Intel’s Reign of Terror
In 1992, Intel made about $5.8 billion in revenue and $1 billion in net income. By 1998, the company did $28 billion in revenue and $6.95 billion in net income. That year, Intel was the third most profitable company on the Fortune 500. Only Exxon Mobil and General Electric did better.
The 1990s made Intel the world's richest and most powerful semiconductor company. And that was the problem. In this video, Intel at the peak. The most powerful semiconductor company of all time. Where do you go from there?
## Beginnings What drove this massive growth boom? Several things. On the demand side. In mid-1992, the biggest PC clone maker Compaq changes its leadership and corporate strategy - drastically cutting prices on their PCs. The consumer PC clone boom that followed birthed a new generation of commodity PC box sellers like Dell and Gateway 2000. They all need an Intel CPU, often paying 20-25% of their total wholesale price for it.
And people paid those prices because they felt like they were getting the best technology. Intel told them so. In mid-1991, there was the famous "Intel Inside" marketing campaign. Launched in response to AMD's second sourcing of the 386 CPU and backed with a staggering $500 million throughout the 1990s, Intel Inside told customers that they should care about the silicon powering their computers. The campaign also had a partnership-marketing fund, in which Intel would share the cost of their marketing expense if an "Intel Inside" logo was included in it. The $125 million fund worked wonders. 1,200 companies signed up, basically the whole PC industry. In 1993, Financial World magazine named Intel the third most valuable brand - after Marlboro (yikes) and Coca-Cola.
## Pentium Then in early 1993, Intel launches the 3.1 million transistor Pentium CPU. And Microsoft, the other half of the Wintel death machine, releases a series of popular, widely adopted operating systems. First in 1993, Windows NT makes inroads into the corporate world. And then Windows 95 completes a long-running merger between the text-oriented MS-DOS and a long-waited Graphical User Interface. The compute-intensive GUI drove people to buy more powerful chips. But more powerful chips gave programmers leeway to do more intensive things. That in turn causes the software to slow down. Which then motivates people to go buy a more powerful chip.
Thus the famous phrase: Grove giveth, and Gates taketh away. ## Taming the Competition At this time, the competition is temporarily tamed. PowerPC - a consortium of Apple, IBM, and Motorola - crumbles thanks to bugs in its initial PowerPC 600 CPU family. Plus Apple's unwillingness to license its operating system and go the clone route put a ceiling on its adoption. AMD is temporarily put to bed. In 1995, Intel makes peace after a flurry of lawsuits. Now
barred from second-sourcing Intel, AMD must make their own x86 products. Their first in-house chip, the K5, suffers delays and then flops due to subpar performance. Meanwhile Intel starts really flexing its capital expenditure muscle. In 1992, they spent $2 billion for new fabs in California and Ireland.
These expenditures bear fruit as the company ramped Pentium. At the same time, Intel pulled off a savvy vertical integration move into chipsets and PC motherboards. The motherboard is the main PCB for a PC, and the chipset maintains communications between the PC and external components.
This move vastly accelerated Pentium CPU adoption while also sucking profits out of a competing part of the value chain. It also massively consolidated Taiwan's big motherboard industry. In 1993, there were 300 motherboard-makers in Taiwan. By 1996, just 20 such companies would be left. Intel's stock skyrockets, with the price growing a thousand-fold since the IPO in 1971. The company now holds a 80% market share on the PC CPU market. ## Leadership Transition At the top, three leaders shared power at the company.
Co-founder Gordon Moore still popped in about three times a week to discuss strategy and long term considerations. But he had always been a bit of an academic - removed from everyday operations. Then-CEO Andy Grove ran the company's marketing strategy and product roadmaps. Every decision on the microprocessors went through him. And then the third leader, there was COO Craig Barrett. A burly former Stanford professor who
joined in 1974. He led the company's formidable operations and engineering side. Whatever Moore and Grove decided to do, it was Barrett's job to make happen and scale it across the globe. Grove finally starts to step back in the late 1990s. If being the CEO of one of the world's most valuable companies was not enough, he was also becoming a well known business guru. In 1996, he published a bestselling book called "Only the Paranoid Survive". It covers a series of decision points in Intel's history including the one to get out of DRAMs. The title is
quite famous today, perhaps as famous as that Clayton Christensen book "Innovator's Dilemma". The next year, 1997, Time named Grove "Man of the Year" - directly tying him to America's ongoing digital and information technology transformation. Peers like Regis McKenna, the Silicon Valley marketing legend, said that he may be the "best manager in the world".
