investment firm. He wrote, seeking assistance. Kleiner's contact wasn't really interested. He was about to retire and he didn't want to manage an entirely new portfolio. So he passed the letter to a young associate named Arthur Rock. This is an amazing Sliding Doors moment in history. Arthur Rock knew about semiconductors and understood their potential. He also knew all about Bill Shockley, both the good and the bad. And he understood what a big deal it was that these eight guys were willing to quit working with a man that many people in the industry referred to simply as God. So Arthur Rock wrote to Eugene Kleiner and over dinner in San Francisco, suggested an idea that even these geniuses never contemplated. He told them that they should start their own company. It was the best way for them to gain the autonomy they craved, while also capturing a fair share of the value that they created. The young researchers realized that Arthur Rock was right. They agreed to his suggestion. So he got to work. But he hit a brick wall. Despite being brilliant, the eight researchers had no profile of their own. And and they were working with technology that may as well have been magic to the average person. Arthur Rock approached 35 different investors and couldn't squeeze a single cent out of them. On the advice of a colleague, he approached Sherman Fairchild. Fairchild was an inventor and keen student of science. He was also very rich. Sherman's father was the co founder and first chairman of IBM. And when he passed away in 1924, Sherman was the sole inheritor of his multimillion dollar estate, as well as millions of dollars worth of IBM stock. Fairchild met with the eight researchers. One of them presented his vision for the future of semiconductors. The substrate should be made from silicon, not germanium. He argued. Silicon had superior physical properties, plus its materials were cheaper to obtain. The visionary researcher also expressed his belief that appliances made from silicon components would simply be replaced rather than repaired, which meant more money for someone like Sherman Fairchild. The businessman was impressed enough to invest. He agreed to loan the researchers $1.5 million to start their company. In return, he'd gained the right to buy the whole company for 3 million, an option he ended up exercising. The eight young men, the so called Traitorous Eight, founded Fairchild Semiconductor in 1957. Two of the eight would go on to become main characters in the intel story. One of them, Robert Noyce, made the presentation that persuaded Sherman Fairchild to invest. Noyce invented the modern integrated circuit, which essentially was the blueprint for implementing transistors in a silicon substrate. Gordon Moore, meanwhile, was such a brilliant scientist that he coined a term still in common use in modern computing. Moore's Law. He was also the man who hired Andy Grove to come work at Fairchild. It was 1963. Moore and his colleagues needed smart employees who could lead the lab's research into embedding transistors on silicon wafers. Andy had just finished his PhD. It was the beginning of a 30 year working relationship that would transform the futures of all three men. The future of Silicon Valley and the future of computing. Andy quickly made a name for himself as a high achiever with high standards. The group he was part of was trying to tackle a persistent problem that, when it was solved, would pave the way for metal oxide semiconductor technology, which is still the most common type of transistor in integrated circuits more than 50 years later. His team also succeeded in reducing the instability of transistors by removing impurities in silicon chips. Bob Noyce and Gordon Moore were watching closely. Fairchild Semiconductor enjoyed a great first decade. They were technologically innovative, but they were also innovative on the business side of things. In 1965, they shocked the industry by announcing they would price their integrated circuit products at just $1, far below cost. But Bob Noyce drove that decision because he understood that the market for integrated circuits was about to explode. Setting the price so low allowed FairChild to achieve two important things. Firstly, it would accelerate demand. Secondly, below cost pricing would make Fairchild the logical choice for companies just entering the market. The insight reminds me of one of the big realizations that Morris Chang, the founder of tsmc, made early on in his own career. When Morris was at Texas Instruments, he implemented learning curve pricing, a model that improved yields by rapidly scaling production at low price. He understood that when manufacturing volumes are small, data and market share are more important than profits, because as the market matures, the profits will come. You've just got to give yourself the biggest chance of being the guy standing underneath the waterfall when they do. This is the kind of freedom that senior staff at Fairchild had. They pushed boundaries