Posts Tagged ‘GlobalFoundries’

India’s Semiconductor Push: More Smoke Than Fire

Wednesday, February 8th, 2023

India has been trying to get into semiconductor fabrication for a while now, and after announcing a $10 billion investment fund, and with China locked out of so much semiconductor technology, there have been a lot of news bubbling up, but I want to focus on the Foxconn/Vedanta fab project.

The Economic Times is reporting that Foxconn and Vedanta are seeking to bring in European chipmaker STMicroelectronics as their technology partner in their proposed India manufacturing unit. The two companies announced their joint venture February 2021, with Foxconn as lead partner. Vedanta are reportedly seeking to onboard a CXO to head their semiconductor business.

Snip.

Vedanta-Foxconn are set to finalize a location for their facility in the next few weeks. The consortium are reportedly seeking a 800-1000 acre land parcel that is also well connected with Ahmedabad. The Gujarat government, as of media reports on September 16, showing sites at Sanand and Mandal-Becharaji in Ahmedabad district, two locations near Vadodara in central Gujarat, Dholera, Himmatnagar, Jamnagar, and Kutch. The plant has to be located at a distance from national and state highways so to cut off any vibration from heavy traffic movement. Further, no other major industry should be located in its vicinity.

Vedanta and Foxconn, in a 60-40 joint venture, will be setting up India’s first semiconductor production plant, a display fab unit, and a semiconductor assembling and testing unit over 1000 acres in Ahmedabad, state of Gujarat. The plant will begin production in two years as Foxconn plays the role of technical partner while Vedanta provides financial backing. The investment is worth over INR 1.54 trillion (approx. US$20 billion) and semiconductor manufacturing will be carried out by the holding company, Volcan Investments Limited.

Foxconn is a serious tech player that has serious mastery over the value-added chain. $20 billion, assuming it actually materializes, is real money, even in semiconductors. It’s right around the threshold to build a state of the art sub-10nm fab, even though it’s apparent that that’s not what they’re aiming for.

Vedanta, on the other hand, is another matter. They’re “a globally diversified natural resources company. We extract and process minerals, oil and gas.” Yeah, a natural resources company generally isn’t who you want running your fabs. Another strike is their talking about “Net Zero Carbon by 2050,” which suggests they may have their fingers in political scam pies.

STMicroelectronics is a real chipmaker that runs real fabs, but not the first company I would turn toward to purchase cutting edge process technology from, nor even the tenth. The fact that STM has already announced plans to team up with Global Foundries to build a new 300mm fab next to their existing fab line in Crolles, France in June 2022 makes me even more suspicious. Information on that existing 300mm Crolles fab is sketchy, and I know that for a long time it was a pilot rather than a production line, and I can see no evidence that it was ever expanded to volume production.

The fact that they plan to set set up a fab, a display fab, and a slice-and-dice packaging facility suggest a certain lack of focus. Flat Panel Display (FPD) fabs use familiar semiconductor steps, but the machines are very different because the substrates are different, and Samsung has huge dedicated display fabs. It’s setting up a modern chip fabrication plant that’s the difficult part, and while this combination could probably put together a solid trailing edge fab, like Bosch’s new 65nm fab. But that only cost $1.2B. Maybe they plan to build something in the 20-10nm range.

“The plant will begin production in two years.” Yeah, that’s not happening. Even giant players like TSMC and Intel generally take 2.5-3 years to stand up a new fab from breaking ground to starting up the line.

This could still happen, but the details are very sketchy. The slice and dice operation could be set up without too much difficulty, but it’s a low volume, low tech spinoff operation. A display fab would be more difficult, but it’s doable, though again, probably not in two years. But a real 300mm wafer, sub-65mn node microchip fabrication plant in India? I don’t see this set of players carrying that off well in three years. Five sounds more realistic, and that’s assuming the deal doesn’t fall apart.

Other India semiconductor plays sound even more nebulous.

Taiwan’s TSMC is also looking to set up a chip-fabricating factory in India, and is currently speaking to various government agencies to check the viability of setting up factory in India. TSMC already has one of its largest offices outside of Taiwan in India in Bengaluru, Karnataka, from where it provides support to its’ existing customers in Asia, Europe and North America and supports and encourages fabless companies in India in design and growth.