Being the face of all that was tiring. Grove was now entering his sixties and was exhausted. One time, he pulled over while driving in San Francisco and asking his wife to take over, saying that he was "just so tired". In 1997, Intel had a market cap of over $110 billion. Thanks to generous stock option compensation, Andy Grove cashed in $95 million the year prior. With Gordon Moore also pulling back, it was time for a transition. Grove asked Barrett to
take over as CEO and stepped up into the boardroom as chairman. To outsiders, Barrett’s demeanor seemed softer than Grove’s. But to steal a quote from Mr. Lee Kuan Yew, whoever governs Intel must have that iron in him and Barrett had iron indeed. Grove said: > I have heard him describe himself as more ambitious and aggressive than I am. His ambition is to start from our position today and move it to a higher category ## Intel and the PC-makers What higher category would that be however? That was the big question.
Throughout the 1990s, there was constant angst between Intel and the big PC makers, with Compaq and its CEO Eckhard Pfeiffer leading the chorus. The two companies maintained a close but tense relationship. Compaq like others constantly hissed at Intel taking a percentage cut of their PCs’ entire wholesale price - a cut sometimes as high as 20-25%. It left Intel with a delicious 55% gross margin in 1994, while Compaq's gross margin languished at just 26.6%. The margins of Dell, Gateway 2000, and AST Research were even lower. Such lopsided margins hints at who wore the pants in the relationship.
Pfeiffer especially disliked Intel's willingness to use its new motherboard business to help rivals like Gateway 2000 and Dell bring Pentium-powered computers to market ahead of Compaq. Fair enough. But didn’t Compaq remember a few years earlier how they and Intel collaborated on the Compaq Deskpro 386 PC to boop IBM? I don’t think he should have been surprised. Ride a scorpion across a river, it’ll sting you.
Anyway. Moves like that made the PC makers paranoid that their vendor and close technical partner - a partner without which they cannot offer a product - will someday compete against them. An Intel-made PC could have deeply undercut their sales. For this reason Compaq and others looked for alternatives. Noticeably in September 1994, they unveiled a line of Presario desktop computers with an AMD 486 chip. The 486 was already an antediluvian six years old in 1994, but it wasn’t Intel. And the fact that the PC vendors would even try such a thing indicates the desperation of their predicament.
In an interview with BusinessWeek, the chairman of a large computer-maker said in 1995: "Everybody has been hassling Gates about Microsoft's monopoly. Microsoft doesn't have half the leverage that Intel has today." Eventually even Pfeiffer bent the knee, because if he did not then his company would not get that year's latest chip - the Pentium Pro. But everyone can see the point,
there was little margin left for Intel to take. ## Intel and Microsoft And then on the software front, Intel bumped up against Microsoft. In the 1990s, Intel was more profitable than its Wintel partner, which made a pitiful $4.49 billion in net income in the fiscal year ending June 1998. But Gates and Microsoft refused to act the junior partner and cede any airspace to Intel. In 1995,
Intel starts working on a multimedia software standard called "native signal processing" or NSP that lets the CPU handle some audio and video compute processing. Seems hardware enough to me. But this software work essentially co-opts Windows. When Microsoft caught wind of that, they considered it a threatening flanking maneuver and immediately announced that they would not support it. CEO Bill Gates sends an email to Andy Grove to express his irritation. Thus Intel backed off, but always kept an eye
out for Microsoft developing a version of Windows that ran outside of Intel's chips. The mobile OS Windows CE, first released in 1996, was a good example. It could run on x86 chips, but ARM-based or MIPS-based ones too. That never really caught on, but nevertheless Intel felt like efforts like these were Microsoft’s warning shots to Intel about entering the software market.
## Sub-$1000 PCs Craig Barrett's real concern at the start of his tenure in 1998, however, was the unexpected popularity of cheaper PCs. By 1996, PC home adoption hit a bit of a plateau, at around 40-50%. As I mentioned in my video about Compaq, all the low hanging fruit had been eaten. By then, CPUs got powerful enough to handle basic tasks like word processing and spreadsheets. More compute-intensive consumer applications like gaming were coming along but had not yet become a big enough market.
