Category Archives: Device Architecture

China has been the largest consuming country for ICs since 2005, but large increases in IC production within China have not immediately followed, according to data presented in the new 500-page 2019 edition of IC Insights’ McClean Report—A Complete Analysis and Forecast of the Integrated Circuit Industry (released in January 2019).  As shown in Figure 1, IC production in China represented 15.3% of its $155 billion IC market in 2018, up from 12.6% five years earlier in 2013.  Moreover, IC Insights forecasts that this share will increase by 5.2 percentage points from 2018 to 20.5% in 2023.

Figure 1

Currently, China-based IC production is forecast to exhibit a very strong 2018-2023 CAGR of 15%.  However, considering that China-based IC production was only $23.8 billion in 2018, this growth is starting from a relatively small base.  In 2018, SK Hynix, Samsung, Intel, and TSMC were the major foreign IC manufacturers that had significant IC production in China.  In fact, SK Hynix’s 300mm China fab had the most installed capacity of any of its fabs in 2018 at 200,000 wafers per month (full capacity).

Intel’s 300mm fab in Dalian, China (Fab 68 that started MCU production in late October 2010), was idled in 3Q15 as the company switched the fab to 3D NAND flash manufacturing.  This conversion was completed in late 2Q16.  Intel’s China fab had an installed capacity of 70,000 300mm wafers per month in December of 2018 (full capacity).

In early 2012, Samsung gained approval from the South Korean government to construct a 300mm IC fabrication facility to produce NAND flash memory in in Xian, China.  Samsung started construction of the fab in September of 2012 and production began in 2Q14.  The company invested $2.3 billion in the first phase of the fab with $7.0 billion budgeted in total.  This facility was the primary fab for 3D NAND production for Samsung in 2017 with an installed capacity of 100,000 wafers per month as of December 2018 (the company plans to expand this facility to 200,000 wafers per month).

Significant increases in IC sales over the next five years are also expected from existing indigenous Chinese companies including pure-play foundries SMIC and Huahong Group and memory startups YMTC and ChangXin Memory Technologies (CXMT, formerly Innotron). DRAM startup JHICC is currently on hold pending the sanctions imposed on the company by the U.S.  Moreover, there are likely to be new companies looking to establish IC production in China like Taiwan-based Foxconn, which announced in December of 2018 that it intended to build a $9.0 billion fab in China to offer foundry services as well as produce TV chipsets and image sensors.

If China-based IC production rises to $47.0 billion in 2023 as IC Insights forecasts, it would still represent only 8.2% of the total forecasted 2023 worldwide IC market of $571.4 billion.  Even after adding a significant “markup” to some of the Chinese producers’ IC sales figures (since many of the Chinese IC producers are foundries that sell their ICs to companies that re-sell these products to the electronic system producers), China-based IC production would still likely represent only about 10% of the global IC market in 2023.

Even with new IC production being established by China-based startups such as YMTC and CXMT, IC Insights believes that foreign companies will continue to be a large part of the IC production base in China.  As a result, IC Insights forecasts that at least 50% of IC production in China in 2023 will come from foreign companies with fabs in China such as SK Hynix, Samsung, Intel, TSMC, UMC, GlobalFoundries, and Foxconn.

Given the sheer size of China’s investment plans over the next five years, it is likely that China will achieve some level of success with their strategy to become less reliant on IC imports.  However, given increased government scrutiny of Chinese attempts at purchasing foreign technology companies and the legal challenges that the Chinese startups are likely to face in the future, IC Insights believes that China’s current strategy with regard to the IC industry will fall far short of the level of success that China’s government has targeted with its “Made in China 2025” plan (i.e., 40% self-sufficiency by 2020 and 70% by 2025).

Governor Andrew M. Cuomo today announced that IBM (NYSE: IBM), a long-time anchor tenant at the SUNY Polytechnic Institute campus in Albany, plans to invest over $2 billion to grow its high-tech footprint at the campus and throughout New York State. This includes the establishment of an “AI Hardware Center” at SUNY Poly for artificial intelligence-focused computer chip research, development, prototyping, testing and simulation. Once established, the AI Hardware Center will be the nucleus of a new ecosystem of research and commercial partners, and further solidify the Capital Region’s position as “Tech Valley” – a global hub for innovative research and development.

