Tag Archives: Top Story Right

By Christian G. Dieseldorff and Eugenia Liu

SEMI FabView update for calendar year Q3 2018

Global fab construction investment shows continuing strength, with 19 new fab projects expected to begin construction in 2019 and 2020, based on the latest data published in SEMI’s World Fab Forecast.

Fab investment is just one indicator of how growing demand in areas such as high-performance computing, data storage, artificial intelligence (AI), cloud computing, and automotive are driving the fourth consecutive year of spending growth in the semiconductor industry. Below are a few highlights* from September’s SEMI FabView:

Memory: Not fading

  • Micron plans to invest $3 billion by 2030 in Manassas, Virginia – These investments, driven by strong demand for automotive applications, are contemplated in Micron’s long-term model. The production ramp is anticipated to be in the first half of 2020.
  • SK Hynix to build new DRAM fab in Icheon (Gyeonggi Province), Korea – The construction, to be completed by the end of 2020, will adopt 1znm node (probably EUV). Total investment is estimated to exceed $13 billion.
  • Nanya Technology doubles 2018 capex plan – The increase is for additional DRAM capacity and more 20nm DRAM conversion (from 30nm).

200mm and below: Not leading edge, but continues to draw investment

  • Vanguard changes fab investment strategy – Vanguard will focus on 200 mm and has scrapped its plan for 300mm expansion.
  • Murata to invest into 150mm expansion – Murata announced a 5 billion Yen investment (US$44.6 million) in a new fab extension in Vantaa, Finland.

Investment, M&A in Analog, Logic, Power and Opto Segments

  • Texas Instruments is looking to invest $3.2 billion in new fab construction in 2019 – Texas Instruments is eyeing Richardson, Texas and also considering sites outside Texas.
  • Bosch 300mm fab in Dresden, Germany – Bosch held a groundbreaking ceremony on April 24. Equipment installation is expected in 2H19.
  • Microchip completes acquisition of Microsemi – Microchip closed its $8.45 billion acquisition of Microsemi on May 29. Microsemi has five fabs in the U.S. with a wide range of semiconductor products and system solutions.

New fabs in China keep on coming

  • Shanghai Jita Semiconductor/Huada Semiconductor – Shanghai Jita Semiconductor, a subsidiary of Huada Semiconductor and China Electronics Corporation (CEC), announced plans earlier this month to build both 200 mm and 300 mm semiconductor fabs for analog and power semiconductors in Shanghai. The combined fab investment will total $5.18 billion.
  • Hamamatsu Photonics building 200 mm fab – Hamamatsu announced that it is building a new facility Investment of 2.8 billion Yen (US$25 million) to boost opto semiconductor capacity. Production is anticipated to start in late 2019.

*Actual FabView updates provide more detail

SEMI FabView, a mobile-friendly, interactive version of SEMI’s popular World Fab Forecast, delivers on-demand fab information such as fab spending and capacity for over 1,200 facilities, including over 60 planned facilities worldwide, across a wide range of product segments including Power, GPU, Memory, Foundry, MEMS and Sensors fabs. Fab data include region, start of construction, operation, construction and equipment spending, capacity, wafer sizes, product types and geometries. SEMI FabView subscribers receive forecast model updates through SEMI’s World Fab Database.  Click here for a trial if you want to experience SEMI FabView first hand.

Christian G. Dieseldorff is senior principal analyst and Eugenia Liu is senior product marketing manager, Industry Research and Statistics, SEMI, Milpitas, California. 

Originally published on the SEMI blog.

By Walt Custer

Global economy

Manufacturing activity continues to expand – but at a slowing pace (Chart 1). The Global PMI was 52.5 in August, down from 52.8 in July and its recent high of 54.5 in December. PMI values >50 indicate an expansion.

World manufacturing growth has slowed but its growth rate varies significantly by region. Chart 2 compares the PMI values over time for the World, USA, Europe and China. Recently China and Europe have registered slower growth but U.S. growth is expanding (based on the Institute for Supply Management’s PMI). How long U.S. manufacturing will continue to accelerate remains to be seen. Geopolitical issues abound.

Semiconductor industry

In the semiconductor industry both semiconductors and SEMI capital equipment continued to register double-digit growth in July (Chart 3), but these growth rates are now moderating. In July, World semiconductor shipments were up 17.4 percent and SEMI capital equipment sales rose 13.9 percent on a 3-month growth basis.

