What
In just 20 days you can get an update on four Mentor Graphics tools as used in the TSMC Open Innovation Platform (OIP). Many EDA and IP companies will be presenting along with Mentor, so it should be informative for fabless design companies in Silicon Valley doing business with TSMC.
Continue reading “Mentor Graphics Update at TSMC 2012 OIP”
Whenever people hear that I travel internationally one week a month they cringe at the thought of crowded airports, 12 hour flights, jet lag, and days packed with meetings. I generally shrug, accept the label of travel warrior, and say it is all part of doing business in the semiconductor ecosystem. But in reality, it is not as bad as it sounds, especially if you include the knowledge gained as part of the ROI calculation.
First and foremost, I rarely travel alone. As a business consultant I work for the CEOs of emerging technology companies advising them on a variety of topics. I wish I could say they learn more from me than I do from them but that is rarely the case. Successful emerging technology CEOs are a unique breed and are almost always a pleasure to work with. These CEOs wear so many hats and work so many hours that I get a headache just thinking about it. Every trip brings new problems and new ways to solve them which is an excellent learning experience on many different levels.
As a frequent traveler, I get perks from airlines and hotels fit more for a king than a regular working person. Executive lounges at airports are a good example. Right now I’m in the EVA airlines Evergreen Lounge sipping champagne and eating rice crispy treats (I don’t like caviar). Seat upgrades, personalized in-flight service, priority baggage handling etc…, it really does take the sting out of air travel. Even the airport garage I park my car in has frequent parking perks: Car washes, oil changes, priority shuttle services etc…
The hotels are the most generous. I get corporate rates for the cheapest rooms and the upgrades just keep coming. The rooms I usually get are huge with every modern convenience you can imagine. Sometimes the bathrooms are absolutely amazing. As you approach the toilet the seat raises, auto flushes, washes and dries, and closes as you walk away, semiconductor technology at its finest! The hotel welcome baskets are quite tasty and quite fattening, so I always bring my gym clothes to burn the extra calories that come with travel. Exercise also helps with jet lag. Another way I avoid jet lag is to NOT eat airline food and drink lots of water, believe it.
The boxes in the picture are Taiwan tea which is an interesting story in itself. A fellow international traveler convinced me that green tea is the key to a long and fruitful life and I have been drinking it ever since. That was more than ten years ago and I still believe it. I started with Japanese green tea but now Taiwan tea is my favorite.
In the hills right above the TSMC fabs in Hsinchu Science Park is one of the more famous green tea farms which I visit quite frequently. As the Taiwanese tea legend goes:
A tea farmer’s crop was infested by little green-leaf worms. Trying to recover from this catastrophic loss, the tea farmer went into town to sell his low-grade tea anyway. As it turns out the tea had a unique flavor and gained huge popularity. This guy was a great salesman for sure! When the tea farmer returned to his village and told the story it was viewed as bragging so they called his tea “puffing tea”, puffing means bragging in Taiwanese. The legend also says that Puffing Tea was a favorite of Britain’s Queen Victoria (1800’s) and she gave it the name Oriental Beauty Tea. European aristocrats later named it Champion Oolong Tea.
The tea is harvested three times a year. In the spring it is green tea (the green box) which has a mild flavor. Next harvest is orange tea. Fall harvest is red tea with the strongest flavor. If you look on the top of the boxes there are pictures of green-leaf worms which determines the price. The more worms the higher the price. There is more on Taiwan Tea Farms HERE. Even though Starbucks has invaded every corner of Taiwan, tea is still a big part of modern life here.
Just a reminder, the semiconductor industry is doing quite well thanks to the fabless semiconductor ecosystem. TSMC, my economic bellwether, reported another great month with a 32% increase over August 2011 and a 16% increase over January-August 2011. TSMC is forecasting Q3 at a 7% increase over Q2, which was an amazing 21% increase over Q1. Congratulations to everyone in the ecosystem on the amazing TSMC 28nm ramp!