New York has always been at the forefront of emerging industries, and this private sector investment to create a hub for artificial intelligence research will attract world-class minds and drive economic growth in the region,” Governor Cuomo said. “Artificial intelligence has the potential to transform how we live and how businesses operate, and this partnership with IBM will help ensure New York continues to be on the cutting edge developing innovative technologies.”

“This investment by IBM will continue to grow New York’s high-tech industry in the Capital Region and across the state,” said Lieutenant Governor Kathy Hochul. “The artificial intelligence hardware center will expand research and partnerships at SUNY Polytechnic Institute, and ensure Tech Valley attracts innovative business and development that drives economic development in the region.”

IBM’s expected $2 billion investment will be made at SUNY Poly and other IBM facilities in New York State. IBM plans to provide at least $30 million in cash and in-kind contributions for artificial intelligence research across the SUNY system, with SUNY matching up to $25 million for a combined total of $55 million. Empire State Development will provide a $300 million capital grant over five years, to the Research Foundation for SUNY to purchase, own and install tools necessary to support the AI Hardware Center.

IBM also plans to expand and extend its partnership with SUNY Poly for the Center for Semiconductor Research (CSR), which is set to expire at the end of 2021, through at least 2023, with an option to extend the CSR for an additional five years through 2028.

The AI Hardware Center will attract new AI industry companies and federal research to the state, while fostering economic development and working to create several hundred new jobs and retain hundreds of other existing jobs at the SUNY Poly campus and at IBM’s and its collaborators’ facilities.

By bombarding an ultrathin semiconductor sandwich with powerful laser pulses, physicists at the University of California, Riverside, have created the first “electron liquid” at room temperature.

The achievement opens a pathway for development of the first practical and efficient devices to generate and detect light at terahertz wavelengths — between infrared light and microwaves. Such devices could be used in applications as diverse as communications in outer space, cancer detection, and scanning for concealed weapons.

The research could also enable exploration of the basic physics of matter at infinitesimally small scales and help usher in an era of quantum metamaterials, whose structures are engineered at atomic dimensions.

The UCR physicists published their findings online Feb. 4 in the journal Nature Photonics. They were led by Associate Professor of Physics Nathaniel Gabor, who directs the UCR Quantum Materials Optoelectronics Lab. Other co-authors were lab members Trevor Arp and Dennis Pleskot, and Associate Professor of Physics and Astronomy Vivek Aji.

A video depicting the research is available here.

In their experiments, the scientists constructed an ultrathin sandwich of the semiconductor molybdenum ditelluride between layers of carbon graphene. The layered structure was just slightly thicker than the width of a single DNA molecule. They then bombarded the material with superfast laser pulses, measured in quadrillionths of a second.

“Normally, with such semiconductors as silicon, laser excitation creates electrons and their positively charged holes that diffuse and drift around in the material, which is how you define a gas,” Gabor said. However, in their experiments, the researchers detected evidence of condensation into the equivalent of a liquid. Such a liquid would have properties resembling common liquids such as water, except that it would consist, not of molecules, but of electrons and holes within the semiconductor.

“We were turning up the amount of energy being dumped into the system, and we saw nothing, nothing, nothing — then suddenly we saw the formation of what we called an ‘anomalous photocurrent ring’ in the material,” Gabor said. “We realized it was a liquid because it grew like a droplet, rather than behaving like a gas.”

“What really surprised us, though, was that it happened at room temperature,” he said. “Previously, researchers who had created such electron-hole liquids had only been able to do so at temperatures colder than even in deep space.”

The electronic properties of such droplets would enable development of optoelectronic devices that operate with unprecedented efficiency in the terahertz region of the spectrum, Gabor said. Terahertz wavelengths are longer than infrared waves but shorter than microwaves, and there has existed a “terahertz gap” in the technology for utilizing such waves. Terahertz waves could be used to detect skin cancers and dental cavities because of their limited penetration and ability to resolve density differences. Similarly, the waves could be used to detect defects in products such as drug tablets and to discover weapons concealed beneath clothing.