Ever since news N Chandrasekaran – chairman of the Tata group, announced that Tata Electronics (TEPL) will set up an Outsourced Semiconductor Assembly and Test (OSAT) facility in India, there has a been a lot of speculation, according to which, TSMC and Tata may enter into a partnership.

Besides TSMC, Powerchip Semiconductor Manufacturing Corporation, a Taiwanese chipmaker, is also in exploratory negotiations with several Indian companies to help establish new chip operations in the country, as per a report by Taipei Times. According to the newspaper, the memory chip maker’s announcement put an end to six months of speculation that it was planning to invest in India to diversify its operations despite Taiwan’s rising geopolitical tensions.

“Speaking to,” “exploring plans,” etc. These are very wishy-washy terms. Powerchip is a memory manufacturer that’s hardly flush with cash. I’m sure TSMC is talking to a lot of countries about fabs, but their newest one is under construction in Arizona.

International Semiconductor Consortium (ISMC), a joint venture between UAE’s Next Orbit Ventures and Israel’s Tower Semiconductor, was supposed to spend $3 billion to get started on a 40-65nm analog fab right about now, but I don’t see signs that’s actually happened. Tower is a real foundry, and the $3 billion pricetag and 4-5 year timeline seems realistic, but I’m not 100% sure they’re still interested in the project after Intel announced plans to buy them about a year ago. And having to bring in Arab petrostate funding for your venture is seldom a sign of strong financial viability.

There’s no reason you can’t build one or more modern fabs in India, but so far no major chip manufacturer has chosen to do so, despite the supposed availability of $10 billion in government subsidies.

Tags:, , , , , , , , , ,
Posted in Economics, Elections, Foreign Policy | 1 Comment »

Ill-Advised Semiconductor Subsidies Pass

Thursday, July 28th, 2022

Semiconductor subsidies passed the Senate and House and now will become law.

The House on Thursday passed the bipartisan Chips and Science Act, which aims to increase domestic production of computer chips to allow the U.S. to become more competitive against China in the global technology market.

The bill passed the House in a 243-187 vote one day after passing the Senate in a 64-33 vote. The legislation now heads to the desk of President Joe Biden.

Biden called the passage of the bill on Thursday “exactly what we need to be doing to grow our economy right now.”

“Today, the House passed a bill that will make cars cheaper, appliances cheaper, and computers cheaper,” Biden said. “It will lower the costs of every day goods. And, it will create high-paying manufacturing jobs across the country and strengthen U.S. leadership in the industries of the future at the same time.”

Twenty-four Republicans voted to pass the measure, despite Republican leadership making a last minute push to discourage GOP lawmakers from supporting the bill. GOP leaders sought to keep the bill from passing after news broke on Wednesday that Senator Joe Manchin (D., W. Va.) had reached a deal with Democratic leaders on a nearly half-a-trillion dollar spending package targeting energy and climate, health care, and increased taxes on the wealthy.

Snip.

The measure includes $39 billion to “build, expand, or modernize domestic facilities and equipment” for semiconductors, $2 billion to specifically manufacture semiconductors and $11 billion for Department of Commerce research and development.

“Research and development” is no doubt going to be a rich conduit of graft to Democratic Party cronies having nothing to do with semiconductors.

For reference, $29 billion is probably just enough to build two state-of-the-art 300mm chip fabrication plants.

As I’ve argued before, the reasoning behind the bill is specious and it won’t result in a single new chip being fabbed in the next two years.

The most recent stats I can find show that the United States has some 47% of the semiconductor market. We (and Taiwan, and South Korea) are kicking China’s ass in semiconductors.

The chips China make are generally either: A.) Cheap, or B.) intended for their internal market. No one sends cutting edge chips to be fabbed in China because they don’t have the tech to do it and everyone know they’ll steal your designs and crank out knock-offs on the sly whenever possible. China’s semiconductor industry is mostly smoke and mirrors all the way down.

Semiconductor subsidies have all the hallmarks of a classic Washington boondoggle: The wrong action at the wrong time for the wrong problem.

First, there are already signs that the automotive semiconductor crunch is easing, thanks not to the Biden Administration but to the actions of the free market.