and sold technology to clients as big as the American military. And as far as Andy and the Traitorous Eight were concerned, there was just one the structure of their deal with Sherman Fairchild meant they didn't have any equity. Gordon Moore and Bob Noyce believed their talents were being undervalued. They also believed that they could take research into computer memory in exciting new directions. A highly simplified model of a computer today includes three main the cpu, which acts as the main processor. There's RAM for short term memory and hard drives, or SSDs to store long term memory. At the time the late 60s, long term memory lived on magnetic disks, not silicon semiconductors. Short term memory didn't exist at all, neither did the cpu. There was basically no demarcation between software and hardware. Moore and Noyce didn't necessarily see every piece of the puzzle, but they believed they could marry silicon semiconductor technology with computer memory. And they were determined to figure it out. So they decided to speak to Arthur Rock. Rock had learned from the experience of founding Fairchild Semiconductor. He could see that while it was a great model for making new discoveries, it only made one person seriously rich. Sherman Fairchild. Rock knew that there had to be a better way to reward successful founders. So he suggested to Moore and Noyce that they start a company and he would lead the fundraising. They agreed, and Rock ended up raising $2.5 million. His reward, in addition to an equity stake, was being appointed the inaugural chairman. On July 18, 1968, the new company was founded in Mountain View. It needed a name. Moore and Noyce initially thought they would go by the convention of the time. That would have meant calling the company Moore Noyce, but yeah, that's a terrible name. It sounds like Moore Noyce. It's not ideal. So they went back to the drawing board. They picked a new integrated electronics. Chances are, you know, the shortened version of this, Intel. Now that they had their name decided, Gordon Moore and Bob Noyce turned their attention to recruitment. They both knew there was only one choice for Intel's first ever employee. The young man who had survived a fascist dictatorship, German military occupation, the siege of Budapest, communism and a brutal crackdown before finding his new home in the United States, Andy Grove. You might think that given the amount of adversity he had already survived Andy and the huge promise he had already shown in the early part of his career, that Andy would be confident about his ability to take the next step. But that's not how he saw it. Andy was, to use his own words, scared to death. He had left a secure job in an innovative company to run R and D for a new startup in an industry that essentially didn't exist yet. But he had complete faith in what the founders were trying to do. Bob Noyce and Gordon Moore were in a hurry. The conglomerate Honeywell was Ready to buy 64 bit static random access memory chips from any company that successfully developed them. That triggered a rush among memory manufacturers. Intel was at the back of the pack. It had only been around for a year, it hadn't released a single product, and it was also establishing semiconductor manufacturing capacity essentially from scratch. But Intel's founders, Bob Noyce and Gordon Moore, believed that they could over overtake their competitors from behind by developing faster and more powerful memory products. So they directed Andy to pursue three different technologies at once. The first was bipolar memory. This approach used relatively proven technology, the bipolar transistor, but it was hard to develop further. The second was Silicon Gate metal Oxide semiconductor memory. And the third was multi chip memory, which involved linking together smaller memory chips. Whichever technology proved its viability the fastest would become Intel's first product. The winner was bipolar memory. As a result, Intel's first ever product was the 3101. The 3101 was a static random access memory chip or SRAM. SRAMs were hard to develop and intel was working on extremely tight deadlines. But the 3101 was still twice as fast as the competition. Its development sent a clear signal to the rest of the industry. Intel was a force to be reckoned with. To underline this, it quickly followed up the 3101 with the 1101. Developed in parallel with the 3101, the 1101 was also an SRAM chip. But it was also the first widely produced chip made using silicon gate metal oxide semiconductor technology. MOS technology was a much more efficient way of manufacturing chips than the bipolar memory architecture it replaced. It enabled the large scale integration that would be required for the PC revolution. If the 1101 was a big step forward, then Intel's next product was a huge leap. Released In October of 1970, the 1103 was the first dynamic random access memory chip. It made magnetic core memory completely obsolete. The device you're watching this video on