Posted on

Making your ARMs POP

Making your ARMs POP
by Paul McLellan on 04-16-2012 at 6:30 am
Categories: Arm, Foundries, IP, TSMC

Just in time for TSMC’s technology symposium (tomorrow) ARM have announced a whole portfolio of new Processor Optimization Packs (POPs) for TSMC 40nm and 28nm. For most people, me included, my first question was ‘What is a POP?’

A POP is three things:

  • physical IP
  • certified benchmarking
  • implementation knowledge

Basically, ARM takes their microprocessors, which are soft cores, and implements them. Since so many of their customers use TSMC as a foundry, the various TSMC processes are obviously among the most important. They examine the critical paths and the cache memories and design special standard cells and other elements to optimally match the processor to the process. They don’t do this just once, they pick a few sensible implementation choices (highest performance 4 core for networking, medium performance dual core for smartphones, lowest power single core for low end devices). A single POP contains all the components necessary for all these different power/performance/area points. Further, although we all casually say things like ‘TSMC 40nm’ in fact TSMC has two or three processes at each node to hit different performance/power points, so they have to do all of this several times.

Then they provide the performance benchmarks that they managed to achieve, along with all the detailed implementation instructions as to how they did it. These are EDA tool chain independent since customers have different methodologies. But the combination of IP and documentation should allow anyone to reproduce their results or get equivalent results with their own implementations after any changes that they have made for their own purposes and to differentiate themselves from their competitors.

Companies using the POPs get noticeably better results than simply using the regular libraries and doing without the specially optimized IP.

About 50% of licensees of the processors for which POPs have been available seem to have licensed them, currently there are 28 companies using them. Here’s a complete list of the POPs (click to enlarge):
Of course ARM has new microprocessors in development (for example, the 64 bit ones already announced) and they are also working closely with foundries at 20nm and 14nm (including FinFETs). So expect that when future microprocessors pop out that a POP will pop out too.

About TSMC
TSMC created the semiconductor Dedicated IC Foundry business model when it was founded in 1987. TSMC served about 470 customers and manufactured more than 8,900 products for various applications covering a variety of computer, communications and consumer electronics market segments. Total capacity of the manufacturing facilities managed by TSMC, including subsidiaries and joint ventures, reached above 9 million 12-inch equivalent wafers in 2015. TSMC operates three advanced 12-inch wafer GIGAFAB™ facilities (fab 12, 14 and 15), four eight-inch wafer fabs (fab 3, 5, 6, and 8), one six-inch wafer fab (fab 2) and two backend fabs (advanced backend fab 1 and 2). TSMC also manages two eight-inch fabs at wholly owned subsidiaries: WaferTech in the United States and TSMC China Company Limited, In addition, TSMC obtains 8-inch wafer capacity from other companies in which the Company has an equity interest.

TSMC’s 2015 total sales revenue reached a new high at US$26.61 billion. TSMC is headquartered in the Hsinchu Science Park, Taiwan, and has account management and engineering service offices in China, Europe, India, Japan, North America, and, South Korea.

Posted on

The Truth of TSMC 28nm Yield!

The Truth of TSMC 28nm Yield!
by Daniel Nenni on 04-15-2012 at 7:00 pm
Categories: Foundries, TSMC

As I write this I sit heavyhearted in the EVA executive lounge returning from my 69[SUP]th[/SUP] trip to Taiwan. I go every month or so, you do the math. This trip was very disappointing as I can now confirm that just about everything you have read about TSMC 28nm yield is absolutely MANURE!
Continue reading “The Truth of TSMC 28nm Yield!”

Posted on

Intel’s Fait Accompli Foundry Strategy

Intel’s Fait Accompli Foundry Strategy
by Ed McKernan on 04-05-2012 at 1:09 am
Categories: Foundries, FPGA, Intel Foundry, TSMC

As many analysts have noted, it is difficult to imagine what Intel’s foundry business will look like one, two or even three years down the road because this is all new and what leading fabless player would place their well being in the hands of one who is totally new at the game. I would like to suggest there is a strategy in place that will soon lead to tectonic shifts in the semiconductor world. The assembled pieces of “no-name” startup chip companies building in Intel’s advanced 22nm trigate process include Achronix, Tabula and now Netronome. Each represent three possible solutions to high performance data path processing that may lead to Intel’s goal of dominance in the combined server, storage, networking platform. Or, perhaps they may serve as a forcing function for leading Altera, Xilinx, Broadcom, Marvell or Cavium away from TSMC and partnering with Intel. Either outcome is a win for Intel.