However, SEMI equipment growth rates also vary widely by region. Per Chart 4, China growth is accelerating, Taiwan and South Korea are contracting, and Europe and the USA are still expanding but at slower rates.

Timely World and regional industry information is key to understanding present and future business conditions and this data requires careful watching in these fast-changing times.

Walt Custer of Custer Consulting Group is an analyst focused on the global electronics industry.

Originally published on the SEMI blog.

By Michael Droeger

Over the past three decades, most of the world’s innovations have centered largely on business models and involved iterative advances of existing technologies, with none matching the global impact of the top 10 semiconductor industry discoveries and advances, Dr. Morris Chang, founder of TSMC and the IC foundry model, said at SEMICON Taiwan 2018 this week.

Few have as clear a perspective on the transformative power of semiconductors as Dr. Chang, founder of TSMC and father of the IC foundry model. Keynoting the IC60 Master Forum celebrating the 60th anniversary of the invention of the integrated circuit (IC), Dr. Chang listed what he considers the 10 key semiconductor industry innovation milestones since 1948:

1. Invention of the transistor by Shockley, Bardeen, and Brattain – 1948

2. Silicon transistor – 1954

3. Integrated circuit – 1958

4. Moore’s Law – 1965

5. MOS technology

  1. MOS FET – 1964
  2. Silicon gate – 1967
  3. CMOS  – 1970

6. Memory

  1. DRAM – 1966
  2. Flash – 1967

7. Outsourced assembly and test (OSAT) – 1960s

8. Microprocessor – 1970

9. VLSI systems design – 1970-1980

  1. IP and design tools – 1980-present

10. Foundry model – 1985

Among the most consequential semiconductor advances may be yet to come, Dr. Chang said, citing innovations including artificial intelligence (AI) and machine learning, new device architectures, Extreme Ultraviolet lithography (EUV), 2.5D/3D packaging, and new materials such as graphene and carbon nanotubes.

Dr. Chang argued that because bringing an innovation into production is immensely more expensive than proving a theory in a lab, innovators are not always the ones to implement and benefit from their novel ideas. Today, innovation costs are skyrocketing, driving more consolidation across the supply chain.

Michael Droeger is director of marketing at SEMI.

Originally published on the SEMI blog.

Products built with microelectromechanical systems (MEMS) technology are forecast to account for 73% of the $9.3 billion semiconductor sensor market in 2018 and about 47% of the projected 24.1 billion total sensor units to be shipped globally this year, according to IC Insights’ 2018 O-S-D Report—A Market Analysis and Forecast for Optoelectronics, Sensors/Actuators, and Discretes.  Revenues for MEMS-built sensors—including accelerometers, gyroscope devices, pressure sensors, and microphone chips—are expected to grow 10% in 2018 to $6.8 billion compared to nearly $6.1 billion in 2017, which was a 17% increase from $5.2 billion in 2016, the O-S-D Report says.  Shipments of MEMS-built sensors are forecast to rise about 11% in 2018 to 11.1 billion after growing 19% in 2016.

An additional $5.9 billion in sales is expected to be generated in 2018 by MEMS-built actuators, which use their microelectromechanical systems transducers to translate and initiate action—such as dispensing ink in printers or drugs in hospital patients, reflecting light on tilting micromirrors in digital projectors, or filtering radio-frequency signals by converting RF to acoustic waves across structures on chips.  Total sales of MEMS-built sensors and actuators are projected to grow 10% in 2018 to $12.7 billion after increasing nearly 18% in 2017 and 15% in 2016 (Figure 1).

Figure 1

In terms of unit volume, shipments of MEMS-built sensors and actuators are expected to grow by slightly less than 12% to 13.1 billion units worldwide after climbing 20% in 2017 and rising 11% in 2016.  Total revenues for MEMS-based sensors and actuators are projected to increase by a compound annual growth rate (CAGR) of 9.2% between 2017 and 2022 to reach $17.8 billion in the final year of the forecast, according to the 2018 O-S-D Report.  Worldwide shipments of these MEMS-built semiconductors are expected to grow by a CAGR of 11.4% in the 2017-2022 period to 20.2 billion units at the end of the forecast.