is almost certainly packed with DRAM chips. They're the memory of choice when you want something with low cost and high capacity. The 1103 made Intel a world leader in memory manufacturing for a decade. Bob Noyce and Gordon Moore capitalized on this incredible early success by taking intel public at a market cap of $58 million in October of 1971. Their company had gone from nothing to being the market leader in just two and a half years. And they weren't done yet. Just weeks later, intel unveiled the world's first widely available microprocessor, the 4004. The 4004 is a great example of how the products that make the biggest impact require a combination of brilliance and circumstance. When a Japanese calculator manufacturer asked intel to design chips for a series of calculators, intel assigned its engineer Ted Hoff to the project. During his search for a cost effective design, Hoff devised the plan of fitting a CPU onto a single chip. By successfully harnessing MOS technology, Intel's engineers fit twice as many transistors onto the 4004 as existing process products, which resulted in five times the operating speed. The 4004 was as powerful as the ENIAC, the first ever programmable digital computer. But it was a bit smaller. ENIAC weighed 27 tons and was as big as a house. The 4004, by contrast, measured 18 by 16 of an inch. By miniaturizing the CPU, the 4004 made it possible for small machines to perform the kinds of calculations that in the past only made mainframe computers could do. In other words, it made the personal computer possible. Intel released four transformational products within the first four years of its history and it did all of this with fewer than 200 staff. Intel followed up the 4004 with the 8008, the first eight bit microprocessor. The seeds of the PC revolution had been sown. They would soon bear fruit. The March 1974 issue of a magazine featured the first ever ad for a personal computer. The Selby. Short for scientific, electronic and biological. The Selby cost $565 and it came with 1k of programmable memory. With an additional 15k of memory available for $2,760, it ran on an 8008 microprocessor made by Intel. Next in July came the Mark VIII, which also ran on an 8080. The 8080 was a critical component of all the first performance personal computers. But Intel's founders and engineers were already looking beyond it. Between the release of the Selby and the Mark VIII, it released the 8080. The 8080 was the first truly general purpose microprocessor. It was a computer on a chip, and at a time when conventional computers cost thousands of dollars, intel sold it for just $360 at release. The 8080 soon became the industry standard, cementing Intel's status as as undisputed market leader. The next noteworthy personal computer to be released was the Altair 8800. Designed and manufactured by MITS, it made it to the market just before Christmas of 1975. It was billed as the world's first minicomputer kit to rival commercial models. The Altair shipped with an Intel 8080 CPU and a 256 byte RAM card, all for $400. You had to assemble it yourself, but it was cheap and accessible and you could write and execute your own software. Two Harvard freshmen realized that the Altair created new possibilities. They contacted Ed Roberts, founder of mits, the company that made the Altair. Within six weeks, the two intrepid young programmers had compiled a version of Basic to run on the Altair. Their names were Paul Allen and Bill Gates. And the company they would soon establish. They named it Microsoft, which of course would go on to Become an important part of the intel story. That same year, a young Steve Wozniak, partly inspired by the Altair, built a microprocessor into his video terminal in order to develop a complete computer. That project became the Apple one, and Steve Wozniak's close friend Steve Jobs convinced him that the two should go into business together. Intel was growing as quickly as the number of transistors it could fit onto its memory chips. In 1978, it introduced the 8086, signaling the beginning of the famous x86 architecture. In the early days, Bob Noyce, Gordon Moore and Andy Grove opted for informal weekly luncheons with their direct reports. That system worked great when intel had 200 employees. But by 1980, intel had 15,000 and a market cap of $2 billion. Such rapid growth required a more scalable management approach. So Intel's founders implemented a program that emphasized openness while also delegating decision making to the lowest levels. They trusted staff to solve problems and to prove they were all in it together. They avoided luxuries like limos and long lunches. Intel rose to the top of the industry and stayed there because it maintained focus on the micro and the macro. At the same time, it was helped by Andy Grove's paranoia. Now, Andy wasn't paranoid in the same way as Bill Shockley. He was paranoid in a way that kept him vigilant. Andy would go on to write a book about his time at Intel. In it, he wrote