Second, the shortage wasn’t the result of a “chip shortage,” it was the result of “a lack of available foundry wafer starts.” Automakers cancelled their orders for display drivers when it looked like Flu Manchu lockdowns were going to depress the economy for a while, and were caught off-guard by the V-shaped recovery under Trump, and got sent to the back of the line to get their product fabbed after they changed their mind. Remember, just about all foundries are running flat-out 24/7/365, pausing only to switch to different chips for different customers. There’s no slack in the system, and those wafer starts are already spoken for (and possibly paid for) by other customers well in advance. Just as nine woman can’t give birth to a fully grown baby in one month, you can’t just “make chips quicker” in an existing fab.

Third, remember that cutting edge semiconductor fabs are hideously expensive. Moore’s second law states that the cost of a new, cutting edge semiconductor plant doubles every four years. Samsung’s planned fab in Taylor, Texas is going to cost $17 billion.

Fourth, nothing about these subsidies will address the real problem with American semiconductors, which is that the overwhelming majority of cutting edge chip designs have to flow through TSMC fabs in Taiwan. What will solve that problem is TSMC opening a state-of-the art fab in Arizona in 2024. No amount of U.S. taxpayer money will make that already-under-construction fab start producing chips any quicker.

Could these subsidies boost American semiconductor manufacturing 2-3 years from now? Possibly. Knowing the cycling nature of the industry and the tendency of government subsidies to backfire, new/upgraded fab lines might come online just as the industry is experiencing a glut.

But the real key to restoring America to the cutting edge of semiconductor manufacturing is the already-in-progress inshoring of cutting edge foreign owned fabs from Samsung and TSMC, and having American semiconductor manufacturers like Intel and GlobalFoundries master sub-10nm chip fabrication processes, something they have heretofore been unable to do. (Intel is closer, having been on the cutting edge until they lost their way, while GlobalFoundries stopped all development on their 7nm node because they couldn’t find a way to make the investment pay off.)

Throwing buckets of budget-busting borrowed taxpayer money around isn’t going to make any of those things happen any faster.

Tags:, , , , , , , , ,
Posted in Budget, Democrats, Waste and Fraud, Welfare State | 3 Comments »

Semiconductor Update for July 18, 2022

Monday, July 18th, 2022

Enough links have filtered into the semiconductor bucket to be worth doing a roundup. This one touches on China and the corruption of our political elites.

  • The congressional Democrats’ attempt to throw money at the problem is going nowhere fast.

    The Biden administration is laser-focused on sending Ukraine billions of dollars in weapons, including the latest round of anti-ship systems, artillery rockets, and rounds of 105 mm ammo for howitzer cannons that it has entirely lost focus on reshoring efforts to boost semiconductor production Stateside.

    Multiple manufacturers of semiconductor wafers have announced plans for new multi-billion dollar factories across the U.S. but are contingent on Congress allocating funds to aid in building facilities under the Creating Helpful Incentives to Produce Semiconductors (CHIPS) for America Act.

    Congress passed the CHIPS Act in January 2021 as part of last year’s National Defense Authorization Act, which proposed $52 billion in funding for increasing the domestic capacity of chip production, though the House and Senate have come to a standstill over disagreements on certain parts of the bill that have sparked so much uncertainty among companies set to build new factories.

    In a letter on June 15, dozens of technology executives from IBM, Intel, Microsoft, Analog Devices, Micron, Amazon, and Alphabet called on Congress to move quickly on the CHIPS Act. They wrote, “the rest of the world is not waiting for the U.S. to act,” and funding for new chip factories must be achieved immediately.

    Taiwan’s GlobalWafers announced a new $5 billion factory in the U.S. on Monday, but contingent on subsidies from the federal government.

    “This investment that they’re making is contingent upon Congress passing the CHIPS Act. The [GlobalWafers] CEO told me that herself, and they reiterated that today,” U.S. Commerce Secretary Gina Raimondo told CNBC, the same day GlobalWafers announced its development plan.

