Paul Otellini, CEO of Intel, in conversation with Berkeley-Haas Dean Rich Lyons
- Thread starter hist78
- Start date
hist78
Well-known member
I created a Google NotebookLM AI generated audio overview of this 53-minute 2012 interview. You can listen to it at the link below:
Source: https://notebooklm.google.com/noteb...tifactId=a0b3bc54-65f0-459b-8b55-a1474217ecc3
Last edited:
He will be known as the man who missed mobile!
semiwiki.com
Paul Otellini Wiki - Semiwiki
Born: October 12, 1950 – San Francisco, California, U.S.Died: October 2, 2017 – Sonoma County, California, U.S.Education: B.A. in Economics, University of San Francisco (1972) M.B.A., University of California, Berkeley (1974) Occupation: Business executiveKnown For: 5th CEO of Intel Corporation...
semiwiki.com
hist78
Well-known member
He will be known as the man who missed mobile!
Paul Otellini Wiki - Semiwiki
Born: October 12, 1950 – San Francisco, California, U.S.Died: October 2, 2017 – Sonoma County, California, U.S.Education: B.A. in Economics, University of San Francisco (1972) M.B.A., University of California, Berkeley (1974) Occupation: Business executiveKnown For: 5th CEO of Intel Corporation...
Paul Otellini and Intel didn’t ignore the mobile opportunity. The problem was that he and Intel insisted on embracing mobile the Intel way, trying to force the world to follow Intel’s steps in the mobile revolution.
In the PC revolution, Intel had two powerful allies, IBM and Microsoft, to help establish its dominance. In the mobile revolution, Intel was a lonely player with the ambition to assert its dominance again. And it failed miserably.
I remember Paul telling a story about an Intel internal discussion about the iPhone. An unnamed executive said it was a joke. Phones belonged on desks next to Their PCs. In a nutshell that was the problem with Intel and their old school executives and board members.
hist78
Well-known member
In this 2012 Berkeley interview, Paul Otellini stated that Africa would probably not go through a wide PC deployment due to its lack of wired communication infrastructure. Instead, he believed mobile devices would be a better fit for the region and he was correct.
However, he may have overlooked the fact that mobile devices (smartphones and tablets) would also outsell or replace PCs in the United States. Many Americans would carry and constantly use their smartphones whole day without ever touching or even owning a PC.
And at the time of this 2012 interview, Paul Otellini was also serving on Google’s Board of Directors. He was an insider through and through!
From the Interview:
"I never thought that until a few years ago, but you can see that there's different models. For example, I don't think you'll ever see large-scale personal computing in Africa. It's unlikely. Africa will grow up with one of the best wireless networks in the world because they don't have to deal with the wireline legacy.
They're probably going to start out with their first computer as a smart phone. And it's very likely that most of Africa will live on increasingly smartphones and tablets or some kind of hybrid device over time.
And the personal computer revolution may not ever have to go there."
user nl
Active member
Interesting also how at some point PO discusses the costs of building new fabs, at that time in Oct 2012 PO mentions some 5 B$, to increase to some 10 B$ around 2018. And how he sees only INTEL and Samsung being able to keep up with that. He basically sees no other company able to keep investing like that in new fabs, he doesn't name TSMC as a potential third player.
Not sure if this was pure arrogance or ignorance or both. But if I understood well INTEL tried at some point to have ASML move to develop 450 mm wafer lithography tools, and have the whole industry move with it. So, that no other company could afford building leading edge fabs anymore, except INTEL and Samsung. TSMC got so concerned about this move to 450 mm that they actively worked to kill this. Finally they announced this in 2013, as told by this TSMC engineer, see eg this link:
https://www.theregister.com/2022/08/08/how_tsmc_killed_450mm_wafers/
Maybe that "killing of 450 mm" was also a crucial move by TSMC enabling them to eventually outcompete INTEL as a foundry of leading edge logic.
How TSMC Contributed to the Death of 450mm and Upset Intel in the Process - Semiwiki
Pinpointing exactly when 450mm died is tricky. Intel’s pullback in 2014 has been cited as a pivotal moment because it was the main backer of the proposed transition, as it had been for the shift to 150mm (6-inch) wafers in the early 1980s. However, the participation of global foundry leader TSMC...
semiwiki.com
hist78
Well-known member
How TSMC Contributed to the Death of 450mm and Upset Intel in the Process - Semiwiki
Pinpointing exactly when 450mm died is tricky. Intel’s pullback in 2014 has been cited as a pivotal moment because it was the main backer of the proposed transition, as it had been for the shift to 150mm (6-inch) wafers in the early 1980s. However, the participation of global foundry leader TSMC...
I believe those writings are based on the Oral History of Shang-Yi Chiang, recorded on March 15, 2022, by the Computer History Museum of Mountain View, California. It’s a very informative and interesting interview.
Source: https://archive.computerhistory.org/resources/access/text/2022/07/102792671-05-01-acc.pdf
In addition to TSMC’s opposition, the interview noted that semiconductor equipment companies, such as ASML, Applied Materials, KLA Tencor, and Lam Research, were unwilling to shoulder the high costs of developing 450 mm technology.
Last edited:
user nl
Active member
How TSMC Contributed to the Death of 450mm and Upset Intel in the Process - Semiwiki
Pinpointing exactly when 450mm died is tricky. Intel’s pullback in 2014 has been cited as a pivotal moment because it was the main backer of the proposed transition, as it had been for the shift to 150mm (6-inch) wafers in the early 1980s. However, the participation of global foundry leader TSMC...