that success breeds complacency. Complacency breeds failure. Only the paranoid survive. The title of the book was borrowed from this final sentence. As a result of his productive paranoia, Andy urged his senior executives to allow people to test new ideas, products and sales channels. When the world changed, he was determined to be ready. On August 12, 1981, IBM unveiled the model 5150. This product set the standard for what we think of now when we think about PCs. Helped by a famous marketing campaign, the IBM PC flew off the shelves. The company originally forecast that it would sell 250,000 units within five years. In reality, it sometimes shipped that many in a single month. It became the first computer widely adopted by regular businesses. And every single one had an intel microprocessor inside. By bringing the personal computer into hundreds of thousands of homes and offices around America, IBM created a vast era ecosystem of software and other peripherals. In December of 1982, flush with the money that the PC was generating, IBM paid $250 million for a 12% share of Intel. The next year, Intel's revenues exceeded $1 billion for the first time. The leadership team celebrated the first $250 million revenue quarter with a rare indulgence. A custom bottled champagne they called vintage Intel. That's not to say there weren't bumps along the road. Intel's first attempt at a 32 bit processor system, the IAPX432, was a radical departure in terms of build and purpose than the 8008 and the 8080 processor. Instead, it was a disaster, running five to ten times more slowly than competitors. But Gordon, Andy and Bob knew that occasional failures were just the cost of creating a culture that reliably produced transformational products. It seemed like nothing could stop Intel. But the story was never going to be that simple. Because in business, your margin is always someone else's opportunity. If you want to build an enduring business, you need a moat that will allow you to maintain a competitive advantage against rivals. And there are a few ways to build a moat. You can exploit some regulation in a way that no one else has. You can be handed a monopoly by the government, so you're the only game in town. Or you can make a product that no one else can. The problem with that approach is that unless you can keep finding ways to protect your moat, someone will eventually get through. And that's what happened to Intel. It pioneered memory chips and scaled their production. But in the process, it also taught other companies how to make them. Every computer needed memory, but memory is just memory. So long as it was delivered on time and on budget, the companies who needed it didn't care much where it came from. Intel's rivals were learning how to make DRAM chips. They were after a big piece of the market that intel had created, and they could see vulnerabilities. The main one was manufacturing. Intel had been plagued by intermittent undercapacity issues for years. Intel's seventh fab took two years to get up to full capacity. Things weren't helped by an economic slowdown, which further undermined its manufacturing capabilities. As a result, the companies that bought Intel's products started asking it to outsource some of its production to other suppliers, a practice known as secondary sourcing. Intel partnered with up and coming Japanese firms. That took some of the pressure off, but it also helped Japanese firms like Toshiba, nec, Hitachi, Mitsubishi Electronic and Fujitsu climb up the learning curve. Learning how intel designed their memory chips was the first big advantage for the Japanese companies. Their second big advantage was their superior access to capital. Every Japanese electronics company that threatened Intel's market Position was part of a conglomerate with a presence in several industries and a close relationship to the Japanese banking and government sectors. They could borrow cheaply to expand their production capacity and had access to high quality engineering talent. And they could absorb financial losses for a very long time. As a result, the top Japanese producers were able to reach production yields that were up to 40% higher than leading U.S. companies. At its peak, DRAM memory accounted for more than 90% of Intel's revenues. It subsidized other parts the of of the business while also serving as a technology driver that underpinned Intel's learning curve. But that was before Intel's moat became a commodity and its secondary suppliers became rivals. In 1974, intel controlled 83% of the market for DRAMS. Ten years later, it controlled just 1.3%. This was a shocking decrease. Opinions about how to respond were divided. The engineers thought the sales team needed to work harder. The sales team thought the engineers needed to come up with better products. Everyone was desperately looking to Intel's leaders for an answer. From having been on top of the world, Intel's founders were finally faced with the reality that the business model that