Today, Intel seeks to remake itself into a full-service foundry with Intel Foundry Services. But did you know that this is not their first foundry? Some 15 years ago, Intel had Intel Custom Foundry, or ICF. Most people today hardly remember ICF. Intel was so low-key about it. But at its peak it had a billion dollars in revenue contracts, strong support by the ecosystem, over a thousand multi-national employees ... And ten plus customers - many won over from TSMC. In today's video, the rise and fall of Intel’s first foundry.
And the lessons to be learned for its second. ## The Best Factories Intel in the 2000s believed that it had two Crown Jewels. The first of which was its leading edge manufacturing tech. For two decades, Intel under its CEOs Andy Grove, Craig Barrett and Paul Otellini developed and scaled a brand new technology to billions of chips. And did this every two years to hit the Moore's Law cadence. Such technologies spanned disciplines from lithography like the excimer lasers to new deposition methods like Atomic Layer Deposition to new materials like Hafnium Oxide.
Several times, the company's engineers made astute, bold technical choices that yielded where others did not. Like for example, the choice of "Gate Last" over the "Gate First" method, backed by the tech giant IBM and its allies. Even TSMC pursued "Gate First" before abandoning it halfway through development. Such choices granted the company a strong 1-2 generation lead over its competitors. And Intel's people certainly liked to hammer that point in their presentations.
But Intel's top management also knew that fabs were getting increasingly more expensive. And at some point, Intel's PC market revenues alone cannot justify the next node. And in those years, the semiconductor industry was still hoping to make the jump to 450 millimeter wafers. Larger wafers reduce per-die cost but with the drawback of massive investment. The proper strategic response then would be to roll up demand from outside fabless customers and use that to help fund the next node.
## Intel Architecture What Intel saw as its other crown jewel was the x86 chip architecture, or Intel Architecture. Before Otellini took over the CEO spot from Craig Barrett, Intel offered an Arm-based product called XScale. Otellini sold XScale to Marvell in 2006, seeking to refocus the company on x86. In an August 2007 interview with the Financial Times, he explained why he was throwing the company's whole weight behind x86: > And rather than rethink or rewrite the internet, which today runs on X86 machines, between us and our competition. I mean, the entire internet is coded for that base. All the browsers, all the Flash applications ... You catch that? I admit it is a bit garbled. But Otellini is
asserting that the whole internet was written on x86 machines, referring to Intel's then growing strength in Linux and Xeon-powered internet servers. He continues. > Rather than rewrite the internet to a new architecture, it’s a lot easier to – from my perspective, to do derivatives of our architecture to meet new form factor and power requirements The iPhone proved Otellini’s sentiments incorrect, and Intel sensed the immediate shift away from x86. In response, they rushed out the Intel Atom brand of low-power processors. These later powered small, cheap netbooks, but that wasn’t the original intention. ## Custom Intel Architecture Foundry Intel's first foundry began as a simple initiative by Paul Otellini to monetize these two crown jewel assets.
A small team was formed to provide "turnkey custom foundry services" for internal and external customers to produce custom or semi-custom chips. The foundry's original name was "Custom Intel Architecture Foundry" or CIAF. The name hints at the goal: Monetize Intel’s manufacturing factories, and also defend the threat to the Intel Architecture franchise from the trend of low-power systems-on-chip. The effort essentially kicked off in 2008 with an Intel Technology & Manufacturing VP discretely approaching existing customers, and touting the benefits of customizing Intel products using Intel manufacturing processes. This VP shortly left this role and was replaced by Sunit Rikhi, a vice president from Intel's Technology and Manufacturing division. Rikhi saw the foundry as an opportunity to build a new business and would serve as its general manager for the next seven years.