For much the past three years the spotlight has shined brightly on everything that is mobile – as it should have. Questions about Intel’s ability to either counter Apple’s ARM based mobile rise or to be its eventual supplier across the board will be on every analysts mind until there is resolution. However, there is another side to Intel’s business that is not well understood. Intel always fights a multi-front war to maximize its advantage and overwhelm competitors without similar magnitudes of resources. Only Intel, historically, has been able to do this.

Today, while it charges ahead with its Medfield processor in the smartphone and tablet space to blunt ARM’s early lead, Intel enters a mopping up phase in the PC market with its Ivy Bridge based Ultrabooks that will neuter AMD and nVidia in what will be the highest volume segment by the end of 2013. And in the background Intel has opened up a third front against the foundries of TSMC, Global Foundries and Samsung who the ARM Camp depends on to win the Mobile Tsunami Marketplace. Without a process within spitting distance of Intel, ARM would be relegated to trailing edge embedded SOCs. Therefore Intel will leverage its Fabs to peel away Foundry customers, cutting off oxygen that pays for future capital expenditures at leading nodes.

The announcements that FPGA startups Achronix and Tabula are utilizing Intel’s 22nm process technology had some guessing where were Xilinx and Altera. With Netronome, the question could be where is Cavium and the Netlogic RMI group acquired by Broadcom. All attack the data processing path that Intel needs to fill out the networking platform. The acquisition of Fulcrum last summer and QLogic’s Infiniband group provide critical functions that should be able to leverage 22nm at the expense of Broadcom and Marvell’s switch chips and Mellanox Infiniband chips.

As Andy Bechtolsheim, the former Sun founder and Google investor and now running Arista, a startup building low-latency high performance switches, said the era of ASIC based switch chips is over. The inevitable march towards merchant Ethernet silicon is on and who can build the fastest chips accessing the latest technologies wins. The Fulcrum acquisition seems to preclude Broadcom and Marvell from a Foundry slot unless they were to sign away product rights.

Traction by Achronix or Tabula could force Xilinx or Altera to seek an entry into the 22nm trigate process. Until now, both Xilinx and Altera have walled off the FPGA market to startups with their software tools, leading edge processes and robust IP. However what happens if a startup competitor gets a 3-year process technology advantage. In 2009, Altera beat Xilinx out the door by 12 months with its high end 40nm Stratix IV and ended up crushing them in the communications space, a segment that represents almost half the revenue and the majority of the profits. You have to wonder if there is any reason that they aren’t both running test wafers at Intel.

Diminishing nVidia and AMD’s stature in the PC and tablet business; pulling away a Xilinx or Altera; outrunning Broadcom and Marvell in the switch chip market all seem to be part of an overriding strategy that has yet to be communicated by Intel but is a factor in their massive capital expenditure that looks to double capacity by the end of 2013 and put some distance between them and the Foundries. If Intel out executes on the process side, then many fabless vendors may be presented with a Fait Accompli.


FULL DISCLOSURE: I am long INTC, AAPL, ALTR, QCOM

Posted on

NVIDIA Claims TSMC 20nm will not Scale?

NVIDIA Claims TSMC 20nm will not Scale?
by Daniel Nenni on 03-25-2012 at 6:00 pm
Categories: Foundries, TSMC

Interesting article from Joel Hruska on ExtremeTech: Nvidia deeply unhappy with TSMC, claims 22nm essentially worthless . The title is a bit dramatic (poetic license) but the charts are accurate to the degree that 20nm costs will be significantly higher from the start and will continue to be higher throughout production and maturity.

One reason is double patterning, which is required at 20nm (double patterning splits a design into separate masks when devices are too close together). If 28nm masks cost $3M, 20nm masks will cost $5M or more. Another reason is design complexity: layout dependent effects (LDE) and process variation will get worse at 20nm, simulation and verification requirements will explode, and the list goes on…

Life in the semiconductor ecosystem certainly gets more complicated when you try and cram 30B+ planar transistors on one chip, absolutely!