    Notes:

    • IBM doesn’t own any fabs any more, having sold them all to GlobalFoundries.
    • Intel runs a huge number of very profitable fabs (troubles with their sub-10nm process yields notwithstanding) and doesn’t need federal subsidies.
    • Microsoft doesn’t own any fabs and is deeply unlikely to build any; their flagship Xbox Series X uses a custom AMD Zen 2 fabbed by TSMC as its CPU.
    • Analog Devices is an Integrated Device Manufacturer that owns several fabs with pretty old technology; they don’t have any 300mm fabs. They closed a small fab in Milpitas they got from their acquisition of Linear Technology last year. Designing analog chips is its own black art, and not everything that applies to shrinking digital circuits applies to the analog realm.
    • Amazon has no fabs and probably won’t be building any, but they do have a chip design division to support Amazon Web Services, and recently designed a cloud computing chip. They work closely with AMD (fabbed at TSMC), Intel (own their own fabs) and Nvidia (another fabless design house that also gets their chips fabbed at TSMC).
    • Alphabet AKA Google has no fabs and probably won’t be building any, though they do have a lot of AI chip design work going on.
    • GlobalWafers isn’t a semiconductor manufacturer, it’s a silicon wafer manufacturer. Making such wafers (the substrates upon which semiconductor fabrication depends) has its own challenges, but they are several orders less difficult than cutting edge chip fabrication. Maybe I’m quite far out of the loop, but I’m deeply suspicious that GlobalWafers planned wafer plant in Sherman, Texas will cost $5 billion. That’s a relatively piddling sum for a new semiconductor fab, but extremely expensive for a wafer factory. This makes me suspect a subsidy grab is afoot.

    So of the companies mentioned, Intel could suck up government funding to build a fab they were going to build anyway, I’m sure Analog Devices would build a fab with government money, but chances of them running an under 10nm process in said theoretical fab is extremely slim, none of the other mentioned copies are going to build a fab, and none of that government money is going to alleviate the main problem that the overwhelming majority of cutting edge chip designs have to flow through TSMC fabs in Taiwan. What will solve that problem is TSMC opening a state-of-the art fab in Arizona in 2024. No amount of U.S. taxpayer money will make that already-under-construction fab start producing chips any quicker.

    As I’ve mentioned previously, semiconductor subsidies are the wrong solution to the wrong problem.

    $250 billion in taxpayer subsidies wouldn’t get you a single additional wafer start this year, and probably would accomplish little more than channeling money to politically connected firms and sticky pockets in a state (New York) that no one wants to build fabs in any more because of high costs, high taxes and union rule requirements.

  • So who expects to earn immediate gains from the taxpayers subsidizing semiconductors? Would you believe Nancy Pelosi?

    I bet you would.

    This past week it hit the terminal that House Speaker Pelosi was doing a little portfolio re-jiggering, including exercising $8 million of call options in Nvidia and selling Apple and Visa calls. The data was per CongressTrading.com and was reported on by Bloomberg.

    The Nvidia LEAPS were bought June 3, 2021 with $100 strikes, set to expire June 17, 2022 and the position appeared to be disclosed on Thursday morning for the first time. $8 million trades seem a little odd for members of Congress to begin with, but who are we to judge?

    But then, what did Speaker Pelosi do just hours after disclosing the trade, on Friday?

    She threw her weight behind a stalled $50 billion CHIPS PLUS bill that “would provide $52 billion in funding for semiconductor manufacturing grants and investment tax credits for the chip industry.”

  • Speaking of TSMC, they’re tired of their customers using their old tech.

    We tend to discuss leading-edge nodes and the most advanced chips made using them, but there are thousands of chip designs developed years ago that are made using what are now mature process technologies that are still widely employed by the industry. On the execution side of matters, those chips still do their jobs as perfectly as the day the first chip was fabbed which is why product manufacturers keep building more and more using them. But on the manufacturing side of matters there’s a hard bottleneck to further growth: all of the capacity for old nodes that will ever be built has been built – and they won’t be building any more.

    Not strictly true. Remember, Bosch just finished building a 65nm fab.

    As a result, TSMC has recently begun strongly encouraging its customers on its oldest (and least dense) nodes to migrate some of their mature designs to its 28 nm-class process technologies.

    Nowadays TSMC earns around 25% of its revenue by making hundreds of millions of chips using 40 nm and larger nodes. For other foundries, the share of revenue earned on mature process technologies is higher: UMC gets 80% of its revenue on 40 nm higher nodes, whereas 81.4% of SMIC’s revenue come from outdated processes.

    That’s because UMC has fallen woefully far behind TSMC, and no one trusts them because they let Chinese spies walk out the door with other company’s IP. SMIC is on Mainland China, sucks even more, and is trusted even less.