## Building a Foundry Yes, Intel could be called the leader in silicon manufacturing technology. But that technology had been tightly optimized by Intel’s Technology and Product groups to make fast x86 CPUs and not much else than that. It was less a question of ability, than rather those processes being in no shape to present to external customers. Think of Intel’s factories as like a pizza kitchen. For years, that pizza kitchen - its equipment, ingredients, processes and cooks - produced pepperoni pizzas per a recipe made by another internal Intel team. The two groups know each other well and how they worked. And together they
made the best pepperoni pizzas in the industry. But now we must reveal the kitchen’s capabilities to external customers, who then might use it to make cheese pizzas, Hawaiian pizzas, Chicago deep-dish pizzas, or New York style pizzas. Might even get the occasional weirdo asking for a calzone. A foundry like TSMC, UMC or GlobalFoundries will have done the extensive work needed to lay out its kitchen’s capabilities for a customer to pick and choose their pizza the way they want it to be. Intel needed to match that. Moreover, the Intel foundry team had to also equip itself with complementary IPs that its customers can use. Think toppings like olives, sausages,
and onions. Building this whole ecosystem required a lot of work. In the case of TSMC, they had been doing this for over three decades. And those years of cumulative knowledge have yielded a vast, widely varied menu as well as a kitchen capable of producing precisely what was ordered off it. A different mindset, and approach, from what Intel had always done before. ## Ditching Intel Architecture Custom Intel Architecture Foundry was the second prong of a two-pronged strategy by Otellini to bring a low-power x86 Intel Architecture, Atom, to mobile products. The first prong had been an alliance with TSMC. In this partnership, fabless customers can license Intel Atom IP and build systems-on-chips on top of them.
Less than a year later, however, this collaboration was put on hold due to low customer demand. The team at Custom Intel Architecture Foundry experienced the same low demand issues. This low demand stemmed from the Atom CPU's immaturity as an ecosystem or platform for building products. And even if there had been customer interest, it would have been an immense technical challenge to turn the CPU into usable IPs akin to those already offered by Arm. So some time in 2009, Otellini was convinced to drop the Intel Architecture part of the name. They would move forward with just Intel Custom Foundry. He reportedly did this reluctantly, and held it in his heart until his retirement that Intel Custom Foundry should be about Intel Architecture designs rather than that of others like Arm. Regardless, he gave ICF support and should be credited for as such.
## Competing With Customers There remained one major thing. Intel going into the foundry business raises the question of competing against your own customers. This issue dates back to the days of the Japanese semiconductor-makers, who gladly rented out spare capacity on their lines in exchange for a pound of flesh, your stock or your IP. TSMC was founded on the basis of never doing this, and it remains enshrined in their corporate charter. Intel would have been very much competing with many of its foundry’s largest potential customers. Companies like AMD, Nvidia, and increasingly, Qualcomm. Even if Intel was not in that business right then, there lingered the worry that that might change. The company’s future aspirations spread quite far and wide.
So the Intel management chose to do a slow and discreet rollout. They restricted the young foundry to only engaging with customers in areas that Intel does not compete in and had no aspirations for. Which was not a very large playing field. In the end, the Intel Custom Foundry team chose to compete in a specific type of chip called a Field-Programmable Gate Array, or FPGA. That was a large enough space that Intel then had no intention of getting into. Later. In the first two quarters after Intel went public with the news that it was
getting into foundry, analysts asked twice about Intel foundry’s impact on capacity. Then-CFO Stacy Smith, who is now on the board, replied both times that Intel was not focused on that business, was only talking to very specialized companies, and was not building a broad-based foundry. ## Big Customers That turned out to be alright, because Intel Custom Foundry was nowhere near ready to accept big customers. Shortly after news first emerged of Intel offering foundry services in late October 2010, Apple emerged as a potential foundry client. There was a brewing conflict between themselves and their current SOC foundry, Samsung. Later in 2011, Apple sued Samsung for infringing on Apple's intellectual property. It was time to go somewhere else.
Morris Chang wrote in his autobiography - which for now is only available in Chinese - that in November 2010, Terry Guo of Foxconn rung him to say that he would bring an Apple executive to dinner. That executive was Jeff Williams, today Apple's COO. Williams showed interest in engaging TSMC to produce iPhone SOCs and the two sides started working on a deal. But just as this work ramped up, it was suddenly put on pause. Jeff Williams called Morris to tell him that Paul Otellini was negotiating with CEO Tim Cook on a potential foundry deal. The TSMC-Apple collaboration would be paused for two months.