There is no doubt in my mind that these slides are authentic. NVIDIA CEO Jen-Hsun is quite the showman and has a reputation for this kind of public grandstanding. I was a little surprised that NVIDIA used some sort of “normalized” $$$/transistor as a key metric, rather than more traditional semiconductor scaling measures, but perhaps that’s part of their internal business model.

One thing you must know, Jen-Hsun Huang and Morris Chang are VERY close, so you have to play a game of chess here and ask yourself what the agenda really is. A public whipping such as that? Openly criticizing your good friend and founding business partner?

Check out this video at the Silicon Valley Computer Museum, Jen-Hsun interviews Morris, it starts at 8:30 minutes in:

This video will give you a good understanding of the relationship between Jen-Hsun and Morris. It also provides a candid view of Jen-Hsun’s personality, which seems somewhat narcissistic to me.

This topic is certainly timely since the annual TSMC 2012 Technical Symposiumwill take place on April 17[SUP]th[/SUP] at the San Jose Convention Center. Trust me, you’re not going to want to miss this one!

Join the 18th annual TSMC Technology Symposium and get first-hand updates on TSMC’s advanced and specialty technologies, advanced backend capabilities and future development plans!!

  • TSMC’s 20nm and 14nm process development status including FinFet and advanced lithography insights
  • TSMC’s New High-Speed Computing, Mobile Communications, and Connectivity & Storage technology development
  • TSMC’s robust Specialty Technology portfolio that includes Backside Illumination (BSI), Embedded Flash Power IC and MEMS
  • TSMC’s new and exciting GIGAFAB™ new programs and improvements that enhance time-to-volume
  • TSMC’s advanced backend technology for 3D-IC, CoWoS (Chip on Wafer on Substrate), and Bump on Trace (BOT)

To me this is a simple case of wafer pricing negotiations gone wild. Since TSMC is in such a dominant position at 28nm and 20nm the industry is telling them, in the Year of the Dragon, to be more like a teddy bear and collaborate on pricing in the same way they collaborate on technology.

But it’s ridiculous for any fabless semiconductor company to say that they will not aggressively transition to the coming process nodes (20nm and 14nm)! I remember back in the day buying an AMD 40MHZ based PC versus an Intel 36MHZ. Seriously, did I really need that extra 4MHZ? Did I pay extra for it? Of course!

At AMD’s Financial Analyst Day, CEO Rory Read made a point of saying that the company no longer intends to aggressively transition to new process nodes given the diminishing marginal returns from doing so.…….

Okay, let me clarify, any leading edge fabless semiconductor company that wants to STAY in business will aggressively transition to the coming semiconductor nodes, absolutely. Just my opinion of course.

Posted on

TSMC absolutely did NOT halt 28nm production!

TSMC absolutely did NOT halt 28nm production!
by Daniel Nenni on 03-07-2012 at 6:18 pm
Categories: Foundries, TSMC

Once again industry professionals get duped! Tabloid journalism runs amok inside the semiconductor ecosystem! As if our industry does not face enough challenges, why are we wasting time on drivel like this? This is a TSMC 28nm wafer by the way and thousands of them are being shipped around the world, believe it.
 Continue reading “TSMC absolutely did NOT halt 28nm production!”

Posted on

Atrenta/TSMC Soft-IP Alliance: 10 companies make the grade

Atrenta/TSMC Soft-IP Alliance: 10 companies make the grade
by Paul McLellan on 03-05-2012 at 7:30 am
Categories: EDA, Foundries, TSMC

Last May, Atrenta and TSMC announced the Soft-IP Alliance Program which uses SpyGlass and a subset of its GuideWare reference methodology to implement TSMC’s IP quality assessment program. TSMC requires all soft-IP providers to reach a minimum level of completeness before their IP is listed on TSMC online. Since TSMC is so dominant in the foundry business right now (Global struggling with process, Intel talking the talk but not yet really walking the walk, UMC…whatever happened to them anyway?) getting approved and listed with TSMC is extremely important.

Atrenta put everything needed to meet TSMC’s requirements in an IP Handoff Kit. Under the hood this uses SpyGlass’s RTL analysis suite to check for syntax and semantic correctness, simulation-synthesis mismatches, connectivity rules, clock domain crossings, test coverage, timing constraints and…lots more.