The other great news is that according toIC Insights, semiconductor R&D spending will hit a record high of $53.4B in 2012. Great news and it came out on September 4[SUP]th[/SUP] which is my birthday. R&D is now 16.2% of the total $329.9B semiconductor industry sales and I am now well into my 50’s. Seriously, I make 52 look good.
It is interesting to see that 7 of the top 12 semiconductor R&D spenders are fabless and all 7 are TSMC customers. In fact every company on this list are TSMC customers including Intel and Samsung. I think another interesting number would be total fabless R&D dollars spent per year which would include all companies that are part of the fabless semiconductor ecosystem. Even if you just took ARM, ARM customers, and the foundries that make the ARM based products it would be greater than ONE TRILLION DOLLARS as compared to Intel’s $8.3B R&D spend in 2011. That is what you call crowd sourcing and the crowd will win every time. In case you had not guessed, I’m short on Intel and that was even before Intel’s revenue warning last week.
In July 2012, TSMC chief executive Morris Chang noted that his company’s 2012 R&D budget is now double the amount spent in 2009 (which was $656 million), in addition to raising capital expenditures 13% to an all-time high of $8.25 billion compared to $7.33 billion in 2011. TSMC’s R&D–to-sales ratio stood slightly above 8% in 1H12 versus 7.9% in 2011, 5.1% in 2005 and 3.1% in 2000.
It is also interesting to see that US based companies spent 57% of the total semiconductor R&D money while Japan spent 17%, EU 10%, Taiwan 8%, South Korea 7%, and China 1%. TSMC hit the top 10 R&D spenders in 2010 and is spending almost twice as much as the 12% published average. Also note that TSMC’s R&D increased 1% while Intel’s decreased 1%.
Capital expenditures (CAPEX), not to be confused with R&D spending, are also on the rise. The chart says $6B for TSMC but that is wrong. In fact, TSMC’s CAPEX in 2012 is currently $8B and could hit $8.5B by year end. TSMC is expected to increase its CAPEX by 25%, from $8B in 2012 to $10B during 2013 for 28nm expansion, 20nm ramp, and 16nm (FinFET) development. Intel is at $12.5B and Samsung is at $12.2B. Both Intel and Samsung’s 2013 CAPEX should exceed $13B but will certainly not grow 25%.
With the continuing migration from IDM (fab) to fab light to fabless expect to see TSMC outperform the semiconductor industry for years to come. Also expect to see me outperform the 50’s well into my 60’s.
Reports out of Taiwan (I’m in Hsinchu this week) have TSMC more than doubling 28nm wafer output in Q3 2012 due to yield improvements and capacity increases while only spending $3.6B of the $8.5B forecasted CAPEX! Current estimates have TSMC 28nm capacity at 100,000 300mm wafers (+/- 10%) per month versus 25,000 wafers reported in the second quarter. Wow! Talk about a process ramp! As I mentioned before, 28nm will be the most profitable process node the fabless semiconductor industry may ever see!
As a fitting postscript to my “Brief History of the Fabless Semiconductor Industry”, semiconductor research company IC Insights compiled a list of the top semiconductor companies for the first half of 2012. As the traditional IDMs go fabless and sink in the ratings, the foundries post record gains led by TSMC at 22%, GlobalFoundries at 18%, and UMC at 16%, WOW! GlobalFoundries also passed UMC for the #2 spot with $2B in Q1/Q2 2012 revenue versus $1.8B. You go GloFo! TSMC is still the foundry market segment leader with $7.9B.
In case you are not familiar with them, IC Insights is a leading semiconductor market research company offering coverage of current business, economic, and technology trends, top supplier rankings, capital spending and wafer capacity trends, the impact of new semiconductor products on the market, and other cool semiconductor industry information. I really like these guys, they are a bloggers best friend.