had carried them to success making memory chips, was no longer an option. D day arrived in early 1985. Andy Grove was in his office with Gordon Moore, pondering. These men were already legends. Together with Bob Noyce, they'd built their company into the undisputed market leader in computer memory. They could have walked off into the sunset content in the knowledge that their legacies and personal fortunes were forever secure. So that's what they did. Gordon Moore and Andy Grove shook hands and walked in the direction of the sunset. Except they didn't get very far, because after just a few seconds, they turned around and walked back in. And they rehired themselves. This whole exchange came about because of a question Andy asked Gordon. If the board kicked us out and brought in a new CEO, what do you think the new CEO would do in response to this situation? Moore looked him straight in the eyes and answered, his voice steady. They'd get us out of the memory business. So that's exactly what they vowed to do. Gordon Moore and Andy Grove symbolically fired themselves from their positions as leaders of an industry incumbent which had lost the agility to respond to a rapidly changing world. And then they rehired themselves as leaders of a company that would become the world leader in microprocessors. With new newfound tenacity, microprocessors were by no means an obviously attractive revenue source. Personal computers were coming into focus as a business opportunity. But the outline was still fuzzy. They didn't have the luxury of working from a blueprint. And they didn't have the luxury of colleagues who saw things the same way they did. Internal resistance was stiff. Intel was defined by memory. It was such a core product. Intel's employees, including many of its top executives, were emotionally attached to the DRAM business and the company was highly dependent on it. The sales team couldn't let go of the belief that they just needed a full product line to do a good job in front of customers. But Andy Grove understood the reality. Intel had to lose its memory or it would lose everything. The history of business is littered with companies that were on top for a long time, but couldn't see how fast the world was changing in ways that threatened their position. Names like Nokia, Kodak and BlackBerry didn't mean anything to Andy, but we understand them as cautionary tales of companies that weren't paranoid enough to survive. He summarized this challenge better than anyone. One of the toughest challenges is to make people see that these self evident truths are no longer true. There weren't and still aren't many examples of a company the size of intel successfully pivoting away from the business model that made it successful in the first place. But that's what Andy was determined to do. He directed the head of Intel's memory division to develop a plan to get out of the memory business. Unsurprisingly, this guy refused the order. So Andy fired him and ordered his replacement to do the same thing. Get out of the memory business. The replacement also refused. So Andy had no choice to fire him too. I mean, this is kind of understandable. No one wants to make themselves obsolete, especially when it means firing the people you've worked with for years. But it was still brutal. It was brutal, but it had to be done. Andy knew he was trying to turn around a huge ship that was rapidly taking on water. Intel lost $173 million in 1986. Until 2022. It was the only year that intel had ever lost money. It was closing factories, laying off thousands of staff. It was headed for bankruptcy. It was in this climate of doom and gloom where intel wasn't just worried about its future in the market for memory chips, but its continued existence as a company that Andy took over the reins as CEO from Gordon Moore. But there was no time to celebrate because Andy was face to face with the biggest challenge of his career. He had a singular focus to do the obvious but hard thing. It was a huge task. But there was one thing even more dangerous than trying to execute such an audacious standing. Still, Intel's founders and engineers had done more than any other company to create the technology underpinning the personal computer. The speed at which intel had lost market share in memory took Andy by surprise. But he had an ace up his sleeve. Since 1982, the company had been quietly working on a new microprocessor called the 3386. It wasn't just the next product in the x86 architecture. It was a paradigm shift. The 386 was the first 32 bit microprocessor. Technologically, it was like when tennis players switched from using wooden rackets to ones made of Graphite. It had 4 kilobytes of memory and could perform 4 million operations per second. It could run more than one program at a time. And critically, it made graphical operating environments for PCs practical. Andy Grove knew the 386 had potential to be Intel's life raft. So he wanted to protect it at all costs. After years of resorting to