And with the rise of the Internet, consumers found a source of content that did not need the latest CPU or Windows operating system for more entertainment. And sadly enough for Intel, browser bloat - and Electron-based desktop apps shipping entire browsers inside their packages - would not be a thing for another decade and a half. A lesser-known corollary of Moore's Law goes: If customers don't demand better performance, they will demand lower prices. In late 1998, the average price of a PC fell below $1000 for the first time. A third of the PCs sold were for less than $800 - driven by low cost providers like Packard-Bell. Intel has seen this before. Customers originally complained that the 386 was unnecessarily more
powerful compared to the 286. That eventually led them to the famous Intel Inside campaign. But this time the stakes were much higher. Since Intel priced its CPUs as a percentage, lower PC prices threatened revenues. Such a market situation also favored Intel's rivals like National Semiconductor and AMD which did not invest so much into leading-edge fabs. The latter managed to sneak back into the
market after their iffy K5 effort with the much better received K6-2 chips. And most of all, lower PC prices threatened Intel's way of doing business. Every few years, Intel spends billions of dollars on another massive factory - praying that if they build them, then the people will come. If sales dry up, then the money for that next factory dries up too. ## Otellini and Segmenting the Line So Craig Barrett assigned one of his rising stars to fix this problem: Paul Otellini, a soft-spoken Bay Area native who first joined Intel through the finance department in 1974 and then rose up the ranks to become a general manager. Andy Grove made Otellini his "technical assistant" in 1989. Otellini was smart
and aggressive but lacked a big ego, and Grove liked that. The role would have been somewhat of a demotion for Otellini but he nevertheless took it. Grove then assigned him onto various thorny business problems across Intel. Otellini was a talented diplomat - his brother apparently once worked in the Vatican as one - and this helped him navigate the company's complicated politics and many power centers. Otellini’s solution to the PC pricing problem was to segment the Pentium line. For sub-$1000 PCs, they brought out a cheaper brand called the Celeron.
After a rough start, Intel refined these low-end Celeron chips - some units were priced as low as just $71 - and they started to eat into AMD's and Cyrix's businesses. Intel did the same for the high-end too. For workstations they had the 64-bit "Merced" CPU. Long in development with the computer-maker Hewlett-Packard, they later renamed it to the Itanium. And since this was 1997, it sparked "Titanic" jokes. They called it the "Itanic". A fitting reference as it turned out.
But another segmentation - one for servers - did very well. In April 1998, Intel unveiled the Pentium Xeon or just Xeon line of chips. They nicknamed it the "warrior princess chip", guess because of the cult-classic TV show Xena Warrior Princess. Intel positioned these powerful chips for beefy servers powering corporate networks or movie-making rigs like those used by the animation studio Pixar. Since these systems cost anywhere from $25,000 to $100,000, Xeon chips were very profitable.
Incumbents in this market included IBM mainframes or DEC's VAX minicomputers running proprietary software. And Sun Microsystems, which offered Unix-based Solaris applications running on self-designed RISC-based Sparc chips. At first Sun ridiculed Intel for entering the market. A group marketing manager said at the time:
> Xeon is just a Pentium chip ... It's desktop technology trying to play in the enterprise market. It looks from this perspective like Pentium's last gasp. How far can they push a 20-year-old architecture?
But Intel used its formidable manufacturing capacity to give their Xeon-based servers fast but cheap compute that competitors could not match. Meanwhile, we have the rising adoption of the open-source Linux operating system. Linux worked quite well and lacked the legal issues wafting over Unix. And being open source, it did not cost anything for people to use. Customers were more than willing to port their proprietary software to Linux.
The Linux + Intel combo beat Sparc and RISC so badly that even Sun caved and started making x86 servers too, perhaps an attempt to beat back IBM and Microsoft. In other words, Intel made Sun compete against its own in-house product. The concoction of free, useful software and cheap, powerful silicon is very potent indeed. ## Communications Barrett's strategy during his CEO tenure had two parts. First, he poured billions of dollars into R&D and new factories. This continued even during the slow years of the early 2000s. Barrett recounts in an oral history with the Chemical Heritage Foundation: > I think we went through ten full-blown up and down cycles in my thirty-two years here. Gordon [Moore] always said, “You don’t save
your way out of the recession. You invest your way out a recession.” These big factory investments covered major semiconductor breakthroughs covered on this channel like copper interconnects and the transition to 300 millimeter silicon wafers. But the sheer size of these investments - up to 30% of revenue - disappointed the financial community. The stock went to record highs during the Dotcom bubble but then fell into a deep doldrum afterwards. Worse yet, the stock traded at a relatively low price to projected earnings multiple, indicating that it was cheap ... and also that the market did not expect much future growth up head. So Barrett also started buying back stock. The buyback program began in 1990 under Andy Grove,
but it was under Barrett's watch that the stock buybacks really surged. In 1997, the year before Barrett took over, Intel bought back $3.37 billion of stock. In 2005, Intel bought back $10.6 billion of stock. They spent more on buying stock than capital equipment for the fabs. The second big thing Barrett had to work on involved those fabs. If the fabs are
not fully utilized, then they are losing money. With the PC market no longer growing as fast, that meant finding new markets for all that new factory capacity. In a 2001 interview with Tampa Bay Times, Barrett said: > During the '90s was the PC revolution and by Intel focusing all of its guns on the PC, I think we were able to ride the wave just at the crest and never fall off, and that's precisely where you want to be. I think we did absolutely the right thing.