One of the biggest changes expected in the five-year forecast period will be greater stability in the average selling price for MEMS-built devices and significantly less ASP erosion than in the past 10 years. The ASP for MEMS-built sensors and actuators is projected to drop by a CAGR of -2.0% between 2017 and 2022 compared to a -4.7% annual rate of decline in the 2012-2017 period and the steep CAGR plunge of -13.6% between 2007 and 2012.  The ASP for MEMS-built devices is expected to be $0.88 in 2022 versus $0.97 in 2017, $1.24 in 2012, and $2.57 in 2007, says the 2018 report.

The spread of MEMS-based sensors and actuators into a broader range of new “autonomous and “intelligent” automated applications—such as those connected to the Internet of Things (IoT) and containing artificial intelligence (AI)—will help keep ASPs from falling as much as they did in the last 10 years.  IC Insights believes many MEMS-based semiconductors are becoming more specialized for certain applications, which will help insulate them from pricing pressures in the market.

Micron Technology, Inc., (NASDAQ:MU) today announced plans to invest $3 billion by 2030 to increase memory production at its plant in Manassas, Virginia, creating 1,100 new jobs roughly over the next decade. These investments are contemplated in Micron’s long-term model to invest capital expenditure in the low thirties as a percent of revenue. The expansion will position the Manassas site — located about 40 miles west of Washington, D.C. — to support Micron’s leadership in the rapidly growing market for high quality, high reliability memory products.

Source: Micron

“Micron’s Manassas site manufactures our long-lifecycle products that are built using our mature process technologies, and primarily sold into the automotive, networking and industrial markets,” said Micron President and CEO Sanjay Mehrotra. “These products support a diverse set of applications such as industrial automation, drones, the IoT (Internet of Things) and in-vehicle experience applications for automotive. This business delivers strong profitability and stable, growing free cash flow. Micron is grateful for the extensive engagement of state and local officials since early this year to help bring our Manassas expansion to fruition. We are excited to increase our commitment to the community through the creation of new highly skilled jobs, expanded facilities and education initiatives.”

“Micron’s expansion in the City of Manassas represents one of the largest manufacturing investments in the history of Virginia and will position the Commonwealth as a leader in unmanned systems and Internet of Things,” said Governor Northam. “This $3 billion investment will have a tremendous impact on our economy by creating 1,100 high-demand jobs, and solidifies Micron as one of the Commonwealth’s largest exporters. We thank Micron for choosing to deepen their roots in Virginia and look forward to partnering in their next chapter of major growth.”

The initial clean room expansion is expected to be completed in the fall of 2019 with production ramp in the first half of 2020. This expansion will add less than 5% to Micron’s global clean room space footprint and will primarily support enablement of DRAM and NAND technology transitions as well as modest capacity increase at the site, in-line with growing customer demand for Micron’s long-lifecycle products.

“As a leading global supplier of automotive electronics systems and components, ZF appreciates the long-standing support of Micron to our business,” said Karsten Mueller, vice president, Corporate Materials Management, Global Commodity Electronics at ZF Friedrichshafen AG. “Meeting the ever-increasing demands for automotive applications will require significantly greater memory as the dual trends of advanced safety and autonomy drive the industry forward. Micron’s decision to expand the manufacturing and R&D capabilities at this IATF-certified facility is another indication that this growth should only accelerate in the future.”

As part of this expansion, Micron will also establish a global research development center in Manassas for the development of memory and storage solutions focused mainly on the automotive, industrial and networking markets. The research and development center will include laboratories, test equipment and a staff of approximately 100 engineers.

The Virginia Economic Development Partnership (VEDP) worked with the City of Manassas and the General Assembly’s Major Employment and Investment (MEI) Project Approval Commission to secure the project for Virginia. Micron will be eligible to receive an MEI custom performance grant of $70 million for site preparation and facility costs, subject to approval by the Virginia General Assembly. Additionally, the City of Manassas and utility partners are providing a broader, comprehensive support package to enable the expansion, including substantial infrastructure upgrades and additional incentives.

By Christopher Morales, Head of Security Analytics, Vectra

On August 3, Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC), the largest chip fabricator globally introduced a WannaCry Ransomware cryptowormvariant onto its information technology/operational technology (IT/OT) networks. A TSMC supplier installed infected software on a new fabrication tool and connected it to the network, facilitating the malware infestation.

The infection spread quickly, taking out 10,000+ unpatched Windows 7 machines that run the chip fab company’s tool automation interface. The crypto worm crashed and rebooted systems endlessly, forcing several plants in Taichung, Hsinchu andTainan to shut down through much of the weekend.