secondary sourcing in order to fulfill clients demands for memory chips, a decision which helped Intel's rivals overtake it and ultimately turned Intel's main product line into a commodity that almost anyone could manufacture, Andy was determined to make intel the sole gatekeeper. He faced up to his responsibility as a leader and made the call. Intel would not license the design of the 386 to any other manufacturers. No more secondary sourcing. The best decisions are the ones that seem totally obvious in hindsight. But Andy was taking a big risk. He was defying IBM, which was Intel's biggest customer and a major shareholder. He was also defying standard industry practice. Secondary sourcing was common because manufacturers consistently had to solve difficult and novel problems. Technical issues resulted in longer wait times for clients, so in order to guard against the risk of intermittent supply, customers insisted that their chips be made by multiple manufacturers. It wasn't an easy decision, nor was it a cheap one. Andy and his team had to spend billions on expensive new fabs to ensure consistent supply. Amd, one of Intel's secondary suppliers for microprocessors, successfully sued for breach of contract and ended up winning more than $10 million in damages. But Andy would be vindicated. His decision was key to successfully turning around Intel. And his decision dated back to a project that began when intel was on the edge of disaster and to Andy's paranoia. In March of 1985, a new computer manufacturer named Compaq started working on a new desktop machine. Compaq was fighting a difficult battle. The main player in the market was IBM Big Blue. But Compaq wasn't working alone. It worked closely with intel, which shared information about its 386 processor architecture. And they were also working closely with Microsoft. So Intel, Compaq and Microsoft signed a non disclosure agreement and got to work. Early prototypes of the computer that they were working on were designed around the earliest batches of the 386, which had a 12 MHz clock speed. Crucially, by the time the computer was ready for release, manufacturing yields of the improved 16 MHz 386 had reached acceptable numbers, resulting in a 25% speed boost. On September 9, 1986, Compaq unveiled the DeskPro 386 at a launch event in New York City. The DeskPro 386 was the most advanced high performance personal computer in the world, and it was a hit. By the following February, Compaq had sold 25,000 units. During the second quarter of the year, another 90,000 were sold. Compaq reported selling out of the DeskPro 386 by the end of 1987. Not bad for a computer that cost $6,500. The DeskPro had successfully challenged the Big Blue. Compaq's president was quoted at the time saying that IBM had six months to respond or they risked losing some of their market share that had been unchallenged since the introduction of the Model 5150. IBM's response, the PS2, didn't come on the market until almost a whole year later. And when it did, it was a total disaster. IBM planned to replace Microsoft's popular DOS operating system with its own OS 2. But this proved to be a big mistake. However, the failure of the PS2, which lost millions of dollars and sent the entire company into a tailspin, effectively left Microsoft without any viable competitors. Bill Gates had this to say when asked about the success of the DeskPro in an interview for PC Magazine in 1997. The folks at IBM didn't trust the 386. They didn't think it would get done. So we encouraged Compaq to go ahead and just do a 386 machine. That was the first time people started to get a sense that it wasn't just IBM setting the standards, that this industry had a life of its own, and that companies like Compaq and Intel were in there doing new things that people should pay attention to. IBM still stung by Andy Grove's decision to not second source production of the 386 fell victim to the same kind of thinking he spent so long trying to overcome. It had a stubborn, irrational belief that intel couldn't get it done. The DeskPro established Intel's 386 processor as the dominant CPU choice. It's possible to tell the story from this point on like it's just one big victory lap for Intel. But there were still significant challenges that required all all of Andy's technical, business and managerial expertise to solve. The first one arrived almost right away when intel began work on a chip that would succeed the 386. At the time, there was a strong push within the semiconductor industry toward a protocol called Reduced Instruction Set Computing or risk. This design eliminated rarely used instructions in order to free up space and boost speed. Although it's retroactively been called Complex Instruction set, that architecture, cisc, the incumbent protocol, didn't really have a name. It was just how things were done. In this design, a single instruction performs all loading, evaluating and storing operations. Hence it's complex. Many smart people within intel were