Morris then wrote about visiting Apple in April 2011 asking about things. Cook simply tells him that "Intel is not good at 'contract manufacturing'" (他們不擅做代工). Let's drill into that a bit. In late 2010, Intel Custom Foundry was only about two years old. They were profoundly
unprepared to take on what is perhaps the most demanding tech company client in the world. I want to remind you. The most important decision that any fabless company can make is their choice of foundry partner. Get that wrong, and nothing else matters. For such a company-defining decision, you don't want to take any risk. Intel Custom Foundry, after two or so years of work, was not yet 90% close to TSMC. Even if they were, that last 10% would be a big enough deal that Apple will still go with TSMC.
Apple's people took the meeting because their top management asked for it. But in terms of actual KPIs and milestones to get the deal, Intel was nowhere near ready yet. It was dead on arrival. There was another opportunity - maybe - for an Intel-Apple deal three years later, but I will get to that in a bit. ## Startup So Intel Custom Foundry was not ready for the major leagues.
But there were a few customers amenable to what Intel was offering. Those guys were the startups. They were tired of being so beholden to TSMC. They were also more tolerant of the competitive risks of working with Intel, thanks to intensely structured contracts that held Intel to severe punishments if they did not protect customer IP from Intel's Product divisions or provide the contracted capacity. The thing most important to these startups was Intel's league-leading silicon manufacturing. Intel Custom Foundry's first Process Design Kits were for the 32 nanometer node, but they quickly switched to 22 nanometers.
Because customers wanted to bring out a league-leading product and leapfrog their competitors. Considering how long it would take to bring a chip to the market, the 32 nanometer process node would already be lagging. Intel Custom Foundry's first reported launch customer was a small fabless FPGA startup called Achronix. Achronix's FPGAs were for specific tasks like network traffic or data encryption. The Wall Street Journal reported in late October 2010 that Intel agreed to give the startup access to about 1% capacity of its leading-edge 22-nanometer FinFET process. In an interview, Achronix's chief executive acknowledged that the Intel wafers cost more, but said, "This is a historic development. Intel is far ahead of anyone else in new process generations".
## 22 nanometers Intel's process lead on its competitors seemed the greatest when it came to the critical transition to 3D transistors. In late 2011, Intel announced its 22 nanometer process node equipped with its "Tri-Gate" variant of FinFETs. The FinFET is a type of transistor which covers the channel on three sides, offering superior power consumption and speed. It took over a decade for Intel to master the various bits of making the FinFET, which included controlling the variations you tend to get in the manufacture step. This work
remains one of the company's crowning achievements in semiconductor history. Intel originally scheduled to ship its "Ivy Bridge" 22-nanometer CPUs in April 2012. But they encountered some problems, and the ship date slipped to June 2012. A few weeks delay, but no big deal at the time. Intel was so far ahead of everyone else - at least two generations - that nobody thought more of it. For Intel Custom Foundry, Achronix’s Speedster22i FPGAs started to ship in volume in early 2013.
And those chips were good - ICF always did make good silicon once it was all said and done. These FPGAs were notable for consuming half the power of other high-end chips from other FPGA makers like Altera and Xilinx. The best working process node those other two FPGA companies had access to in those days was TSMC's 28-nanometer node - which was then still somewhat volume-constrained. So 22 nanometer was maybe one or two generations ahead. In February 2012, Intel added a second foundry customer, Tabula. Like Achronix, Tabula was another FPGA-maker startup, and they had been using TSMC's 40-nanometer process node. And then in April 2012, another start-up customer called Netronome. They previously
used TSMC's 65-nanometer process node. The news makes it clear. From the start, ICF was stealing customers from TSMC. ## Design-Manufacturing Integration Intel's external spokespeople regularly touted their design-manufacturing integration as the future of foundry. In October 2012, Intel scientist and senior fellow Mark Bohr said: > The traditional foundry model is running into problems. In order to survive, the foundries will have to become more like an integrated device manufacturer. Even some of the chief spokespeople for the foundries have said something similar.