The biggest losers here are the Japanese IDMs; Toshiba, Fujitsu, and Renases which declined 26%, 23%, and 10% respectively. While Toshiba is riding the memory pricing rollercoaster, Fujitsu and Renases are in transition to the fabless business model moving manufacturing to TSMC. AMD also took an 11% hit which is not surprising as the company is still in “flux”. Check the blogs “The Coming Battle for AMD’s x86 Hidden Cache” and “AMD 2Q Financials — The Good, The Bad, and The (Pretty) Ugly” for more information on AMD.
U.S. Memory maker Micron moved up one spot with a 5% increase in sales. As Micron digests Japanese memory maker Elpida, expect another spot or two jump with an additional $2-3B in sales. Elpida is the last Japanese DRAM maker so this is an end of an era for the Japanese semiconductor industry. How long before Japan is completely fabless? Certainly within my lifetime (I’m 51).
Qualcomm, the fabless semiconductor wonder, posted a surprising 6% decline. Rather than taking personal responsibility, Qualcomm CEO, Dr. Paul E. Jacobs Ph.D, blames 28nm supply shortages even though Qualcomm is working with multiple foundries and I can tell you that TSMC has fulfilled ALL contractual commitments for 28nm. Either Qualcomm is bad at forecasting demand or this finger pointing is a diversion for deeper problems at Qualcomm. As TSMC clears all 28nm backlog in Q3 let’s see where the finger points at the next Qualcomm conference call. You can bet I will blog it because I really hate finger pointing. In stark contrast, the non-finger pointing Broadcom posted an 8% gain. And the mighty Intel posted a disappointing 5% increase based on lackluster Ultrabook sales and no position at all against ARM in the mobile market as of yet.
All-in-all a good financial start for what will be a great year in semiconductors. 28nm is breaking ramping and revenue records, 20nm tape-outs will commence in Q1 2013, the mobile market segment continues its meteoric rise, and the semiconductor ecosystem has never been stronger. So, do you still think that the fabless business model is broken Mr. Mark Bohr of Intel? I’m also not a fan of asinine PR stunts so I will continue to haunt Intel on this one.
In 1985 Morris Chang was recruited by the Taiwanese government to help develop the emerging semiconductor industry. In 1986 Morris joined the Hsinchu based non profit research institute ITRI as Chairman and President and launched what would be TSMC’s first semiconductor wafer fabrication plant on the ITRI campus. Taiwan Semiconductor Manufacturing Company Ltd. was officially formed in 1987 as a joint venture between the Taiwan government (21%), Dutch multinational electronics giant Philips (28%), and other private investors.
Without a doubt, TSMC created what is today’s semiconductor foundry business model. While at TI, Morris Chang pioneered the then controversial idea of pricing semiconductors ahead on the cost curve, sacrificing early profits to gain market share to achieve manufacturing yields that would result in greater long-term profits. This pricing model is still the foundation of the fabless semiconductor business model.
Even starting 2 process nodes behind competing semiconductor manufacturers (IDMs), TSMC was able to attract customers. 4-5 years later TSMC was only behind 1 process node and the orders started pouring in. In 10 years TSMC caught up with IDMs and the fabless semiconductor industry blossomed enabling a whole new era of semiconductor design and manufacturing. Today TSMC is the undisputed leader with a reported 49% share of the $30B foundry market segment. UMC is second with just over a 12% share, GlobalFoundries is a close third with 12%, and SMIC is fourth with 4.4%.
As the story goes, Morris Chang made the first TSMC sales calls in 1987 with a single brochure:
TSMC Core Values: Integrity, commitment, innovation, and customer trust.
It is interesting to compare the consistency of this statement with the 1997 TSMC mission statement:
We are building the world’s Virtual Fab! We provide the best quality technology, the greatest capacity and the highest standard of service. We are the most reliable choice as a partner in semiconductor manufacturing.
And again with the vision/mission statement on the TSMC website today:
Our vision is to be the most advanced and largest technology and foundry services provider to fabless companies and IDMs, and in partnership with them, to forge a powerful competitive force in the semiconductor industry. Our mission is to be the trusted technology and capacity provider of the global logic IC industry for years to come.