convinced that risk was the future. One of Intel's rising stars, a 25 year old engineer named Pat Gelsinger, actually quit because he thought the company was going to move away from the CISC architecture. He believed that what risk chips gained in speed, they lost in flexibility. More to the point, they weren't compatible with the software already on the market. The same software that intel believed would guarantee the success of the follow up to the 386. Andy Grove shared the young engineer's doubts. He also wanted to keep him on staff. So he put Gelsinger in charge of the follow UP project, the 486. The 486, which was designed using the CISC architecture, was released in 1989. It doubled the performance of the 386 without increasing the clock rate, cementing Intel's technological advantage and leading market position. And today Pat Gelsinger serves as CEO of Intel. Of course, Andy wasn't the kind of guy to put all his eggs into one basket. Intel also released a risc chip, the 860. But it was the 486 that would emerge as the preferred option for buyers. It took until March, now 1991, more than five years after the release of the Intel 386 for AMD to release its own AM386 processor. Partly that was because of Intel's success tying it up in the courts. The relationship between the two semiconductor companies had become acrimonious. AMD's processor was arguably the Superior product. But intel won the race to be first to market with a product that met the surging demand for personal computer processors. The success of the 386 and the 486 allowed Andy and his team to plow massive sums of money into creating new and faster designs, testing the limits of the law coined by its founder, Gordon Moore. It also gave them the money to launch one of the most brilliant marketing campaigns of all time. It's the sound logo intel created as part of its intel inside campaign. It was the perfect campaign. Launched at the perfect time, intel could see that PCs were about to become an essential device. So it sought to bolster consumer interest and demand for computers that featured intel inside. This is kind of crazy. Most people knew nothing about computers, let alone the components inside them. But they identified their computer by the type of chip it had inside, not just by who made the box. Intel inside wasn't just aimed at consumers, though. Intel understood that its positive brand recognition gave it huge leverage over PC manufacturers like Compaq, Dell, and even even IBM. So it would offer to pay for half of the manufacturer's marketing costs for their new computer if they prominently advertised the intel processor inside the machine. The results were astronomical. Intel's annual net income topped $1 billion for the first time in 1992, and they made almost 6 billion in revenue that year. By this stage, intel had created a rock solid partnership with Microsoft. Every PC with an intel processor shipped with Windows as standard. Many people believed this was anti competitive and that it undermined innovation. Whether or not it was good for consumers, it was certainly good for Intel. The partnership between the two companies became known as Wintel in 1993, helped by the release of their fifth generation Pentium processor, Intel's revenues increased by more than 50% to over $8 billion. Using the massive proceeds of its X86 series sales. Andy and his team continued developing increasing increasingly powerful processors. And all the while, the money pile just grew bigger. Revenues broke 20 billion in 1996, and over 5 billion of that was profit. The 90s was the decade of intel and Microsoft. Both effectively became monopolists for millions of people around the world. A good computer had Microsoft Windows and an intel processor. Both companies knew this, and they knew they could charge accordingly. In May of 1998, 11 years after being made CEO following Gordon Moore's retirement, Andy Grove stepped down. He drove Intel's transformation from a maker of memory chips into the world's dominant producer of PC microprocessors, servers, and general purpose computing. During his tenure Intel's market cap grew from 4 billion to 197 billion. Revenues grew more than 10x. When Andy stepped down, intel was the world's seventh biggest company and employed 64,000 people people. Time magazine named him their man of the Year. He continued serving as Intel's chairman until 2004, when he finally retired. Andras Istvan Grof passed away in March of 2016 at the age of 79. By that time, intel had lost its status as the world's preeminent semiconductor manufacturer in the face of competition unlike anything it had seen for decades. Can intel ever win back its crown? That's a story for another day, but the story of Andy Grove, a Jewish kid born in Budapest in 1936, who lived through a fascist dictatorship, Nazi occupation, and the Final Solution, escaped repressive communism, and built one of the world's greatest companies. That story deserves to be told for centuries to come. Thanks for watching.