Suk Lee of TSMC (my successor at running IC marketing when we were both at Cadence) sees this as measurably improving IP quality. Of course TSMC isn’t directly responsible for IP quality but if IP fails and chips don’t go into production TSMC don’t make any money. Anyway, ten companies have now jumped through all the hoops and qualified their IP for inclusion in the TSMC 9000 IP library.

The companies in this initial program are a veritable who’s who of the IP world (with the notable exceptions of ARM and Synopsys). In alphabetical order so as not to offend anyone:

  • Arteris (NoC)
  • CEVA (DSP cores)
  • Chips&Media (video IP)
  • Digital Media Professionals (graphics IP)
  • Imagination Technologies (GPU cores)
  • Intrinsic-ID (security IP)
  • MIPS Technologies (CPU cores)
  • Sonics (NoC)
  • Tensilica (reconfigurable processors and cores)
  • Vivante (GPU cores)

Now that the dominant way to build an SoC is through assembling IP, the issue of IP quality is is a huge problem and a mixture of tools, methodologies, standards and certification is for sure the way to address it.

Posted on

TSMC 28nm Yield Explained!

TSMC 28nm Yield Explained!
by Daniel Nenni on 03-04-2012 at 4:00 pm
Categories: Foundries, TSMC


Yield, no topic is more important to the semiconductor ecosystem. After spending a significant part of my career on Design for Manufacturability (DFM) and Design for Yield (DFY), I’m seriously offended when semiconductor professionals make false and misleading statements that negatively affects the industry that supports us.
 Continue reading “TSMC 28nm Yield Explained!”

Posted on

The Qualcomm PUT and The FABulous Year Ahead

The Qualcomm PUT and The FABulous Year Ahead
by Ed McKernan on 01-19-2012 at 5:14 pm
Categories: Foundries, Intel Foundry, TSMC

Humor can arise in surprising ways and yet still be disguised to many. As I was researching Qualcomm the other day, I came upon the transcript of their last quarterly earnings and I had to laugh. In the midst of last summer’s European crises, when the Club Med (Greece, Italy, Spain and Portugal) Sovereign Debt was trying to be rolled over with few takers and stocks swooned across the globe, there was Qualcomm injecting a little humor into the markets. You see, in the midst of everyone selling, did a very bullish thing: they sold over $500M of PUTS and collected $75M doing so. At the last earnings call, none of the analysts inquired. Why does a company with $21B in the bank sell PUTS to earn $75M? The answer, I believe has to do with Qualcomm looking to raise its profile even further as they separate themselves from the rest of the mobile ARM camp. However, 2012 will require an even bigger bet the company move.

Throughout the 1990s, Alan Greenspan, disciple of Ayn Rand, and perhaps the most knowledgeable person on the planet in terms of the gives and takes of the economy liked to befuddle congress with presentations that were incomprehensible with the result being that he had great degrees of freedom in pursuing a monetary policy that he believed generated optimum economic growth with low inflation and a stock market that for the most part headed north. An economy, though is a complex thing and even though he thought he could calibrate the health and dynamism by monitoring things like weekly cardboard box production, outside forces could take the stock market down (e.g. the Asian Crises in 1997 and LTCM in 1998). When a crises occurred he would swing into action by implementing what became known as “The Greenspan PUT”. The Fed would immediately lower interest rates and stocks would head higher to the cheers of the investment class. The PUT has limitations, as is seen in our current crises, when debt loads get too large.

Qualcomm’s PUT, in the midst of the European crises was a signal to Wall St. that they believe very good times lie ahead regardless of whether Italy, Greece, France or the whole EU goes in the tank. Although market indications lately show that Europeans will forsake everything to get their hands on an Apple iPAD and iPhone. Perhaps even Italian three hour lunches. Not to worry the money printing has already begun. So, Qualcomm’s selling of PUTS in Q3 was a bullish signal that at the time of expiration in 2H 2012, the stock will be higher than the strike price (I am guessing between $45 and $50 as Qualcomm’s stock flirted with a low of $46). Qualcomm said the breakeven price of the PUT option is roughly $43 a share. Below that they have to write a check to buy the stock back.

Qualcomm’s $21B stash of cash is greater than Intel’s and will soon be three times that of nVidia, Broadcom and Marvell combined who make up the ARM camp that not only are Qualcomm’s chief competitors but licensees. Qualcomm stands as a tall Redwood in a forest of seedlings as it is more than an order of magnitude larger in sales than any of the current ARM campers. But there are major business decisions coming down the pike.