The infection crippled materials handling systems and production equipment as well as Windows 7 computers. Some of the plants were producing SoC chips for the AppleiPhone 8 and X models. The incident’s connection to Apple and the iPhone heightened its visibility in the news media.

According to TSMC CEO C.C. Wei, patching for the Windows 7 machines requires computer downtime and collaboration with equipment suppliers. The absence of currentpatches created an environment where WannaCry could easily propagate.

The 2018 Spotlight Report on Manufacturing published by Vectra a few weeks before the incident foretold TSMC’s infection, which could cost the company as much as $255 million.

Smart manufacturer cybersecurity risks are increasing

According to the TSMC website, the company had “introduced new applications such as IoT, intelligent mobile devices and mobile robots to consolidate data collection, yield traceability, workflow efficiency, and material transportation to continuously enhance fab operation efficiency.” Further, TSMC had “integrated automatic manufacturing systems,” according to its website.

These innovations are typical in the evolution of Industry 4.0, which has increased the risk of cyber attacks against manufacturers.

But as manufacturers moved from air-gapped industrial systems to cloud-connectedsystems as part of the IT/OT convergence – using unpartitioned networks and insufficient access controls for proliferating IIoT devices – they created a massive, vulnerable attack surface, according to the Vectra report.

While air-gapped systems such as industrial controls have no connections by design to guard against malicious tampering, IT/OT convergence has connected these systems to information technologynetworks with little accounting for security vulnerabilities.

Many factories connect IIoT devices to flat, unpartitioned networks that rely on communication with general computing devices and enterprise applications. Since IIoT devices support few if any native cybersecurity measures, connecting them to easily infected applications, computers and unsegregated IP networks only invites trouble.

In the past, manufacturers relied on more customized, proprietary protocols, which made mounting an attack more difficult for cybercriminals. The conversion from proprietary protocols to standard protocols makes it easier to infiltrate networks to spy, spread and steal.

Few if any cyberattackers know and understand the proprietary protocols those closed legacy systems used. But it’s easy for most criminal hackers and their exploits to access standard IP network protocols just as WannaCry abuses the SMB protocol where there is no patch.

Real-time network visibility is crucial 

Industry 4.0 brings with it a new operational risk for connected, smart manufacturers and digital supply networks. The interconnected nature of Industry 4.0-driven operations and the pace of digital transformation mean that cyber attacks can have far more damaging effects than ever before, and manufacturers and their supply networks may not be preparedfor the risks.

Wherever cyber attacks interfere business continuity for business and information processes, they can also disrupt operational technologies that render products and get them out the door.

For cyber-risk to be adequately addressedin the age of Industry 4.0, manufacturing organizations need to ensure that proper visibility and response capabilities are in place to detect and respond to events as they occur. As in the case of the TSMC ransomware debacle, anything less than real-time detection and response is too little, too late to avoid production downtime.

There is no visibility into these systems to enable real-time detection before cyber attacks spread. Visibility into these internal connected systems is necessary to curtail the extent of damage from a cyberattack.

Manufacturing security operations now require automated, real-time analysis of entire networks to proactively detect and respond to in-progress threats before they do damage.

The Vectra 2018 Spotlight Report on Manufacturing

The 2018 Spotlight Report on Manufacturing delineates the many attack types and behaviors that the Cognito platform captured. The Cognito threat-detection and hunting platform monitored traffic and collected rich metadata from more than 4million devices and workloads from customer cloud, data center, and enterprise environmentsto reveal the cyberattacker behaviors.

Cyber attacks on manufacturers increased in severity from January to June 2018 based on data that the Vectra Cognito platform collected. The Vectra report confirms that all manufacturing industries are at equal risk of cyberattacks.

To learn about other findings pertinent to your Industry 4.0 cybersecurity risk, download the 2018 Spotlight Report on Manufacturing.

Christopher Morales is the head of security analytics at Vectra, a San Jose, Calif. cybersecurity firm that detects hidden cyberattacks and helps threat hunters improve the efficiency of incident investigations.

A team of researchers led by the University of Minnesota has developed a new material that could potentially improve the efficiency of computer processing and memory. The researchers have filed a patent on the material with support from the Semiconductor Research Corporation, and people in the semiconductor industry have already requested samples of the material.