> Now that that business has slowed down some, it's important for us to diversify, do some other things, and look for some other growth areas to supplement that sort of growth we had from just the PC. So that has required us to take some risk or gambles in moving into new areas, Barrett was very aware of the questions that this might bring about for the company. Was the company losing its focus? So he added this bit. > But you shouldn't read this so much as Craig Barrett became CEO a couple of years (ago) and therefore Intel's changed directions. > It's really the whole environment has changed. And I think we did absolutely the right thing under Andy Grove's CEO tenure - focus, focus, focus - because we had a huge opportunity.
> And now the environment's changed and so we should focus this change. Call that risk-taking, I call it being a pragmatist, a realist responding to the environment. Barrett eventually narrowed onto the communications market as Intel's next thing.
That was then a vertically integrated industry - highly fragmented with no dominant player. Intel intended to turn that into a horizontal structure like that of the PC industry. Barrett envisioned Intel chips powering data networking gear, infrastructure, handheld computers, mobile phones, and more.
Even after the bust, he remained bullish on the internet, e-commerce, mobile, and silicon in general. He noticed how chips were getting into everything, especially in the home, and positioned Intel to provide the central intelligence to connect and control all these things. This platform strategy would eventually come to be known as "Intel Inside Everywhere" or just "Intel Everywhere". In 2003, then-CTO Pat Gelsinger said at a company event: > Before I retire from Intel, I want a piece of Intel technology touching every human on earth, every minute of every day, in every aspect of their lifestyle The issue is that all of these industries already had players. Considering Intel's existing unequal relations with its PC-maker clients, did Barrett recognize how alarming such ambitions can sound to them? Asked about this, Barrett said in another interview in 2001: > This can be threatening. But it can give companies an opportunity to realign themselves
and make use of these building blocks and, just as importantly, adapt to a new economics ## Acquisition Spree Without time to organically enter these industries, Barrett started doing something that Intel had never really done before: Acquisitions. Intel would buy good products from startups and then leverage its big factories and integration to out-manufacture the incumbents. Intel's general manager of communications Sean Maloney said in 2002: > The name of the game in the next three to five years is high-end capacity and what you do with it ... We're the guy with the factories. We believe that capacity is a big deal in the next few years. And integration of design and manufacturing is a big deal
Between 1998 and 2002, Intel spent over $8 billion to assemble a communications and cell phone chip business. Some of the biggest acquisitions came prior to the Dotcom bust at rather elevated prices. For example, in 1999 Intel bought Level One Communications for $2.2 billion in stock. Level One provided chips for LAN/WAN applications, Ethernet controllers and other networking chips. That same year, Intel paid $1.6 billion to buy DSP Communications,
which made digital chipsets, designs, and software for mobile handsets. They were targeting the wireless processor market. They renamed DSP to XScale and extensively marketed it to Motorola, Palm and BlackBerry, to iffy results.
There were some encouraging signs. By 2003, Intel captured 11% of the Wi-Fi chipset market. Good progress after just a few months, but still counted for fifth place behind Broadcom, Globespan-Virata, Texas Instruments, and Atheros. However, many of these initiatives did not work out. Former employees grumbled about Intel putting semiconductor people in charge of these new businesses, who lacked proper mindset for the market. For example, Intel's move into chips for the DSL market. In 1998, they moved into
this market with a new standard they made with the help of Compaq and Microsoft called G.Lite. Back in the day, internet required a "splitter" to separate internet and phone service. A chip on the G.Lite standard theoretically would not have required a splitter, which was nice.
But Intel did not get the backing of the critical gatekeepers, the telecoms. Having not been asked for their input, they refused to adopt it. The technology later encountered technical issues in field tests, was never shipped, and sort of faded away. This happened repeatedly. By 2003, the communications division made $4 billion of revenue but lost $800 million in profit. Years later in 2009, long after his retirement, Barrett was asked his thoughts on these acquisitions and said: > "I bought high and sold low. But at least money was cheap in those days" ## AMD's Challenge So throughout the late 1990s and early 2000s, Intel was experimenting like a small-town freshman at the big-city university.