> The foundry model worked well when traditional scaling was being followed and everybody knew where we were headed. In this era, where you continually have to invent new materials and new structures, it’s a lot tougher being a separate foundry and maskless design house. > Being an IDM, we have design and process development under one roof. That’s really a significant advantage. A fully-focused Intel foundry can provide more than just wafer fabrication. They can also leverage the organization's full line of design and advanced packaging services too. They can do the whole course from soup to nuts. Some of the packaging stuff was special. A bit later, Intel's teams invented a
new type of advanced packaging called Embedded Multi-die Interconnect Bridge. Per the name, this format embeds a small silicon chip into the package to connect different dies without the need for a more expensive silicon interposer. None of the Intel product teams could use it, but Foundry had customers who did. A perfect example of Intel's integration strengths. ## Behind the Scenes So theoretically, this all made sense. But GlobalFoundries sang the same tune of having IDM-like capabilities when they first spun off from AMD. Only to discover that in practice, the two sides - locked in their individual kingdoms - would struggle to come together.
Intel was no different. Many within Intel thought that Intel Custom Foundry was a waste of money, a side project distraction from what was really generating revenue. They were also worried what it might mean for their own KPIs. Intel’s Product people worried that having a foundry might raise outside concerns regarding Intel's own internal evaluation of its products.
Imagine if Intel Custom Foundry taped out another CPU and it turned out to be faster than what Intel Product could do. Imagine the egg on their faces. So they cried foul. And then Intel's Technology group. They sponsored and funded Intel Custom Foundry in part because they thought it would showcase the company's superior technology ... and maybe put those pesky Intel Product guys back in their place for complaining so much. But when that did not immediately turn out to be the case - when customers had feedback on things they wanted changed - the Technology Group's support fell out too. This left the Intel Custom Foundry team struggling to get their peers to change in order to better serve the outside world. The Technology people took a request from Intel
Product at least somewhat seriously because it impacted revenue. But Foundry? No chance. For example. Most larger fabless customers - from their experiences working with TSMC and other foundries - are used to getting some level of insight into a node. But the Intel Technology R&D team was extremely secretive and often refused to divulge such details. They also resisted making transistor changes along the power-performance curve to accommodate customers - seeing such things as distracting from their main goal of advancing the leading edge.
To manage customer frustration, the Intel Custom Foundry team regularly brought in senior management like the aforementioned Mark Bohr to meet with such customers and use his leverage to get R&D to cooperate. ## Altera The FPGA market's two largest players were Altera and Xilinx. The two were fierce rivals. Altera had been a long-time big customer of TSMC - one of their
top ten customers, actually. Meanwhile, Xilinx used TSMC's rival foundry UMC. But when UMC started to wither - a rumored issue at its 65 nanometer node delayed several Xilinx products - Xilinx defected. In February 2010, they announced that they had added Samsung and TSMC as their new leading edge partners. In an attempt to win the deal over Samsung and their 40-nanometer node, TSMC gave Xilinx access to its then-leading edge 28-nanometer node.
This pissed off Altera, which now needed a technical advantage they can leverage against Xilinx in the high-end FPGA space. This brought them to Intel. In 2013, Altera agreed to make FPGAs on Intel's 14-nanometer node - which followed 22 nanometers. Some trailing edge stuff would still be made at TSMC. CEO John Daane cited Intel's process lead as the reason for the move: > "Intel’s 14-nm is a second generation FinFET process, while others are just starting to implement their first" At the start, TSMC did not take ICF all that seriously. I heard second-hand of
a prominent former executive there being quite dismissive. So the Altera defection came as a real shock for the Taiwanese. Daniel Nenni recalled in a blog post on his SemiWiki website: > I was having coffee with a friend in TSMC Fab 12 when it was announced. If my memory serves it was Dr. Morris Chang who made the announcement and it honestly felt like parents were divorcing. In the company's next earnings call on April 2013, Morris Chang said this about the loss: > We have gained many customers in the last few years, but I really hate to lose even a part of an old one. We want them all really. I regret it.