The findings are published in Nature Materials, a peer-reviewed scientific journal published by Nature Publishing Group.

This cross-sectional transmission electron microscope image shows a sample used for the charge-to-spin conversion experiment. The nano-sized grains of less than 6 nanometers in the sputtered topological insulator layer created new physical properties for the material that changed the behavior of the electrons in the material. Credit: Wang Group, University of Minnesota

“We used a quantum material that has attracted a lot of attention by the semiconductor industry in the past few years, but created it in unique way that resulted in a material with new physical and spin-electronic properties that could greatly improve computing and memory efficiency,” said lead researcher Jian-Ping Wang, a University of Minnesota Distinguished McKnight Professor and Robert F. Hartmann Chair in electrical engineering.

The new material is in a class of materials called “topological insulators,” which have been studied recently by physics and materials research communities and the semiconductor industry because of their unique spin-electronic transport and magnetic properties. Topological insulators are usually created using a single crystal growth process. Another common fabrication technique uses a process called Molecular Beam Epitaxy in which crystals are grown in a thin film. Both of these techniques cannot be easily scaled up for use in the semiconductor industry.

In this study, researchers started with bismuth selenide (Bi2Se3), a compound of bismuth and selenium. They then used a thin film deposition technique called “sputtering,” which is driven by the momentum exchange between the ions and atoms in the target materials due to collisions. While the sputtering technique is common in the semiconductor industry, this is the first time it has been used to create a topological insulator material that could be scaled up for semiconductor and magnetic industry applications.

However, the fact that the sputtering technique worked was not the most surprising part of the experiment. The nano-sized grains of less than 6 nanometers in the sputtered topological insulator layer created new physical properties for the material that changed the behavior of the electrons in the material. After testing the new material, the researchers found it to be 18 times more efficient in computing processing and memory compared to current materials.

“As the size of the grains decreased, we experienced what we call ‘quantum confinement’ in which the electrons in the material act differently giving us more control over the electron behavior,” said study co-author Tony Low, a University of Minnesota assistant professor of electrical and computer engineering.

Researchers studied the material using the University of Minnesota’s unique high-resolution transmission electron microscopy (TEM), a microscopy technique in which a beam of electrons is transmitted through a specimen to form an image.

“Using our advanced aberration-corrected scanning TEM we managed to identify those nano-sized grains and their interfaces in the film,” said Andre Mkhoyan, a University of Minnesota associate professor of chemical engineering and materials science and electron microscopy expert.

Researchers say this is only the beginning and that this discovery could open the door to more advances in the semiconductor industry as well as related industries, such as magnetic random access memory (MRAM) technology.

“With the new physics of these materials could come many new applications,” said Mahendra DC (Dangi Chhetri), first author of the paper and a physics Ph.D. student in Professor Wang’s lab.

Wang agrees that this cutting-edge research could make a big impact.

“Using the sputtering process to fabricate a quantum material like a bismuth-selenide-based topological insulator is against the intuitive instincts of all researchers in the field and actually is not supported by any existing theory,” Wang said. “Four years ago, with a strong support from Semiconductor Research Corporation and the Defense Advanced Research Projects Agency, we started with a big idea to search for a practical pathway to grow and apply the topological insulator material for future computing and memory devices. Our surprising experimental discovery led to a new theory for topological insulator materials.

“Research is all about being patient and collaborating with team members. This time there was a big pay off,” Wang said.

The demise of Qualcomm’s pending $44 billion purchase of NXP Semiconductors in late July along with growing regulatory reviews of chip merger agreements, efforts by countries to protect domestic technology, and the escalation of global trade friction all suggest semiconductor acquisitions are hitting a ceiling in the size of doable deals.  It is becoming less likely that semiconductor acquisitions over $40 billion can be completed or even attempted in the current geopolitical environment and brewing battles over global trade.

IC Insights believes a combination of factors—including the growing high dollar value of major chip merger agreements, complexities in combining large businesses together, and greater scrutiny of governments protecting their domestic base of suppliers—will stifle ever-larger mega-transactions in the semiconductor industry in the foreseeable future.  Figure 1 ranks the 10 largest semiconductor merger and acquisition announcements and underscores the growth in size of these M&A transactions.  Eight of the 10 largest announcements occurred in the last three years with only the biggest deal (Qualcomm buying NXP) failing to be completed.