And just like those kids, Intel took their eyes off the ball. And suddenly found itself on its back foot, thanks to its old foe AMD. Back then, consumers used clock speed as a proxy for performance, especially across different processors. Not exactly the most correct way to do it - there is a reason why Steve Jobs called it the megahertz myth - but that is how they did it.
In 1999, AMD finishes their Fab 30 plant in Dresden running a 180 nanometer node, and announces a new line of Athlon chips. First at 900 megahertz and then potentially at 1 gigahertz in early 2000. In March 2000, AMD announced that they had beat Intel to the coveted 1 gigahertz number with their Athlon chips. Later that year, Intel brings out a Pentium III chip with a 1.13 gigahertz clock - just
slightly faster. They can now reclaim the fastest x86 crown. But tests by Tom's Hardware and [H]ard|OCP found that the new chip was unstable and did not run benchmarks properly. When told about this, Intel gallingly said at first, "If you really feel you have to write about it, then we can't keep you". Intel was later forced to pull the chip off the market. The clock speed battles continued throughout the next few years. Intel released Pentium 4 and the NetBurst architecture, which targeted even higher clock speeds.
But such speeds made the chips go so hot that they can literally boil pasta. The 2000s were wild days for computers. ## AMD and the Server And then there was the battle for the server. As I mentioned, Xeon had some success in the workstation and smaller server spaces, which appreciated the cheap compute. But the real money was in the large server space, big systems dominated by custom solutions made by DEC, Sun and IBM. Such systems were differentiated with proprietary components, software, and most of all - 64-bit capability.
Intel tried to create a 64-bit server chip with the Itanium. It was not only long delayed, but not backwards compatible with older software since it was a new architecture. AMD took a different route. They opted to extend the existing 32-bit x86 architecture to 64-bit. The resulting Opteron chip was not only very fast, but also compatible with existing operating systems, software, and hardware.
But despite a large marketing push and a pretty well-received product that many industry analysts believed to be as good or better than what Intel had, AMD did not make headway in the market. ## Retaliation Now Intel's people have long been a-holes. Stories abound but here is one recently told to me. One of the few companies with an x86 license is Taiwan's VIA Technologies. VIA produced some Pentium 4 chipsets - the earlier-mentioned chips
that control communications between the CPU and the rest of the computer. Intel did not want companies to use non-Intel chipsets, and sued VIA for patent infringement. They also allegedly intimidated partners to not use VIA chipsets. Then came the Computex trade show in 2001. VIA's marketing team floats a
bunch of VIA-branded balloons around the show. The Intel marketing people go around cutting the lines to the balloons, causing them to float to the showroom ceiling. Contemporary media dubbed it the "Battle of the Balloons". VIA sued Intel in
response. A Taipei court found Intel not guilty for destruction of property. So yeah, there was a reason why they were so unliked. But as PC market growth slowed and competitors clamored to finally steal their lunch, Intel seemed to go a bridge too far. AMD's inability to gain share against Intel eventually led them to sue Intel for alleged anti-competitive practices in 2005. Practices like paying vendors rebates if they only used Intel chips. Or threatening vendors with pulling rebates or even funding competitors if they should use AMD. A Dell presentation in 2003 said that Intel's retaliation to an AMD switch "could be severe and prolonged with impact to all Lines of Business".
Or modifying its own products to lock out players that challenged its position. Like redesigning compilers to work slower on non-Intel hardware. Nvidia accused Intel of introducing barriers that made it harder for Nvidia GPUs to interoperate with Intel CPUs. Not to mention bundling practices that favored Intel's own integrated graphics over Nvidia GPUs. Intel for its part fought vigorously. Intel and AMD settled their private lawsuit later in 2009. The US FTC then filed an anti-monopoly suit in 2009 which settled in 2010.