Also according to the current website TSMC has:
While the original intention of TSMC was to aid the Taiwanese semiconductor design houses, Dr. Morris Chang clearly had much larger aspirations which transformed the global semiconductor industry into what is today a $300B+ business that is mission critical to modern life.
A Brief History of Semiconductors
A Brief History of ASICs
A Brief History of Programmable Devices
A Brief History of the Fabless Semiconductor Industry
A Brief History of TSMC
A Brief History of EDA
A Brief History of Semiconductor IP
A Brief History of SoCs
Fabless semiconductor ecosystem bellwethers, TSMC and ARM, buck the trend reporting solid second quarters. Following “TSMC Reports Second Highest Quarterly Profit“, the British ARM Holdings “Outperforms Industry to Beat Forecasts“. Clearly the tabloid press death of the fabless ecosystem claims are greatly exaggerated.
“ARM’s royalty revenues continued to outperform the overall semiconductor industry as our customers gained market share within existing markets and launched products which are taking ARM technology into new markets. This quarter we have seen multiple market leaders announce exciting new products including computers and servers from Dell and Microsoft, and embedded applications from Freescale and Toshiba. In addition, ARM and TSMC announced a partnership to optimize next generation ARM processors and physical IP and TSMC’s FinFET process technology.” Warren East, ARM CEO.
ARM enters the second half of 2012 with a record order backlog and a robust opportunity pipeline. Relevant data for the second quarter, being the shipment period for ARM’s Q3 royalties, points to a small sequential increase in industry revenues. Q4 royalties are harder to predict as macroeconomic uncertainty may impact consumer confidence, and some analysts have become less confident in the semiconductor industry outlook in the second half. However, building on our strong performance in the first half, we expect overall Group dollar revenues for full year 2012 to be in line with market expectations.
Even more interesting is the recently announced TSMC / ARM multi-year agreement that extends beyond 20 nm technology (16nm) to enable the production of next-gen ARMv8 processors that use FinFETtransistors and leverages ARM’s Physical IP that currently covers a production process range from 250 nm to 20 nm.
“By working closely with TSMC, we are able to leverage TSMC’s ability to quickly ramp volume production of highly integrated SoCs in advanced silicon process technology,” said Simon Segars, executive vice president and general manager, processor and physical IP divisions, ARM. “The ongoing deep collaboration with TSMC provides customers earlier access to FinFET technology to bring high-performance, power-efficient products to market.”
“This collaboration brings two industry leaders together earlier than ever before to optimize our FinFET process with ARM’s 64-bit processors and physical IP,” said Cliff Hou, vice president, TSMC Research & Development. “We can successfully achieve targets for high speed, low voltage and low leakage, thereby satisfying the requirements of our mutual customers and meeting their time-to-market goals.”
This agreement makes complete sense with 90% of ARM silicon going through TSMC and the PR battle Intel is now waging against both ARM and TSMC. But lets not forget the Intel Atom / TSMC agreement of March 2009:
We believe this effort will make it easier for customers with significant design expertise to take advantage of benefits of the Intel Architecture in a manner that allows them to customize the implementation precisely to their needs,” said Paul Otellini, Intel president and CEO. “The combination of the compelling benefits of our Atom processor combined with the experience and technology of TSMC is another step in our long-term strategic relationship.”
Sorry Paul, clearly this was not the case. TSMC is customer driven and Atom had no customers. So there you have it. The agreement was “put on hold” less than a year later:
Intel spokesperson Bill Kircos said no TSMC-manufactured Atoms are on the immediate horizon, though he added that the companies have achieved several hardware and software milestones and said they would continue to work together. “It’s been difficult to find the sweet spot of product, engineering, IP and customer demand to go into production,” the Kircos said.