    Mature nodes are cheap, have high yields, and offer sufficient performance for simplistic devices like power management ICs (PMICs). But the cheap wafer prices for these nodes comes from the fact that they were once, long ago, leading-edge nodes themselves, and that their construction costs were paid off by the high prices that a cutting-edge process can fetch. Which is to say that there isn’t the profitability (or even the equipment) to build new capacity for such old nodes.

    This is why TSMC’s plan to expand production capacity for mature and specialized nodes by 50% is focused on 28nm-capable fabs. As the final (viable) generation of TSMC’s classic, pre-FinFET manufacturing processes, 28nm is being positioned as the new sweet spot for producing simple, low-cost chips. And, in an effort to consolidate production of these chips around fewer and more widely available/expandable production lines, TSMC would like to get customers using old nodes on to the 28nm generation.

    “We are not currently [expanding capacity for] the 40 nm node” said Kevin Zhang, senior vice president of business development at TSMC. “You build a fab, fab will not come online [until] two year or three years from now. So, you really need to think about where the future product is going, not where the product is today.”

  • This video asks whether China can produce their own chips:

    Obviously, they already produce some of their own chips, but the video covers most of the issues China has with fabbing more complex chips that I’ve already discussed here and here. They’re still dependent on the same three leading fab companies (TSMC, Intel and Samsung) everyone else is for sub 10nm feature chips, and are overwhelmingly dependent on both foreign talent and foreign semiconductor equipment manufacturers like ASML and Applied Materials.

  • Speaking of TSMC and Intel, India would really like them to build fabs there. The problem is, despite a whole lot of technical talent there, it doesn’t have a terribly large domestic electronics manufacturing base.

    • Tags:, , , , , , , , , , , , , ,
    Posted in Democrats, Economics, Elections, Foreign Policy, Texas, video | 9 Comments »

    Q: Can You Double-Pattern Rather Than Use EUV? A: You Don’t Want To

    Sunday, April 10th, 2022

    This is going to be pretty esoteric for many of my readers, but in previous semiconductor posts covering ASML, some commenters have suggested that fabs can do multi-patterning for smaller nodes rather than having to use ASML’s extreme ultraviolet stepper. The following video explains why, below a certain threshold, no, you really can’t.

    I’m not going to summarize every point, but the largest takeaway is that multi-patterning is computationally prohibitive. Double-patterning splits a single mask into two masks, each of which only create half of the mask pattern on the die. Double-patterning was fine for a while, but triple patterning and self-aligned double-patterning start making finding optimal solutions to the mask splitting problem exponentially more difficult.

    Take a square. A square has four nodes in it. With double patterning, each of the two masks handle opposing sides of the square. And with this four-node shape, there are two double patterning options available for coloring. The EDA software thus has to check through them for design rule violations and whatnot. With triple patterning, the number of variations explodes exponentially. For that same square four node structure, triple patterning has 18 variations rather than just two with double patterning. A five node structure, 30. And so on. A semiconductor design can have hundreds of different nodes and design variations. The software needs to check through at least a good portion of these. This problem is not solvable in polynomial time. In other words, for you computer science nerds out there, it is an NP complete problem.

    And then there’s the cost. “Depending on whose cost model you consult, [10nm]’s triple patterning makes its lithography module 3.85x higher than [28nm].” And the non-EUV 7nm node required triple-patterning and something called “self-aligned quadruple patterning.” And on Intel: “Brian Krzanich has said that in certain cases the company needs to use quad (4x), penta (5x), or hexa (6x) patterning for select features, as they need to expose the wafer up to six times to “draw” one feature. I am not super surprised that it wouldn’t yield. No wonder GlobalFoundries ditched their 7nm node.”

    And this summary glosses over big differences between different fab technologies on different companies. TSMC’s 7nm isn’t the same as Intel’s 7nm.

    Anyway, all this goes a long way to explain: Multi-patterning is much more painful than simply ponying up the cost for an ASML EUV stepper. And if you want to do 6nm, you have to use EUV.

    Tags:, , , , , , ,
    Posted in Uncategorized, video | No Comments »

    Semiconductor Subsidies: The Wrong Solution For The Wrong Problem

    Thursday, January 20th, 2022

    There’s no problem that the federal government throwing money at it can’t make worse.