Terahertz transmitters and receivers could also be used for faster communication systems in outer space. And, the electron-hole liquid could be the basis for quantum computers, which offer the potential to be far smaller than silicon-based circuitry now in use, Gabor said.

More generally, Gabor said, the technology used in his laboratory could be the basis for engineering “quantum metamaterials,” with atom-scale dimensions that enable precise manipulation of electrons to cause them to behave in new ways.

In further studies of the electron-hole “nanopuddles,” the scientists will explore their liquid properties such as surface tension.

“Right now, we don’t have any idea how liquidy this liquid is, and it would be important to find out,” Gabor said.

Gabor also plans to use the technology to explore basic physical phenomena. For example, cooling the electron-hole liquid to ultra-low temperatures could cause it to transform into a “quantum fluid” with exotic physical properties that could reveal new fundamental principles of matter.

In their experiments, the researchers used two key technologies. To construct the ultrathin sandwiches of molybdenum ditelluride and carbon graphene, they used a technique called “elastic stamping.” In this method, a sticky polymer film is used to pick up and stack atom-thick layers of graphene and semiconductor.

And to both pump energy into the semiconductor sandwich and image the effects, they used “multi-parameter dynamic photoresponse microscopy” developed by Gabor and Arp. In this technique, beams of ultrafast laser pulses are manipulated to scan a sample to optically map the current generated.

The Semiconductor Industry Association (SIA), representing U.S. leadership in semiconductor manufacturing, design, and research, today announced the global semiconductor industry posted sales of $468.8 billion in 2018, the industry’s highest-ever annual total and an increase of 13.7 percent compared to the 2017 total. Global sales for the month of December 2018 reached $38.2 billion, a slight increase of 0.6 percent over the December 2017 total, but down 7.0 percent compared to the total from November 2018. Fourth-quarter sales of $114.7 billion were 0.6 percent higher than the total from the fourth quarter of 2017, but 8.2 percent less than the third quarter of 2018. All monthly sales numbers are compiled by the World Semiconductor Trade Statistics (WSTS) organization and represent a three-month moving average.

“Global demand for semiconductors reached a new high in 2018, with annual sales hitting a high-water mark and total units shipped topping 1 trillion for the first time,” said John Neuffer, SIA president and CEO. “Market growth slowed during the second half of 2018, but the long-term outlook remains strong. Semiconductors continue to make the world around us smarter and more connected, and a range of budding technologies – artificial intelligence, virtual reality, the Internet of Things, among many others – hold tremendous promise for future growth.”

Several semiconductor product segments stood out in 2018. Memory was the largest semiconductor category by sales with $158.0 billion in 2018, and the fastest growing, with sales increasing 27.4 percent. Within the memory category, sales of DRAM products increased 36.4 percent and sales of NAND flash products increased 14.8 percent. Logic ($109.3 billion) and micro-ICs ($67.2 billion) – a category that includes microprocessors – rounded out the top three product categories in terms of total sales. Other fast-growing product categories in 2018 included power transistors (14.4 percent growth/total sales of $14.4 billion) and analog products (10.8 percent growth/total sales of $58.8 billion). Even without sales of memory products, sales of all other products combined increased by nearly 8 percent in 2018.

Annual sales increased substantially across all regions: China (20.5 percent), the Americas (16.4 percent), Europe (12.1 percent), Japan (9.2 percent), and Asia Pacific/All Other (6.1 percent). For the month of December 2018, year-to-year sales increased in China (5.8 percent), Europe (2.8 percent), and Japan (2.3 percent), but fell in Asia Pacific/All Other (-0.7 percent) and the Americas (-6.2 percent). Sales in December 2018 were down compared to November 2018 across all regions: Japan (-2.2 percent), Asia Pacific/All Other (-3.1 percent), Europe (-4.9 percent), China (-8.1 percent), and the Americas (-12.4 percent).