And because of this we have thoroughly critiqued ourselves ... it's a lesson to us and ... we'll try our very best not to let similar kinds of things to happen again Chang then addresses Intel Custom Foundry's growing impact to TSMC itself.
> I still view Intel as a selective picker among customers. As a foundry competitor they will pick their targets and so on ... And I don't view them as a general competitor ... > But they are a very serious competitor to our customers. That really I would say applies even greater pressure on us than Intel as a direct foundry competitor. They are a very serious competitor to our customers.
You see what he did there? Morris is a very careful speaker. He dismisses ICF as a general foundry competitor, and then shifts the message again to his core point: Intel competes with its own customers. To me, it says everything: Intel Custom Foundry's progress was indeed making an impact on TSMC.
## 14 Nanometers Altera was without a doubt ICF’s biggest win yet. And it was not the only one. These early wins were promising enough to Intel senior management to slowly loosen the reins on the foundry's sector restrictions. New business came in from Panasonic and then LG, which contracted ICF to make specific system-on-chips for them. Switching to a new foundry always takes time. But it took a few months for Altera to even get started on their 14-nanometer design. This was because another FPGA project - the Arria 10, built on TSMC's 20-nanometer process node - had sucked out design resources. The Intel Custom Foundry
team had to dedicate many of their own design resources to ready the design for the fab. The design delay turned out to be fortunate because Intel's 14-nanometer node was a year late. The node's technology definition had been too aggressive. Not fatally so, but aggressive enough to force additional experiment cycles and thus extra time to mature into good yield. At the end of 2012, Intel announced that 14-nanometer should be ready by 2013. It did not reach high volume production until mid
to late 2014 - entirely missing the two-year Moore's Law cadence. The year-long delay also gave foundries like TSMC and GlobalFoundries - whose 14-nanometer node was being done in collaboration with Samsung - some time to close the gap. But like I said, not a fatal error. Just took some extra time to get out. Once mature, the node was fine - a second-generation FinFET process denser than TSMC’s first-gen FinFET 16-nanometer. The main issue was the impact on the node that came after it. ## 10 Nanometers In 2013, Intel and Apple met once more for discussions on a possible foundry deal. Apple was again in the market. Perhaps with the
intention to dual-source their next iPhone chip, the A9. Rumors abounded in the financial media that Intel was targeting Apple as their big customer. They were probably just kicking the tires. Intel Custom Foundry had gotten a few customers by now and was learning from them, but it was still a bit too early. The A9 second-source deal eventually went to Samsung.
Critically in those meetings, Apple indicated that it wasn't interested in the Intel 14-nanometer node, which in 2013 was then going through its slow maturation. But rather the node after that: 10-nanometer, originally due in 2016. Had it ramped on time, it would have done so 2 years ahead of TSMC's N7 node - which it was quite similar to. But famously, it didn't.
What went wrong with 10-nanometer? Intel has never said publicly, and the story is probably best reserved for another day. This is what I understand. Intel develops its process nodes with two teams working in parallel. One team works on what would be 14 nanometer gen+1 while the other works on 10, gen+2. So by the time senior management realized that 14-nanometer had a serious problem - perhaps some time in 2014 - the definitions for 10-nanometer had already been solidified. And 10-nanometer was a full 50% area shrink - with a 2.7x increase
in transistor density - on top of what had already been a very aggressive 14. SemiAnalysis's Dylan Patel added in a 2022 report that Intel’s use of cobalt in the vias and interconnects played a major factor too. The deposition tools for that were not ready yet, but Intel blazed ahead despite warnings.
And meanwhile, EUV - which I remind you that Intel funded and always believed in - was not ready in 2014 or even 2015. It would not be inserted into a process node until 2019, when TSMC did it for their N7+ node. Even if it was available earlier than that, it wouldn’t have been trivial to redesign the node to add EUV.