    Today’s example: Democrats pimping billions in taxpayer subsidies for the semiconductor industry.

    As the COVID-19 pandemic exacerbates supply chain backlogs and global computer chip shortages

    Correction: It wasn’t the pandemic itself, it was government lockdowns and other overeactions that did that.

    Democratic leaders in Congress as well as President Joe Biden want Congress to fast track a $250 billion bill to develop American independence from China and other competitors in chip manufacturing.

    The Capital Region – home to SUNY Polytechnic Institute, the only publicly owned 300-millimeter semiconductor research and development center in the U.S. – stands to reap significant benefits from the enactment of Senate Majority Leader Charles E. Schumer’s multi-billion dollar bill, which he envisions as a direct investment in his home state’s economy.

    “Sen. Schumer wrote this legislation with upstate New York always at the forefront of his mind,” Schumer’s spokeswoman Allison Biasotti said. “We are already seeing the excitement in major employer expansions and thousands of jobs on the horizon from GlobalFoundries’ planned expansion (in Malta) and (his) push for Albany Nanotech to be a hub for the National Semiconductor Technology Center.”

    A focal point of the bill, which the New York Democrat co-sponsored with Sen. Todd Young, R-Ind., is a historic $52 billion investment in stateside semiconductor research and development to address a global chip shortage plaguing the automotive industry.

    Lawmakers began to focus more on the low domestic production of semiconductors when the COVID-19 pandemic cut off supplies from overseas. Without access to chips, several automakers shut down their production lines, and manufacturers of essential medical devices and consumer electronics struggled to meet increasing demand.

    Roughly 12 percent of the world’s semiconductors are manufactured in the United States, down from 37 percent in 1990, according to the Semiconductor Industry Association.

    Either these stats are false or misleading (probably the latter). The most recent stats I can find show that the United States has some 47% of the semiconductor market. It’s possible that the 12% refers to the entire worldwide number of individual chips produced, including discrete components (transistors, resistors, etc.). Those are indeed semiconductors, but they’re produced on old amortized fabs (inside the industry these are referred to as “jelly bean factories”) and sell for pennies a piece (or less). If you’re already in that industry, those old fabs make small, steady profits every year, but nobody jumps into that business with new fabs.

    The chips China make are generally either: A.) Cheap, or B.) intended for their internal market. No one sends cutting edge chips to be fabbed in China because they don’t have the tech to do it and everyone know they’ll steal your designs and crank out knock-offs on the sly whenever possible. China’s semiconductor industry is mostly smoke and mirrors all the way down.

    Semiconductor subsidies have all the hallmarks of a classic Washington boondoggle: The wrong action at the wrong time for the wrong problem.

    First, there are already signs that the automotive semiconductor crunch is easing, thanks not to the Biden Administration but to the actions of the free market.

    Second, the shortage wasn’t the result of a “chip shortage,” it was the result of “a lack of available foundry wafer starts.” Automakers cancelled their orders for display drivers when it looked like Flu Manchu lockdowns were going to depress the economy for a while, and were caught off-guard by the V-shaped recovery under Trump, and got sent to the back of the line to get their product fabbed after they changed their mind. Remember, just about all foundries are running flat-out 24/7/365, pausing only to switch to different chips for different customers. There’s no slack in the system, and those wafer starts are already spoken for (and possibly paid for) by other customers well in advance. Just as nine woman can’t give birth to a fully grown baby in one month, you can’t just “make chips quicker” in an existing fab.

    Third, remember that cutting edge semiconductor fabs are hideously expensive. Moore’s second law states that the cost of a new, cutting edge semiconductor plant doubles every four years. Samsung’s planned fab in Taylor, Texas is going to cost $17 billion.

    Fourth, if you go to a random semiconductor company and go “Here’s 20 billion! Go build a state-of-the-art 5nm wafer fabrication plant!”, then:

    A.) You’re looking at a very minimum of 2-3 years before the first production wafer comes off the line. You can’t just take an existing building and turn it into a fab, it has to be specially built from the ground up with exacting standards for cleanroom air filtering, concrete slab level uniformity, etc. And 2-3 years is probably the lead time to get an ASML EUV stepper.