Over in Japan, their FTC ruled in 2005 that Intel violated anti-monopoly rules and required changes. Intel accepted the ruling and instituted the changes, though it didn't do much for consumers. The European Commission fined Intel about a billion dollars for these abuses, and this is the case that Intel battled in court for nearly 20 years. And I think in October 2024, they finally won. Props to Intel for sticking with it all those years. ## Otellini Takes Over When Barrett first joined Intel in 1974, the company generated about $50 million in revenues per year. By 2005, the company was doing $100 million revenue each day. The size
question continued to plague the company. For Intel, achieving 10-20% annual growth on a $30+ billion revenue base meant adding the equivalent of an AMD every single year. Barrett was now 64 years old and nearing retirement. And thus in 2005,
a new CEO took power: the aforementioned Paul Otellini, the company's first leader without an engineering or even technical background. He got his degree in business. Even at the time, people questioned Otellini’s qualifications to lead such an engineering-heavy company. But both Andy Grove and Craig Barrett publicly backed Otellini. Grove told Fortune Magazine in 2005: > Paul is a self-taught technologist. He knows the products better than I ever did
Barrett admitted that Otellini "is not so steeped in the basic technology as some of us were" but argued that such knowledge was not what Intel then needed. In an oral history he said: > My contention would be that the company is not just making transistors any longer. It’s making more complicated solutions for customers, and no matter who is CEO, the CEO is probably not going to have an intimate understanding of every aspect of our business. > I think Paul’s got a very good cross section of understanding. He’s got a lot of smart engineers. He can talk to them ... But I don’t think it’s a problem. If I thought it was a problem we probably wouldn’t have put him in that job.
He is not wrong by the way. Most CEOs do not need to and should not have to know the details of their business - Jensen Huang perhaps being the exception. Hire good people and trust them to do their jobs. The market welcomed Otellini's moves to continue pursuing the x86 server business and undoing some of his predecessor’s excessive diversifications. Including a 10,000 person layoff to streamline a corporate structure that many analysts felt gotten to be too bureaucratic. Things were starting to fall by the wayside,
and items on the roadmap slipping back and back. ## The iPhone Thing About a year into Otellini's tenure, Intel missed the design win for what would eventually be the iPhone. What has been said about this miss is very well known. Otellini did that infamous interview
with the Atlantic in 2013 where he spins it as a price and forecasted cost thing. If he ever followed up on those comments, it was not widely reported and he sadly passed away a few years later in 2017 from cancer. Eff cancer. Doubly sadly, Grove preceded him by a year for undisclosed reasons. I think Babbage over at the wonderful Chip Letter newsletter clarifies the mythology with the information we have at our disposal. A truly excellent piece.
Among other things, Babbage points out exactly how people at the time responded anytime Intel offered them something: They worried that Intel was going to lock them in and do to them what Intel did to the PC industry. Kind of reminds me of how the movies industry refused to cooperate with Apple after seeing what iTunes did to the music industry. He wrote that Intel was then in the midst of a refocusing on the x86 architecture - which included a sale of XScale to Marvell in 2006. This move to x86-all-the-things was brought upon by AMD's competitive pressure and the failure of their intended x86 successor Itanium. But the power consumption profile of such a device was nowhere near what Apple wanted. Ben Thompson of Stratechery - I affectionally call him
Ben Ben - touches on this in a 2022 interview with Tony Fadell, the father of the iPod. Fadell recalled that Intel came in with a PC mindset. One centered on x86 and repackaging whatever they had on the desktop for the mobile. He likened it
to how Microsoft during the Ballmer era was trying to make everything Windows. It is natural for big, successful companies to stick to what made it so big in the first place. But mobile turned out to be entirely different from the PC. Apple's success induced both Intel and Microsoft to spin their wheels over the next few years, repeatedly throwing out "PC-like mobile" products like Atom mobile processors and Windows Mobile/CE in search of something that can pass muster. ## Conclusion There is one last thing I want to add about that miss. In his biography, Jobs said about Intel:
> They wanted this big joint project to do chips for future iPhones. There were two reasons we didn’t go with them. One was that they are just really slow. They’re like a steamship, not very flexible. We’re used to going pretty fast. That part harkens back to the point I said earlier about the company's growing inability to execute on ambitious roadmaps.
> Second is that we just didn’t want to teach them everything, which they could go and sell to our competitors. Intel worked with IBM, and then turned around and worked with Compaq to bring up a competitor. Then they worked with Compaq's own competitors to compete against Compaq. They shut down potential challengers to their own dominant position. And by the
time it dawned on the PC market who was really running the show, it was too late. And I am sure that Jobs must have remembered how Intel once made their own web tablet, called the iPad, back in 2001. Intel's titanic success in the PC industry greatly enriched them. But it backfired on them in that when they tried to expand, their reputation preceded them. Incumbents in other industries did not want to work with them. Partners building the future
did not want to use them. Intel at the peak had to dance alone.
2025-01-07 18:33