Long story short, it was too much. A step of hubris that took them off the cliff. But Intel's Product group had already taken the assumptions Technology group made for what 10-nanometer was promised to offer and developed products with them. Only to find out that those promises cannot be kept.
It was like building a 50-story skyscraper only to find out on the 20th floor that the foundations can't support the whole building. You can't just tweak this. You must tear it all down and start again - a 2-3 year process. Once the news about 10-nanometer's delays started reaching outside parties in 2015, the backlog of available business for the foundry rapidly began to dry up. By the time
10-nanometer fabbed Intel chips finally started to ship in 2019, the foundry no longer existed. ## The Altera Acquisition And then in June 2015, Intel announced that it would buy its biggest foundry customer Altera for $16.7 billion. CEO Brian Krzanich’s galaxy brain thinking behind buying Altera was that they can integrate FPGAs into their data center CPUs. Such an arrangement could let people offload certain workloads to those FPGAs - which meant that they wouldn’t need to buy a GPU to do those workloads. A relatively sound idea. And indeed after AMD acquired Xilinx, there have come out some interesting SoCs with similar arrangements.
But for Intel, the work of merging these two previously discrete chips - CPUs and FPGAs - turned out to be too much. The strategic notion soon dissolved and the mooted product never made it to market. From the perspective of Intel Custom Foundry, the Altera acquisition was a disaster. Assimilating Intel’s marquee external foundry customer sends a bad message to all of the foundry’s potential customers. It implied to those potential customers - guys like Qualcomm and such who were only starting to open up to an Intel foundry - that using ICF might mean Intel later trying to buy them too. Not to mention all the competition issues Morris talked about.
## End Alone, the Altera buy was a bad move. Combined with 10-nanometer's ongoing failure, it was crippling. In 2015, a number of major retirements, including that of General Manager Sunit Rikhi - who had run Intel Custom Foundry for nearly eight years - delivered the final blow. He was replaced by someone more amenable to Technology's perspectives. In other words, making the people in the Technology look good, rather than advocating for the actual foundry customers. So when the customers inevitably started to get angry with their inability to get through Foundry, senior management arranged for those customers to work directly with the Technology people. In other words, Intel Custom Foundry got cut out of the loop.
With the writing on the wall now obvious, the division - which once employed over 1,100 people around the world - quietly began to dissolve. By 2018 it was no more. ## Conclusion Circumstances have greatly changed between Intel Custom Foundry in 2008 and Intel Foundry Services in 2021. It seems like Pat Gelsinger's approach upon returning to Intel was to rebuild from scratch. I think that was a mistake. Basically throwing out the years of progress made by ICF. Intel Custom Foundry was no joke. Over eight years, they built up a book of 10+ customers and a billion plus dollars of contracted revenue. If TSMC considered them a real
competitive threat, then we should take their lessons and progress seriously too. So I will share a few takeaways of my own. First, the foundry service needs to have organizational support from across the whole organization. It should not be left awkwardly stuck between Technology and Product. At the very least, it should have the autonomy to modify its own technology.
Second, Intel's most significant technological pull is a working leading-edge node. That is what gets people in the door. Intel Custom Foundry could lean on Intel's multi-year process node lead on TSMC, Samsung and the rest of the industry. That lead no longer exists, and retaking it will be extremely difficult. Get 18A and 14A working right as a node for foundry customers - meaning more than just for making CPUs. That will not be easy.
Third, the only thing that closes the book is bankruptcy. Prepare for the long run. Let investors know that this is a very long term story. TSMC was not built in four, five or even ten years. Competing with them will be a ten year endeavor. Fourth, a little humility. Intel Custom Foundry felt like an unusually humble part of the company. Perhaps because it was a ragtag side project that had to serve outside customers. Perhaps because they hired over half of their workforce out of the fabless-foundry ecosystem.
Perhaps because they closely studied Morris Chang whenever he talked about the foundry business. Bring that humility to the rest of the company.
2025-01-16 21:55