Given that wrong turn, the current Intel strategy is to offer ASIC services versus the traditional foundry COT (customer owned tooling) for Atom SoCs using a CPU centric 22nm process. This turnkey ASIC service is currently called Intel Foundry Services to which we have heard plenty but have yet to see any silicon. Just my observation of course.
Twenty eight nanometer is progressing very well. Our output and our yields are both above the plans that we set for ourselves and the plans that we communicated to our customers early in the year. Early in the year means January-February of the year, we set our plans in output and in yields and we, of course, ever since then we tried to exceed the plan and we had also communicated the plan to our customers at the time. And we have indeed exceeded the planned in both output and yields.
We expect to ramp up to about 68,000 wafers per month by the end of the year, 28 nanometer, 68,000 12-inch wafers per month by the end of the year. And by fourth quarter, we will be nearly caught up with the demand and we expect to fully meet the demand from the first quarter of 2013 on we will fully meet the 28 nanometer demand. It is also then that we expect that the 28 nanometer gross margin will catch up with the corporate average.
As I said today, both the defect density and use are better than 40 nanometer at the same stage of the volume ramp and they are also better than what we have – what we planned early in the year and what we communicated to our customers at that time.
Now, next a few words on 20 nanometer, 20 SoC. We have made very good progress on the 112 megabit SRAM yield. Now there are still challenges to overcome in meeting our yield plan of the entire chip. We have made very good progress on 112 megabit SRAM, there are still challenges to overcome in meeting our yield plan of the entire chip, which has both the logic and the SRAM on that, of course.
Now, our 20 nanometer SoC, we believe, is fully competitive with industry leaders, other companies’ 22 nanometer for the served available markets that we serve. For our markets, we believe our 20 SoC is fully competitive with anyone’s 20 nanometer or 22 nanometer offering.And, one important point to make is that our 20 nanometer has the industry’s leading metal pitch of 64 nanometers. Our leading competitors have 80 nanometer metal pitch. That allows an advantage in the device’s density and die size.
Now, as for the timing, we expect our 20 nanometer technology to be qualified by the end of this year and will be ready to support customers (inaudible) in Q1 of 2013. I think that we’ll start some production of 20-nanometer next year, but the small scale, very, very low, what we would call risk type of production basically, but 2014 will be a ramp year for 20-SoC.
Now today, last time I mentioned that we will have a FinFET product after 20 SoC. And today, I’m glad to say that we have been planning the 16 nanometer FinFET. Right after our 20 nanometer (inaudible), which is the 20 SoC, we will offer FinFET at 16 nanometer for significant active power reduction. We expect to achieve speed and density, speed and logic density levels comparable to industry’s leading players 14 nanometer FinFET.
So, we expect our 20 SoC to be competitive with competitors’ 22 nanometer or 20 nanometer products and we expect our 16 nanometer FinFET to be competitive with our competitors’ 14 nanometer FinFET products. You might ask why are we calling it 16. The only reason, in fact, until two days ago, we were undecided on whether to call it 14 or 16 FinFET. Now the only reason we decided to call it 16 FinFET is first, we want to be somewhat modest; second, we are told quite a few major customers ask the 16 FinFET, that designation and we didn’t want to confuse our customers by now switching to 14. But we expect it to be competitive with other people’s 14 nanometer offerings.
Now 16 nanometer FinFET, our 16 nanometer FinFET, is expected to deliver about 25% speed gain given the same standby power over the 20 nanometer SoC. It is expected to give 25% to 30% power reduction at the same speed and the same standby power, and for mobile products, it is expected to give 10% to 20% speed gain at the same total power. As for timing, we expect it to be about one year after 20 SoC namely it should be ready for risk production at the end of 2013 or early 2014, about one year later than the 20 SoC.