Qualcomm, along with Intel and Apple have the most direct impact on the shaping of the smartphone, tablet, ultrabook mobile Tsunami marketplace and yet each impact it differently. Intel is driving ultrabook to be the form factor that separates them from nVidia and AMD giving them a de-facto Monopoly position in the PC space while also pursuing Apple for the iPhone and iPAD processor business. Apple, we know owns the iTunes Walled Garden Ecosystem that gives them the upper hand in selecting from a cornucopia of suppliers for its next products. You can say that Qualcomm is the winner no matter what communications solution is chosen – whether it is its own chipset or a royalty bearing solution from Broadcom, Intel, Marvell and others. However the big money is in supplying the chips and that can be a problem or opportunity.

As mentioned in previous blogs, the economics of Mobile Tsunami are different than the PC market. Apple and Samsung continue to go Vertical in their supply chain to remove excessive margins. In return for a capital investment and guaranteed demand, Apple gets vendors to drop ASPs and margins. Intel is approaching Apple with a production model that can retain its standard 50-60% Gross Margins but ASPs lower than Samsung due to their 2-3 year process lead. Qualcomm on the other hand sells chips that include the TSMC margin on top of their own 60%+ gross margin.

Bottom line: Does Qualcomm use its $21B cash to build a fab to eliminate TSMC margins and build next generation communications chips that aren’t available elsewhere? Do they approach Intel to Fab next generation standalone chips while offering Intel first rights on “volume integrated communications.” I don’t see Qualcomm moving to a complete IP model. However, the maturation of the very high volume mobile market combined with the economics suggest that the winners will either own Fabs or be IP Houses and a shakeout will take place among the Fabless. There is room for one profit margin, not two – unless you build at Intel. 2012 could be a very decisive year for Qualcomm.

FULL DISCLOSURE: I am Long AAPL, INTC, ALTR, QCOM.

Posted on

Tracking the Big Semiconductor Story of 2012

Tracking the Big Semiconductor Story of 2012
by Ed McKernan on 01-06-2012 at 3:56 pm
Categories: Foundries, Intel Foundry, TSMC

It’s just a matter of time – perhaps just a few months – before the greatest mystery of the semiconductor industry is revealed and the peaceful co-existence of the Fab vs Fabless world is blown apart. An arms race was started by Intel to challenge TSMC and Samsung on who would control not only the high valued processor but soon to be much higher valued NAND Flash and SRAM that define mobility and data center peak performance efficiency. It is not, as many may claim, a battle between x86 and ARM architectures, it is rather an SRAM – NAND dominant memory based platform battle. The shift in our understanding will help explain the mystery as to why Intel is doubling Fab capacity starting in 2012 with the 22nm process even though the PC and server markets are expected to grow in low double digits.

I have written previously about the $10.8B capex build that Intel just completed in 2011. During these past 12 months at every earning call or analyst conference, Intel executives have been cross checked on the wisdom of the massive build out and whether there would be delays, slowdowns, cancellations, pushouts etc… Otellini and company stuck to the strategy and to add a little gasoline to the fire decided to bedevil the analysts with dividend hikes and massive stock buybacks, which will accelerate earnings in 2012 and beyond. This is like FDR telling the war department to get cranking on 50 more aircraft carriers in 1944-45 after having built 83 in the two years prior even though the remaining Japanese carriers numbered in the single digits and the Germans had none to begin with.

It finally occurred to me over the holidays as I was reading the estimates of the Ivy Bridge die size at roughly 25% smaller than the 32nm Sandy Bridge die that Intel had much bigger plans in store for 2012. Ivy Bridge is a shrink of Sandy Bridge with a beefed up graphics controller in order to support DirectX 11 but unlike previous shrinks – there are no additional processor cores. And yet Intel has rushed the conversion of three fabs to 22nm, added a 4[SUP]th[/SUP] fab for 22nm, broke ground on a 14nm development fab that is 60% larger than the previous in order to ramp production immediately and then to top it off decided to build a new 14nm production Fab for late 2013 production ramp. According to Intel, total fab square footage by end of 2013 will be twice what it is today. There are three possible markets, beyond what Intel is building for today that will take advantage of this capacity – and I believe all three will see the light of day in 2012.