    B.) Unless you’re TSMC, Samsung or (maybe) Intel, the answer is probably “Uh, we’ll try, but no promises,” because those three companies are the only ones that actually having wafer fabs running 10nm or smaller process nodes. GlobalFoundries, mentioned in the article, has Fab 8 in Malta, NY, running 14nm, which is not horribly far off the state-of-the art, but not good enough to fab the really cutting-edge chips demanded of companies like Apple, NVIDIA, etc. Tiny problem: In 2018, GlobalFoundries stopped all work on 7nm development.

    The contract maker of semiconductors decided to cease development of bleeding edge manufacturing technologies and stop all work on its 7LP (7 nm) fabrication processes, which will not be used for any client. Instead, the company will focus on specialized process technologies for clients in emerging high-growth markets. These technologies will initially be based on the company’s 14LPP/12LP platform and will include RF, embedded memory, and low power features.

    So it was too hard a game for them to play, but with a big heap of taxpayer subsidies, I’m sure they’d be willing to give it another go.

    Of course, you don’t need a cutting edge fab to build display drivers. Bosch just opened a $1.2 billion, 65nm fab in Dresden to do just that. But you don’t need subsidies to build trailing edge fabs.

    $250 billion in taxpayer subsidies wouldn’t get you a single additional wafer start this year, and probably would accomplish little more than channeling money to politically connected firms and sticky pockets in a state (New York) that no one wants to build fabs in any more because of high costs, high taxes and union rule requirements.

    It’s a bad idea congress should reject.

    Tags:, , , , , , , , , ,
    Posted in Democrats | 3 Comments »

    Semiconductor Update: GlobalFoundries Gives Up On 7nm​

    Thursday, August 30th, 2018

    GlobalFoundries has given up work on their 7nm process node. This is a direct result of AMD choosing TSMC over GlobalFoundries to fab their next generation microprocessor.

    GlobalFounderies was always something of an odd duck. It was spun out from AMD in 2009 to turn their manufacturing arm into a foundry because AMD itself could no longer afford the huge upfront capital investment state-of-the-art wafer fabrication plants demanded. As it exists today, GlobalFounderies​ is a Frankenstein’s monster of agglomeration, having gobbled up Singapore-based Chartered Semiconductor and what remained of IBM’s fab infrastructure (back in the day, IBM had some of the best semiconductor design capabilities in the world) in New York and Vermont. (SK Hynix, NXP and ON Semiconductor, all integrated device manufacturers rather than foundries, are similar merger-assembled aggregations.) GlobalFounderies actual owner is the Emirate of Abu Dhabi.

    With UMC screwing the pooch by letting Chinese spies walk out the door with Micron design IP, there was an opening for a (sorta, kinda) American chip foundry to provide a viable rival to TSMC, but GlobalFoundries evidently found it too difficult to do profitably.

    TSMC has already broken ground on a fab that will theoretically take them down to 5nm and is expected to cost $500 billion NT, which works out to over $16 billion US at current exchange rates. That’s more outlay than all the profit TSMC made all of last year.

    Some thoughts (partially based on scuttlebutt, gossip, etc.):

  • Right now there’s no non-TSMC foundry choice if a fabless chip company wants to attempt a sub 14nm design. It’s Taiwan or nothing.
  • To the best of my knowledge, no one outside TSMC, Intel and Samsung are even attempting 7nm. Word is that TSMC’s 7nm is actually closer to 10nm, and Intel is evidently in a world of hurt getting yields up on its 10nm process.
  • Samsung says they’re going to 7nm in 2019 using Extreme Ultraviolet (EUV) lithography, a long, long awaited technological shift that will probably involve its own painful learning curve. Others have speculated that, despite those plans, Samsung seems pretty happy sitting at 14nm with high yields for most of its own chip needs (as opposed to its foundry customers).

    • What this means is that the cutting edge of wafer fabrication technology is probably going to be centered on the Pacific rim for the foreseeable future. China won’t be on that cutting edge, because they can’t steal technology fast enough or hire enough enough qualified process techs to get it done.

    We may finally have reached a point that building a cutting edge, state-of-the-art wafer fabrication plant is a money-losing proposition for everyone.

    That means fabless chip designers working at the cutting edge will be dependent on Taiwan and South Korea for the foreseeable future, a fact that has a lot of foreign policy relevance, especially in relation to China…

    Tags:, , , , , , , , , ,
    Posted in Economics, Foreign Policy | 1 Comment »