Now, why are we having such high capital intensity? Now, well, I think this is actually a focus point of our internal discussion among our top level managers for the last two years now. And basically, we invest in capacity to get future growth. So you look back at history. If you look at our TSMC’s history, during ‘97 and ‘02, between 1997 and 2002, during that six-year period, TSMC’s capital intensity ratio stayed mostly above 60%, 60% during that six-year period. And then was, as you recall, there was a high-tech bubble bursting in late 2000 and early 2001. But in spite of that, our revenue CAGR between ‘97 and ‘07, here after having spending a lot of capital, having sustained high capital intensity for six years, ‘97 to ‘02, our revenue CAGR between ‘97 and ‘07, that’s a 10-year period, was 20%, compounded annual growth rate of 20% in revenue during the 10-year period; the first six of which was marked by high capital intensity. During that 1997 to 2007 period, foundry industry growth was 16% in the same period and ours was 20%. As a result, our market share rose from – foundry market share – from 31% in ‘97 to 43% in ‘07.
And as far as this year’s CapEx is concerned at this point, we are still following the guidance that we gave you last time; I believe NT$8 million to NT$8.5 million. At this point we are still following that. But next year, I am – we are not going to forecast until early next year. But I think I have already given you a view of our reasoning and our strategy and our objectives. So but as to the exact number, I will not give you until early next year.
13 page transcript is HERE.
We all knew this quarter would be big but maybe not this big. Not all good news though so keep on reading. The news coverage is all over the map, mostly because they have no idea what a pure-play foundry really is. They also underestimate the power of mobile computing which should be a “Revenue by Application” market segment itself. I can’t wait to read what the tabloid analysts/journalists have to say about the results.
Yes, I’m still in Taiwan thus the early morning post. Tomorrow I go to Taipei for some hard earned rest and relaxation. I’m a big fan of trying new things but I just could not bring myself to eat this cricket. The person who did eat it said, “Not as good as I expected.” So there you have it. The fish I did eat and it was very good. I lost the staring match though, that fish did not blink ever. Squid on a stick? No thank you. Double click to enlarge at your own risk.
“The world’s biggest contract chipmaker, and rivals including Samsung Electronics and AMD, could face a downturn in global technology spending as Europe’s woes continue to dent demand and China’s economy slows.” Reuters
AMD is TSMC’s customer not rival and Samsung’s $390M 2011 foundry revenue does not rival TSMC’s $14.5B. Intel and Samsung are rivals, Intel and AMD are rivals, Xilinx and Altera are rivals, etc… I can’t wait to read John’s Cooley’s expert analysis. That should be a hoot. Last time John was in a fab was never. Here is the call transcript, good luck John!
“It makes complete sense to dedicate a whole fab, or two whole fabs, in fact, to just one customer,” Morris Chang said today. He didn’t name any clients it may offer a whole factory to or say whether it has immediate plans to do so. Bloomberg Businessweek.
I will give you three guesses who that customer is and they had all better be fruit.
While all this looks good I do see that some clarification is required:
That is a very interesting process ramp. Stay tuned, I will ask around a bit before I leave Taiwan. Post your guesses in the comment section or email it to me privately. I’m sure this will turn into great tabloid fodder.
Second-quarter net income rose to NT$41.8 billion, missing the NT$42.2 billion average of 19 analyst estimates compiled by Bloomberg. The earnings result included a NT$2.68 billion charge for its 5.6 percent stake in Shanghai-based SMIC. Consolidated revenue, reported earlier, rose 16 percent from a year earlier to NT$128.1 billion, beating the NT$127.2 billion average of analyst estimates and surpassing the company’s own forecast of NT$126 billion to NT$128 billion.
Looking forward:
Revenue in the three months ending Sept. 30 may be NT$136 billion ($4.5 billion) to NT$138 billion, the Hsinchu, Taiwan- based company said today, compared with the NT$134.8 billion average of 22 analyst estimates compiled by Bloomberg. Second- quarter operating income rose 23 percent to NT$46.7 billion, beating the NT$45 billion average of 15 estimates.