Figure 1

It is important to note that IC Insights’ M&A list only covers semiconductor suppliers, chipmakers, and providers of integrated circuit intellectual property (IP) and excludes acquisitions of software and system-level businesses by IC companies  (such as Intel’s $15.3 billion purchase of Mobileye, an Israeli-based developer of digital imaging technology for autonomous vehicles, in August 2017).  This M&A list also excludes transactions involving semiconductor capital equipment suppliers, material producers, chip packaging and testing companies, and design automation software firms.

Qualcomm’s $44 billion cash purchase of NXP would have been the largest semiconductor acquisition ever if it was completed, but the deal—originally announced in October 2016 at nearly $39 billion and raised to $44 billion in February 2018—was canceled in the last week of July because China had not cleared the transaction.  China was the last country needed for an approval of the merger, and it was believed to be close to clearing the purchase in 2Q18, but growing threats of tariffs in a brewing trade war with the U.S. and moves to block Chinese acquisitions of American IC companies caused China to taken no action on the $44 billion acquisition in time for a deadline set by Qualcomm and NXP.  U.S.-based Qualcomm canceled the acquisition on July 26 and quickly paid NXP in the Netherlands a $2 billion breakup fee so the two companies could move on separately.

Prior to Qualcomm’s failed $44 billion offer for NXP, the largest semiconductor acquisition was Avago Technologies’ $37 billion cash and stock purchase of Broadcom in early 2016.  Avago renamed itself Broadcom Limited after the purchase and launched a failed $121 billion hostile takeover bid for Qualcomm at the end of 2017.  It lowered the unsolicited bid to $117 billion in February 2018 after Qualcomm raised its offer for NXP to $44 billion.  In March 2018, U.S. President Donald Trump blocked Broadcom’s $117 billion takeover bid for Qualcomm after concerns were raised in the U.S. government about the potential loss of cellular technology leadership to Chinese companies, if the hostile acquisition was completed. After the presidential order, Broadcom executives said the company was considering other acquisition targets, with cash, that would be smaller and more focused.

The global semiconductor industry has been reshaped by a historic wave of mergers and acquisitions during the past three years, with about 100 M&A agreements being reached between 2015 and the middle of 2018 with the combined value of these transactions being more than $245 billion, based on data collected by IC Insights and contained within its Strategic Reviews database subscription service and in The 2018 McClean Report on the IC Industry.  A record-high $107.3 billion in semiconductor acquisition agreements were announced in 2015.  The second highest total for semiconductor M&A agreements was then reached in 2016 at $99.8 billion.   Semiconductor acquisition announcements reached a total value of $28.3 billion in 2017, which was twice the industry’s annual average of about $12.6 billion in the first half of this decade but significantly less than 2015 and 2016, when M&A was sweeping through the chip industry at historic levels.  In the first six months of 2018, semiconductor acquisition announcements had a total value of about $9.6 billion, based on IC Insights’ running tally of announced M&A deals.

IC Insights recently released its Mid-Year Update to The McClean Report 2018.  The update includes a revised forecast of the largest and fastest-growing IC product categories this year.  Sales and unit growth rates are shown for each of the 33 IC product categories defined by the World Semiconductor Trade Statistics (WSTS) organization in the Mid-Year Update.

The five largest IC product categories in terms of sales revenue and unit shipments are shown in Figure 1.  With forecast sales of $101.6 billion, (39% growth) the DRAM market is expected to be the largest of all IC product categories in 2018, repeating the ranking it held last year.  If the sales level is achieved, it would mark the first time an individual IC product category has surpassed $100.0 billion in annual sales. The DRAM market is forecast to account for 24% of IC sales in 2018.  The NAND flash market is expected to achieve the second-largest revenue level with total sales of $62.6 billion this year. Taken together, the two memory categories are forecast to account for 38% of the total $428.0 billion IC market in 2018.

Figure 1

For many years, the standard PC/server MPU category topped the list of largest IC product segments, but with ongoing increases in memory average selling prices, the MPU category is expected to slip to the third position in 2018.  In the Mid-Year Update, IC Insights slightly raises its forecast for 2018 sales in the MPU category to show revenues increasing 5% to an all-time high of $50.8 billion, after a 6% increase in 2017 to the current record high of $48.5 billion.  Helping drive sales this year are AI-controlled systems and data-sharing applications over the Internet of Things.  Cloud computing, machine learning, and the expected tidal wave of data traffic coming from connected systems and sensors is also fueling MPU sales growth this year.