“A strong semiconductor industry is critical to America’s economic strength, national security, and global technology leadership,” said Neuffer. “We urge Congress and the Trump Administration to enact polices in 2019 that promote continued growth and innovation, including robust investments for basic scientific research, long-overdue high-skilled immigration reforms, and initiatives that promote free and open trade, such as the U.S.-Mexico-Canada Agreement (USMCA). We look forward to working with policymakers in the year ahead to further strengthen the semiconductor industry, the broader tech sector, and our economy.”

For comprehensive monthly semiconductor sales data and detailed WSTS Forecasts, consider purchasing the WSTS Subscription Package. For detailed historical information about the global semiconductor industry and market, consider ordering the SIA Databook.

KLA Corporation (NASDAQ: KLAC) today announced the appointment of Victor Peng to its board of directors.

Mr. Peng is president and chief executive officer of Xilinx, where he also serves as a member of the board of directors. Peng has over 30 years of experience leading technology transformation, defining and bringing to market FPGAs, SoCs, GPUs, high performance microprocessors and chip sets, and microprocessor IP products.

“We’re excited to have Victor Peng join the KLA board,” commented Edward W. (Ned) Barnholt, chairman of the board of KLA Corporation. “Victor is an accomplished semiconductor industry leader who brings significant business knowledge, technical expertise, and operational experience that will be invaluable to KLA as we execute our strategies for growth and market leadership.”

Peng joined Xilinx in 2008 and was named president and CEO of the company in January of 2018. Previously he was Xilinx’s chief operating officer, with responsibility for global sales, global operations and quality, product development, and product and vertical marketing. Prior to that, he served as the company’s executive vice president and general manager of Products, a position he held since July 2014. Mr. Peng previously held executive roles at AMD, ATI, and MIPS Technologies.

Peng serves on the board of the Semiconductor Industry Association. He earned a Bachelor of Science, Electrical Engineering from Rensselaer Polytechnic Institute and a Master of Engineering, Electrical Engineering from Cornell University. He holds four U.S. patents.

Today, Mobile Semiconductor announced a new 55nm HD (High Density) memory compiler targeted at the cost sensitive IoT market. The new memory compiler boasts one of the highest density footprints in the industry dramatically reducing the die area and reducing customer product costs for sensors, smart locks, trackers and smart light bulbs.

Cameron Fisher, CEO and Founder of Mobile Semiconductor, said, “We believe that our success in the current 55nm Memory Compilers sets us apart from competitive offerings.  This new high-density product is well positioned to support our customer’s IoT products as they grow in features and capabilities. Our goal is to ensure that our customers can meet and exceed their silicon area goals and therefore reduce their costs.”

Key features include:

  • 15% to 33% smaller than previous 55nm compilers
  • At least 11% smaller than competitive solutions
  • Built on Mobile Semi’s volume designs at 55nm and 65nm
  • Available off the shelf today

Fisher continued, “Mobile Semiconductor remains the leader in providing memory compliers that target the needs of specific industries. We are proud of the fact that repeat customer purchases are close to 100%.  This includes customers moving to the next smaller node or building new products on the same node. Reducing the memory size offered by this new 55nm memory compiler gives our customers a compelling reason to choose Mobile Semiconductor for their cost sensitive IoT products.”

The 55nm HD memory compiler takes advantage of industry standard Bitcells provided by the top foundries.  All Mobile Semiconductor memory compilers are supported by a wide range of industry leading licensing options.

The Semiconductor Industry Association (SIA), representing U.S. leadership in semiconductor manufacturing, design, and research, today announced the addition of Broadcom Inc. (NASDAQ: AVGO) as an SIA member. Broadcom Chief Legal Officer Mark Brazeal is expected to be elected to the SIA board of directors at the association’s next board meeting on April 4.

“We are thrilled to welcome Broadcom, a leading player and valued voice in our industry, into the SIA tent,” said John Neuffer, SIA President and CEO. “The semiconductor industry is foundational to America’s economic and innovation base. Broadcom, headquartered in San Jose, California with over 11,000 employees in the U.S., will greatly strengthen SIA’s work to advance government policies that promote growth and innovation.  We are excited about the knowledge and experience Mark Brazeal will bring to the SIA board.”