Other reported news is that fabless inventory levels have been increasing due to a decreasing global economy, which would lead to an inventory correction in the fourth quarter and a dip in TSMC’s revenue, which could continue into the first quarter of 2013. Don’t believe a word of it. It is going to be a very merry mobile Christmas! An iPhone5 and iPad Mini for all! Ho ho ho…
What does mango beer have to do with semiconductor design and manufacturing? At a table of beer drinkers from around the world I would have never thought fruity beer would pass a taste test, not even close. As it turns out, the mango beer is very good! Same goes for 20nm planar devices. “Will not work”, “Will not yield”, “Will not scale”, as it turns out 20nm is very good!!! The leading edge fabless people who were scratching their heads six months ago are now scheduling 20nm tape-outs for Q1 2013. Crowd sourcing wins again!
The 20nm process node represents a turning point for the electronics industry. While it brings tremendous power, performance and area advantages, it also comes with new challenges in such areas as lithography, variability, and complexity. The good news is that these become manageable challenges with 20nm-aware EDA tools when they are used within end-to-end, integrated design flows based on a “prevent, analyze, and optimize” methodology.
I agree with this statement 100%. It comes from a Cadence white paper A Call to Action: How 20nm Will Change IC Design. 20nm was definitely a turning point for the semiconductor ecosystem. Now that the technical wrinkles have been ironed out let’s look at how the fabless business model has evolved.
“There is no doubt we are at a crossroads at the most advanced process technology nodes. In order to take positive steps forward, significant monetary and collaborative investments and resources are required from both the manufacturing and design sides of the equation,” said Ana Hunter, vice president of Samsung’s North American foundry services.
I spoke with Ana before I left for Taiwan. We are in agreement: The industry is at an inflection point and the business model is changing. A simulated IDM environment is required for fabless semiconductor companies to be competitive at the advanced process nodes, absolutely. Check out the new Samsung “Best of Both Worlds: IDM and Pure-Play Foundry” brochure.
Before, during, and after DAC 2012 I asked various engineers from the top fabless semiconductor companies what has changed for them in regards to how they work with the foundries on new process technologies. I also asked the foundries. The IDM-like answers did not surprise me since working with EDA and IP companies on new process nodes is what I do during the day. It may however surprise others who are on the sidelines and believe the latest Intel PR nonsense.
UMC actually pioneered this simulated IDM environment with Xilinx from .25 micron to 40nm. Xilinx employees literally consumed an entire floor of UMC HQ for more than a decade and acted as the dutiful wife delivering many new processes. After the Xilinx 40nm divorce, UMC is no longer monogamous and now has multiple process development mistresses including TI, Qualcomm, and IBM.
TSMC took a different approach which many people overlook. The early days of collaboration started with Reference Flow 1.0 which is now version 12.0 with many new sub flows to be rolled out at TSMC OIP in October. This collaboration included both established and emerging semiconductor and EDA companies. Next came the IP effort with TSMC developing physical IP for reference and production at zero cost to customers. Commercial IP was next with the TSMC “silicon proven” program. IP companies big and small completed an exhaustive qualification program to make it into the TSMC IP catalog. Next came the TSMC “Early Access Program” where a select group of customers and partners were included in process development activities. Correct me if I’m wrong but I believe this started at 90nm. The qualification process for Early Access was also daunting but clearly it was critical to the evolution of the fabless semiconductor ecosystem. The result being the TSMC DTP (Design Technology Platform) Division which employs hundreds of people and has spent hundreds of millions of dollars (my guess) building the industry leading “Simulated IDM platform” you see today!
That brings us to 40nm. At 28nm and even more so at 20nm the fabless people have taken up residence in Hsinchu and the foundry people have “Early Access” to the top fabless companies. Bottom line: you will be hard pressed to differentiate between a Qualcomm, Broadcom, Nvidia, or Xilinx and a modern day IDM, except of course for the actual ownership of the manufacturing equipment. And lets not forget, the now fabless TI, AMD, Fujitsu, LSI Logic and others used to be IDMs, right?
Thank you again UMC and TSMC for leading the way!