The first and obvious one is that Apple will be moving into Intel’s Fabs starting sometime in 2012. Apple’s incredible growth planned for 2012 based on roughly 300M CPUs (Closing in on the number of x86 chips Intel ships) can only be stopped by the words “Lines Down.” The Japanese earthquake and Thailand Floods have exposed the fallacy of the taut just-in-time manufacturing philosophy. No amount of build-ahead stocking can be as effective or economical to Apple as having multiple Fabs building the same product worldwide with guarantees of upside to replace any loss due to “Acts of God” or the imposition of trade barriers that are in the thoughts of many politicians. Furthermore, one can not overlook the trashing of the US dollar that has taken place during the Bush-Obama years that works to the favor of companies who have production facilities in the US (thus Samsung’s new Fab in Texas). Ensconced in Taiwan, TSMC’s business model is at risk to being uncompetitive if the dollar continues to decline. Intel, on the other hand is way ahead of the curve with Fabs in Oregon, Arizona, New Mexico, Ireland and Israel. And when all the dust settles, will probably pickup a certain fab in Dresden at fire sale prices in order to setup a very public cross Atlantic Ocean auction to decide who will be home for the world’s first 450mm wafer. Unless America has a Sputnik Moment, my bet is on the region with the better Bureaucrats.

Tim Cook’s master logicians have implemented the world’s first Virtual Vertical Manufacturing Operation with relatively little money down but first dibs on the latest technologies at Costco prices. They will not have the same leverage with Intel but they will gain access to the process technology that Qualcomm, nVidia, Broadcom and the rest will not have. So Intel can still make 60% Gross Margins while Apple undercuts their rivals in die size, packaging size, power and cost. 22nm is the process that is sending fears through the fabless community (as Morris Chang recently noted) as Intel has reduced standby power by 10X, meaning ARM’s mobile advantage is gone. Though they operate on different game plans, Intel and Apple have a common goal of knocking Samsung out of the game in order to dominate their respective markets and so they will become Allies in this endeavor. Longer term, Apple will need Intel’s process technology in order to smother Amazon’s tablet effort before it gains steam.

Micron’s latest earnings conference call can be described in short as toss the DRAM overboard, it’s Game-On in the battle for NAND Flash Supremacy. The Intel-Micron joint venture looks to leverage Intel’s process lead with Micron’s low overhead business model to win the battle for not only SSDs but also in the die stacked x86CPU+Chipset+NAND configuration that will see uplift in the ultrabook market in 2013. Intel will drive ultrabooks in ever smaller, thinner formfactors while raising their semiconductor content to the exclusion of AMD and nVIdia. If Samsung wants a piece of the action, then they may need to acquire AMD and head down the same road. The market dynamics here offer tremendous upside to Intel and will be the 2[SUP]nd[/SUP] area of Fab capacity utilization.

The third market segment that Intel will attack with its new fab capacity is one that is so logical that I can’t believe it is not being discussed more in the semiconductor techno-sphere. And yet to understand it you have to make a connection to the business opportunity that has been discussed in bits and pieces by Otellini and Data Center VP Kirk Skaugen. Back in the first half of 2012, Skaugen mentioned that Intel’s highest end Xeon processor that sells for over $4000 offered the best ROI to data center customers. The payback they saw in using high end Xeons was in months not years. Furthermore, Intel has raised prices on Xeon and not seen push back, meaning prices will continue to march higher. The ROI is driven by the large Level 3 (L3) caches that allow workloads to stay on chip instead of going to DRAM. So the logical conclusion is that Intel should build even bigger caches and charge more money. However the current Xeons are maxed out on die size.

This is where I see Intel building L4 caches stacked with the Xeon die and offering a new level of performance/watt that can be priced hundreds to thousands of dollars more than the current Xeons. What’s more this new business opportunity allows Intel to ramp 14nm sooner (as they have stated is their goal) with an SRAM product that serves as the process pipe cleaner and is a natural as a much faster yield ramp than any processor Intel builds. The new L4 Cache can eat up lots of wafers, effectively that generate 80-90% gross margins and become a new multi-billion $ business.