In his current role at Broadcom, Brazeal is responsible for the legal, governance, and compliance functions across the company. Previously, he served as the Chief Legal Officer and Senior Vice President, IP Licensing for SanDisk Corporation through 2016. Before joining SanDisk, Brazeal spent 15 years at Broadcom Corporation, most recently as the Senior Vice President and Senior Deputy General Counsel.

“Semiconductors are at the heart of the greatest technological breakthroughs of our time,” said Brazeal. “I look forward to representing Broadcom on the SIA board and collaborating with my colleagues to promote smart government policies that will help us maintain the blistering pace of semiconductor innovation and move our great industry forward.”

Intel names Robert Swan CEO


January 31, 2019

Intel Corporation (NASDAQ: INTC) today announced that its board of directors has named Robert (Bob) Swan as chief executive officer. Swan, 58, who has been serving as Intel’s interim CEO for seven months and as chief financial officer since 2016, is the seventh CEO in Intel’s 50-year history. Swan has also been elected to Intel’s board of directors.

Intel Corporation has named Robert Swan as its chief executive officer. His promotion was announced Jan. 31, 2019. Swan, who previously served as the company’s chief financial officer and interim CEO, is the seventh CEO to lead the company based in Santa Clara, Calif. (Credit: Intel Corporation)

Todd Underwood, vice president of Finance and director of Intel’s Corporate Planning and Reporting, will assume the role of interim chief financial officer as the company undertakes an internal and external search for a permanent CFO.

“As Intel continues to transform its business to capture more of a large and expanding opportunity that includes the data center, artificial intelligence and autonomous driving, while continuing to get value from the PC business, the board concluded after a thorough search that Bob is the right leader to drive Intel into its next era of growth,” said Chairman Andy Bryant. “The search committee conducted a comprehensive evaluation of a wide range of internal and external candidates to identify the right leader at this critical juncture in Intel’s evolution. We considered many outstanding executives and we concluded the best choice is Bob. Important in the board’s decision was the outstanding job Bob did as interim CEO for the past seven months, as reflected in Intel’s outstanding results in 2018. Bob’s performance, his knowledge of the business, his command of our growth strategy, and the respect he has earned from our customers, our owners, and his colleagues confirmed he is the right executive to lead Intel.”

“In my role as interim CEO, I’ve developed an even deeper understanding of Intel’s opportunities and challenges, our people and our customers,” Swan said. “When I was first named interim CEO, I was immediately focused on running the company and working with our customers. When the board approached me to take on the role permanently, I jumped at the chance to lead this special company. This is an exciting time for Intel: 2018 was an outstanding year and we are in the midst of transforming the company to pursue our biggest market opportunity ever. I’m honored to have the chance to continue working alongside our board, our leadership team, and our more than 107,000 superb employees as we take the company forward.”

Swan is a proven leader with a strong track record of success both within and outside Intel. As interim CEO, Swan has managed the company’s operations in close collaboration with Intel’s senior leadership team. Swan has been Intel’s CFO since October 2016. In this role, he led the global finance, mergers and acquisitions, investor relations, IT and corporate strategy organizations. Prior to joining Intel, Swan served as an operating partner at General Atlantic LLC and served on Applied Materials’ board of directors. He previously spent nine years as CFO of eBay Inc., where he is currently a director. Earlier in his career, he was CFO of Electronic Data Systems Corp. and TRW Inc. He also served as CFO, COO and CEO of Webvan Group Inc., and began his career at General Electric, serving for 15 years in several senior finance roles.

Water molecules distort the electrical resistance of graphene, but a team of European researchers has discovered that when this two-dimensional material is integrated with the metal of a circuit, contact resistance is not impaired by humidity. This finding will help to develop new sensors -the interface between circuits and the real world- with a significant cost reduction.

The many applications of graphene, an atomically-thin sheet of carbon atoms with extraordinary conductivity and mechanical properties, include the manufacture of sensors. These transform environmental parameters into electrical signals that can be processed and measured with a computer.