Two special purpose logic categories—computer and peripherals, and wireless communications—are forecast to round out the top five largest product categories for 2018.

Four of the five largest categories in terms of unit shipments are forecast to be some type of analog device.  Total analog units are expected to account for 54% of the total 318.1 billion IC shipments forecast to ship this year.  Power management analog devices are projected to account for 22% of total IC units and are forecast to exceed the combined unit shipment total of the next three categories on the list.  As the name implies, power management analog ICs help regulate power usage and to keep ICs and systems running cooler, to manage power usage, and ultimately to help extend battery life—essential qualities for an increasingly mobile and battery-powered world of devices.

By Ajit Manocha

At a Glance

“Software is eating the world … and AI is eating software.” Amir Husain, author of The Sentient Machine, at SEMICON West 2018

We’re living in a digital world where semiconductors have been taken for granted. But, Artificial Intelligence (AI) is changing everything – and bringing semiconductors back into the deserved spotlight. AI’s potential market of hundreds of zettabytes and trillions of dollars relies on new semiconductor architectures and compute platforms. Making these AI semiconductor engines will require a wildly innovative range of new materials, equipment, and design methodologies.

Moore’s Law carried us the past 50-plus years and as we’re now stepping into the dawn of AI’s potential, we can see that the coming Cognitive Era will drive its own exponential growth curve. This is great for the world – virtually every industry will be transformed, and people’s lives will get better – and it’s fantastic for our industry. This truly is the very best time to be working in our industry. I’m excited to be at SEMI in this inflection period and at the center of the collaborative platforms that bring the electronics manufacturing supply chain together to Connect, Collaborate, and Innovate to realize the new Cognitive Era. I invite you to partner with SEMI in building the foundation for the Cognitive Era to increase the growth and prosperity of our industry.

The World Wakes Up

Our lives have become digital. An Amazon Echo wakes us up and answers questions about the weather and traffic. Google Maps tells us the best way to get to a meeting. Yelp finds the best nearby restaurant. A Tweet now even informs us of the latest change in government policy. It’s a digital world that we live in – and the world already takes it for granted.

We in the industry know that the digital world only works because of the semiconductors we make and because of our integrated electronics manufacturing supply chain. We make the materials and equipment that, in turn, make the chips that become the beating hearts of the digital economy.

But, semiconductors have been largely invisible – hidden away under and inside a smart speaker, locked deep within a phone, buried in data centers and out of view. Meanwhile, the internet companies like Google, Amazon, Alibaba, Tencent, and Facebook stole the meaning of “Tech” and were given most of the credit for our digital world.

But, finally, things are changing – it’s all coming back to semiconductors!

AI Changing Everything

Over $400B in semiconductors were sold in 2017 – those unseen chips like hearts beating away in Apple computers, in mobile phones for online shopping and social media, and in televisions showing Netflix. Now internet companies Alphabet, Alibaba, Amazon, Facebook, Microsoft and others are rushing to develop their own chips. Silicon is back in the Silicon Valley! Hardware is, once again, the place to be. Why? We are now entering the epoch of Artificial Intelligence (AI) – and semiconductors, and new compute architectures, are the key to AI. At this moment, hardware, not software, is the AI enabler to make leaps in performance and to usher in new architectures to become brain-like with neural networks.

Beyond major AI chip investments like Google’s (Alphabet) $300M+ program to develop its Tensor Processing Unit (TPU) chip, there’s been a surge in new chip startups and VC funding. Last year, VCs (with corporate investors) invested more than $1.5B in new AI chip startups – doubling the rate from the prior year.

After years of consolidation, there is, as some have described, a “Cambrian Explosion” of semiconductor startups with names like Cerebras, Graphcore, Wave Computing, Horizon Robotics, Cambricon Technologies, and DeePhi from the US, Europe, and China. Cambricon (China) has already become the first AI chip “Unicorn” (startup valued $1B+) with a valuation of more than $2.5B after their recent Round B financing. It’s a new silicon world and a new race, as Cade Metz (The New York Times, 1/14/2018) said, “… everyone is starting from the same place: the beginning of a new market.”