In just two short years, as 14nm begins to ramp, Intel will be completely transformed as a company. The x86 vs ARM battle will not be fought as the analysts expect. Dominance in the semiconductor business will have been fought over the superiority of the closely coupled CPU – SRAM – NAND platform of which the latter two will play a greater roll in the value of a platform and end up consuming the most die area and wafers in the Fab. The CPU instruction set and architecture are small potatoes in the overall platform.

Of the three contenders, Intel is the best positioned with process technology and current market standing to move mobiles and servers over to the new memory dominating platforms. Samsung is fighting a two front war with Apple and Intel. TSMC is behind in process technology and SRAM design and does not have a play in NAND – perhaps we will see a partnership form with a leading NAND player.

And so now we can sit back and watch how the biggest semiconductor story of 2012 unfolds as Intel launches its 22nm process.

FULL DISCLOSURE: I am Long AAPL, INTC, ALTR, QCOM

Posted on

GlobalFoundries Versus Samsung!

GlobalFoundries Versus Samsung!
by Daniel Nenni on 11-27-2011 at 7:00 pm
Categories: Foundries, GlobalFoundries, TSMC

Some call it co-opetition (collaborative competition), some call it keeping your enemies close. Others call it for what it is, unfair competition and/or other types of legally actionable behavior. GlobalFoundries calls it“Fab Syncing”, which in reality will SINK their FABS!

“With this new collaboration, we are making one of the industry’s strongest manufacturing partnerships even stronger, while giving customers another platform to drive innovation in mobile technology. Customers using this new offering will gain accelerated time to volume production and assurance of supply, and they will be able to leverage significant learning from the foundry industry’s first high-volume ramp of HKMG technology at 32nm in H1 2011,” said Jim Kupec, senior vice president of worldwide sales and marketing at Globalfoundries.

Unfortunately Jim Kupec no longer works for GlobalFoundries and Samsung may be one of the reasons why. In 2010 Globalfoundries and Samsung Electronics said they would synchronize global semiconductor fabrication facilities to produce chips based on a gate-first implementation of 28nm HKMG technology. They will do the same at 20nm switching to Gate-last HKMG. As a result, Globalfoundries and Samsung will be able to make 28nm and 20nm chips for the SAME customers?!?!?!? Putting aside the gritty technical details, what this means is that GFI will have to compete not only with superpower TSMC, but also their PARTNER Samsung. Samsung is not only the second largest semiconductor company, Samsung is also one of the most fiercely competitive companies in the world. Is that really a good idea?

As it turns out it was a very bad idea for a number of reasons. First and foremost is yield. Samsung is the only “Fab Syncing” partner yielding at 28nm Gate-First HKMG (IBM and GFI are not). Remember Samsung is the largest memory maker so they know how to ramp yield quickly at any node. Are they sharing that manufacturing expertise with GFI and other Common Platform members? Not now, not ever. Samsung is aggressively targeting TSMC and GFI 28nm top customers including AMD, Nvidia, Qualcomm, Broadcom, Marvell, and Xilinx.

Cost and delivery are the key components of a wafer manufacturing contract and Samsung is an expert in both areas. Especially since margins for the Samsung foundry business are not broken out so they could literally dump wafers to get market share. TSMC on the other hand has the biggest wafer margins in the industry which they could cut in half and still make money.

The Samsung cut throat culture is inside the company as well. Multiple Samsung groups compete for a given market. Samsung has phones and tablets based on Nvidia, Qualcomm, and TI processors as well as having their own ARM based processors. They compete in the same way with their largest customer Apple. Apple will purchase close to eightBILLION dollars in parts from Samsung for the iSeries of products this year alone, making Apple Samsung’s largest customer. Samsung is also Apple’s largest competitor and now they are engaged in a mega legal battle which will literally change the face of consumer electronics, believe it.

Even the marketing guys are mixing it up with Samsung firing the first shot:

I’m looking forward to Apple’s response and the Samsung response to that etc…

Let’s not forget the Samsung corruption scandalthat engulfed the government of South Korea. Let’s not forget the chip dumping probes. The book “Think Samsung” by ex-Samsung legal counsel accuses Samsung of being the most corrupt company in Asia.

This battle will be bloody entertaining to say the least! Not so much for GFI though, or the other second source foundries as they see already thinning margins get thinner. For us consumers however it means two things: Semiconductor manufacturing innovation and CHEAP CHIPS! w00t!