Due to their two-dimensional structure, graphene-based sensors are extremely sensitive and promise good performance at low manufacturing cost in the next years.

To achieve this, graphene needs to make efficient electrical contacts when integrated with a conventional electronic circuit. Such proper contacts are crucial in any sensor and significantly affect its performance.

But a problem arises: graphene is sensitive to humidity, to the water molecules in the surrounding air that are adsorbed onto its surface. H2O molecules change the electrical resistance of this carbon material, which introduces a false signal into the sensor.

However, Swedish scientists have found that when graphene binds to the metal of electronic circuits, the contact resistance (the part of a material’s total resistance due to imperfect contact at the interface) is not affected by moisture.

“This will make life easier for sensor designers, since they won’t have to worry about humidity influencing the contacts, just the influence on the graphene itself,” explains Arne Quellmalz, a PhD student at KTH Royal Institute of Technology (Sweden) and the main researcher of the research.

The study, published in the journal ACS Applied Materials & Interfaces, has been carried out experimentally using graphene together with gold metallization and silica substrates in transmission line model test structures, as well as computer simulations.

“By combining graphene with conventional electronics, you can take advantage of both the unique properties of graphene and the low cost of conventional integrated circuits.” says Quellmalz, “One way of combining these two technologies is to place the graphene on top of finished electronics, rather than depositing the metal on top the graphene sheet.”

As part of the European CO2-DETECT project, the authors are applying this new approach to create the first prototypes of graphene-based sensors. More specifically, the purpose is to measure carbon dioxide (CO2), the main greenhouse gas, by means of optical detection of mid-infrared light and at lower costs than with other technologies.

In addition to the KTH Royal Institute of Technology, the companies SenseAir AB from Sweden and Amo GmbH from Germany are likewise participants in the CO2-DETECT project, as is the Catalan Institute of Nanotechnology (ICN) from Barcelona.

Vanguard International Semiconductor Corporation (VIS) and GLOBALFOUNDRIES (GF) today announced that VIS will acquire GF’s Fab 3E in Tampines, Singapore. The transaction includes buildings, facilities, and equipment, as well as IP associated with GF’s MEMS business. GF will continue to operate the facility through the end of 2019, providing a transition period to facilitate technology transfers for VIS and existing GF customers. Fab 3E currently manages a monthly capacity of approximately 35,000 8-inch wafers. The transaction amounts to $236 million USD and the transfer of ownership is set to be completed on December 31st, 2019.

VIS and GF have already reached consensus on the transfer of Fab 3E’s employees and customers. Both companies believe that employees are the most important assets of a company, so their interests should be put as the first priority during the transition; while ensuring no disruption to customers whose products are in production at the fab. Under this premise, VIS will extend employment offers to all employees currently working at Fab 3E, as well as continuously provide existing customers at Fab 3E with its foundry service, including MEMS customers.

“I appreciate the support of GF’s board and management team for this transaction, giving VIS an opportunity to continue expanding its capacity and reinforce momentum for future growth,” said Mr. Leuh Fang, Chairman of VIS. “Since its foundation, VIS has already had three separate experiences of successfully transforming a DRAM fab into a foundry fab. We believe this transaction is a win-win for both VIS and GF; and to VIS, it is also a decision that benefits all of our customers, employees, and shareholders. VIS will uphold its philosophy and principles to continue satisfying customers’ demands in capacity and technology, sustaining profitability and growth, and rewarding our shareholders.”

“This transaction is part of our strategy to streamline our global manufacturing footprint and increase our focus in Singapore on technologies where we have clear differentiation such as RF, embedded memory and advanced analog features,” said GF CEO Tom Caulfield. “Consolidating our 200mm operations in Singapore into one campus will also help reduce our operating costs by leveraging the scale of our gigafab facility in Woodlands. VIS is the right partner to leverage the Fab 3E asset going forward.”

VIS’s capacity has been fully utilized since 2018, and it is in the interests of its customers that VIS expands capacity to meet growing demands. The new fab is expected to contribute more than 400,000 8-inch wafers per year. This acquisition demonstrates the determination and commitment of VIS to accelerate capacity expansion.