Winning at AI is very big business. John Kelly, SVP Cognitive Solutions and Research at IBM, in his SEMICON West keynote earlier this month, said, we’re in the era of Artificial Intelligence with more than a $2T opportunity for AI decision making support on top of the $1.5T IT business in 2025. McKinsey estimates deep learning could account for between $3.5T and $5.8T in annual value.

As John Kelly presented, AI will transform entire industries – not just our personal devices and lives. The $2T AI decision making support opportunity in 2025 is projected to transform the major economy industries as follows:

Moore’s Law describes the exponential increase in the number of transistors per area that has driven growth, and has been the engine for digital innovation, through first the computer era and then the mobility era and now into the dawn of the data era. While the Dennard scaling approach to Moore’s Law may be slowing, the data-centric era continues to drive demand and the industry continues to find new ways to pack more transistors into less volume. Chip sales are forecast to pass $0.5T in 2019 and I predict they will surpass $1T before 2030.

It turns out the Smart is not enough – we must reach “Beyond Smart.”

Beyond Smart – The Cognitive Era

As we move further into the data-centric age, we see it is more than Big Data and AI, it is, instead, the dawn of a wholly new cognitive era. SEMICON West’s 2018 theme was “Beyond Smart” because we are standing at the inflection from sensors triggering actions (smart) to systems that learn and make decisions (cognitive). Devices are moving “beyond smart” to being “cognitive or aware.” Gary Dickerson (CEO of Applied Materials) at SEMICON West said, “… we are in the beginning of the first inning of a major inflection.”

Even in the early dawn of the cognitive era, the volume of data is simply astonishing. In the last 24 months, we create more than 90% of all historic digital data. By 2025 we expect AI to generate 160 zettabytes – with 80% of that unstructured data. Moore’s Law is an exponential, but as John Kelly points out, AI’s deep learning is driving its own exponential with performance/watt increasing 2.5X each year.

AI was the focus of SEMICON West’s Day 1 keynotes – and a common theme through much of the events programming. There was a common language in the keynotes by John Kelly, Gary Dickerson, and William Dally (Chief Scientist and SVP of Research NVIDIA), and others. We heard how AI is based on data, algorithms, and compute. I was inspired by these talks and for the potential for AI and the cognitive era.

Looking ahead, I believe data + algorithms + compute + machine learning = knowledge and cognition. My vision is that this AI knowledge and cognition will be the catalyst to create new modes of systems transformations that will usher in the next Industrial Revolution. As the 4th Industrial Revolution becomes a reality, I look forward to working with others in SEMI Think Tanks to imagine the 5th Industrial Revolution – and its opportunities for our industry. I believe that it will make our lives better, healthier, more prosperous, and more fulfilled.

A sentiment shared by many speakers at SEMICON West was – this is the most exciting time to be in the semiconductor manufacturing industry. Many wished they were just now starting in the industry as this is the most interesting inflection and transformation ever. There is a flood of new architectures, new materials, new equipment, new processes – and a new system-based design approach to enable the Cognitive Era. We, in hardware manufacturing, are in the driver’s seat for this incredible ride.

SEMI is working to help its members speed their time to better business results – and to take full advantage of the Cognitive Era and AI opportunity. At SEMICON West 2018, SEMI provided a broad and deep slate of program education and spotlighted AI expertise across the electronics manufacturing supply. In case you missed it, SEMI also provided

  • Seven keynotes and dozens of expert panelists
  • Semiconductor venture funding program – problems and solutions for the ecosystem
  • SEMI Smart Workforce Pavilion with over 600 students registered to learn about the industry
  • Smart Pavilions including Smart Manufacturing and Smart Automotive

SEMI highlighted the five key vertical application platforms where our industry needs to collaborate across the full supply chain and streamline the supply chain for efficiency. The five are: IoT, Smart Transportation, Smart Manufacturing, Smart MedTech, and Smart Data. These verticals drive huge business potential and are just one of the reasons that SEMICON West has become the gathering place of the extended electronics manufacturing supply chain.

With SEMI, together we can realize the potential of the coming Cognitive Era. SEMI members can advance the industry with SEMI collective action in Workforce Development, Advocacy (public policy and regulatory), Standards to synchronize the industry, and in the many SEMI technology communities and special interest groups – to increase the global industry’s rate of growth and overall level of prosperity. For more information, please visit www.semi.org; to become a member, please visit http://www.semi.org/en/become-member-join-semi.

Ajit Manocha is President and CEO of SEMI

Originally published on the SEMI blog.