Dan: thanks, I was not aware of that extensive discussion on SemiWiki, it seems one can write many history books using all the material of this webportal!
Last edited:
user nl
Active member
Yes!
That still counts as a miss.
Intel faced a similar issue with x86 CPU competitors in the mid to late 1990s and they did the unthinkable. Released low cost Celeron chips to chase off the competition - and it worked. They were willing to forgo their high margins to address the long term threat. They also risked devaluing the brand with "lesser products". Instead, Cyrix, Transmeta, and all other x86 clones were made irrelevant in relatively short order, and AMD only survived due to multiple Hail Mary's. (Nexgen acquisition and K7).
It's a little questionable that the abolishment of 450mm wafers was the reason why TSMC won against Intel.
I know it's not the decisive factor, but I think it's a little different from the rise of TSMC.
hist78
Well-known member
I too wondered why Paul Otellini didn’t mention TSMC as one of the few companies that could afford to build the most advanced fabs in the coming years. Unfortunately, Mr. Otellini passed away in 2017 at the age of 66, just five years after retiring from Intel. We can only guess his reasons at the time, it could have been a mix of factors.
He might have underestimated how soon a foundry like TSMC could catch up to Intel in semiconductor manufacturing technology. And it wasn’t just him; for a long time, Intel executives believed the fabless/foundry model would not be able to sustain the scale, volume, upgrade frequency, and high cost of fab equipment. Obviously, that assumption proved wrong. Smartphones/Mobile revolution and TSMC’s pure-play foundry business model ultimately turned the tide.
As a side note, TSMC and Intel maintained a very friendly relationship since TSMC’s incorporation in 1987, with the exception of the interruption during Pat Gelsinger’s years as Intel CEO.
In 1988, Intel became one of TSMC’s customers after Andy Grove’s visit to TSMC, and after TSMC implemented more than 100 improvements suggested by Intel. TSMC founder Morris Chang and Intel’s founders Robert Noyce and Gordon Moore had been friends since the 1950s. On one occasion, at Morris Chang’s request, Paul Otellini even recommended Craig Barrett, then retiring as Intel’s chairman, for a seat on TSMC’s Board of Directors. In the end, Craig Barrett declined the invitation.
user nl
Active member
No, but perhaps it enabled TSMC to catch up. I know that ASML was very relieved that although they got the money by these three companies to do both, 450 mm development and EUV development, a pause and final silent death to 450 mm happened.
ASML's engineers had their hands full improving the source power of the EUV tool and after they bought CYMER to get this EUV source project under control, they had no desire to spend more money and engineers on 450 mm. And for ASML as tool supplier it was economically just smarter to have more 300 mm EUV tools than fewer 450 mm tools, besides all the technical issues of accelerating such huge wafer stages at incredible speeds.
There were always close ties between ASML and TSMC, they both got kick-started by Philips NL.
About ASML's Customer Co-Investment Program
Three ASML customers - Intel, TSMC and Samsung - have agreed to contribute EUR 1.38 billion to ASML's research and development of next-generation lithography technologies over five years, specifically aimed at accelerating EUV lithography and 450mm lithography development. As part of the Customer Co-Investment Program, but separate from the R&D contribution, ASML has now received EUR 3.85 billion for issuing shares to the three participating customers. This cash will be returned to shareholders (excluding participating customers) via a Synthetic Buyback later this year, which through a reverse split will ensure that no dilution occurs from the co-investment program.
hist78
Well-known member
Another side note:
Morris Chang recalled that Carly Fiorina, who joined TSMC’s board in 2006 and was a former executive at Lucent Technologies and CEO of HP, responded to Craig Barrett’s decision not to join the TSMC Board of Directors by saying, “We’re already doing great without him.”
gc_epictetus
New member
We always assume Apple's mobile business was Intel's to miss, but Tony Fadell said Intel wouldn't have won even if Otellini said yes. Intel chips were made to be plugged in and were too power hungry; plus Intel moved too slow and was too inflexible - Apple wanted design control.He will be known as the man who missed mobile!
Paul Otellini Wiki - Semiwiki
Born: October 12, 1950 – San Francisco, California, U.S.Died: October 2, 2017 – Sonoma County, California, U.S.Education: B.A. in Economics, University of San Francisco (1972) M.B.A., University of California, Berkeley (1974) Occupation: Business executiveKnown For: 5th CEO of Intel Corporation...
“We ended up not winning it or passing on it, depending on how you want to view it. And the world would have been a lot different if we'd done it.” Paul Otellini
tooLongInEDA
Moderator
The Otellini quote is interesting and revealing. Even long after the event, he is either unable or unwilling to own up to a clear view of what happened and why (it was either A or B ...). And coupled with that, he doubles down on the view that history would have been significantly different had Intel built the initial Apple chips.
Looking back now, I'm far from convinced that this last view is correct. Intel might have held the line for a while. But the fundamental forces that TSMC and the fabless design ecosystem unleashed combined with the strong desire of the design ecosystem not to have Intel eating all the industry profits (as with the PC) added to the fundamental drive for low power (never an Intel strength) would surely have forced much the same eventual outcome. I think we're just arguing about the timescale and not the final result.