Top 20 Semiconductor Manufacturers 2014 Some IC Packaging Firms
1. Intel $50B (US) 11. Renesas $6.9B (JP) - ASE (TW)
2. Samsung $38B (KR) 12. SanDisk $6.1B (US) - Amkor (US)
3. Qualcomm $19.3B (US) 13. Infineon $6.1B (DE) - OSE (TW)
4. Micron $16.4B (US) 14. NXP $5.5B (NL) - Promex (US)
5. SK Hynix $15.7B (KR) 15. Avago $5.4B (US) - Shinko (JP)
6. TI $11.4B (US) 16. AMD $5.4B (US) - Kyocera (JP)
7. Toshiba $8.5B (JP) 17. Freescale $4.6B (US) - Signetics (KR)
8. Broadcom $8.4B (US) 18. Sony $4.5B (JP) - Chip-Pak (US)
9. ST Micro $7.4 (FR-IT) 19. UMC $4.3B (TW) - ISI (US)
10. Media Tek $7.1B (TW) 20. NVIDIA $4.0B (US) - Corwil (US)
Several Meetings have been held with government officials - in Congress and Agencies, in Industry and in Academia. No responses from the Trump Administration have been received. The gist of substantive discussions is a "sympathetic ear" and some good suggestions; but no call to action. It appears that if anything is to be done, I will have too be the catalyst for it to happen. There are disturbing inputs:
Needed: Industry Commitment and Input
Included in Electronic Manufacturing are: Smart Phones, Mobile, Tablet, 2-in-1s, Desktop and Server Computers, e.g. PCs and Macs, Servers and Mainframes, Semiconductors, Other Components, Medical Electronics including Medical Devices, Audio/Video Equipment. Transportation Equipment including Automotive, Industrial Equipment including Robotics, Telecommunications Infrastructure, Data Communications, Cellular Communications including Smart Phones, Military Electronics, Appliances, Industrial and Home Security and Automation, Cable TV and Internet Infrastructure, Electric Power Infrastructure, Instrumentation and Specialty Electronics.
All of these technologies, many of which were originally developed in the US, have been globalized. In many cases the leading manufacturing footprints for these devices and equipment are now international, many are assembled in China. Typical numbers you hear about electronics manufacturers revenues are: 60-40 or 70-30 International vs. Domestic. Most of the ‘International’ revenues come from foreign operations – some from export. The positive here is that many companies have established strong global market positions, much of it based on foreign manufacturing operations and have provided leadership in their fields.
The negative is that what could have been domestic employment – including those manufacturing lost to OEMs by outsourcing. Over 100,000 electronics jobs have been lost to the U.S., with some segments of the domestic industry now moribund, others emerging as major forces in the future. The sneaker here is that there are many lost pieces of a domestic manufacturing infrastructure that could take years to replace. Equipment would be relatively easy. Qualified workers would be more difficult in a robotics revolution involving high tech programmed manufacturing.
In fact, there are two poles to this issue, both of which are difficult for domestic manufacturing:
High Tech Automation evolving to “Lights Out” Manufacturing. Major Issues: Capital Investment in an ‘OEM-Outsourced’, Thin Margined Contract Mfg. Marketplace; and with few workers trained to operate advanced robotics equipment and processes.
There has been a long history of new developments and internecine cannibalization in the electronics industry. Perhaps more than any other industry, electronics seems to reinvent itself every 5 years or so, crushing formerly dominant technologies. Outsourcing of manufacturing has been a major recent sea change, begun in the 1990s with IBM PC production. It is now pervasive in the U.S. Very few large OEMs now manufacture t their own products. This is because outsourcing has advantages to the Original Equipment Manufacturer: lower or no capital investments in equipment, significantly lower employee head counts - and ability to re-direct resources toward engineering and marketing. This has undoubtedly helped U.S. companies be globally competitive and gain market penetration in other regions. How would this change with U.S. mfg?
For the contract manufacturer, often called an EMS firm, (Electronic Manufacturing Services manufacturer), they get to focus on manufacturing - often on a global scale. The top EMS firms have huge factories in China, some in SE Asia and Vietnam. However, this is a high volume, low margin business - or with low volume, high mix and quick turn. There is typically limited RD&E investment capability with EMS firms. They are dependent on suppliers and OEM direction to upgrade equipment and processes. It seems Asian firms have specialized in these contract manufacturing businesses, particularly board assembly - working hard and diligent, turning on a dime, with few layers of management, but often 1,000s of factory employees doing bench assembly of electronic parts. Foxconn, HQed in Taiwan, is doing most assembly work in China, employs 1 miliion factory workers in sprawling Chinese plants, with dormitories, soccer fields, (and nets over windows to prevent jumpers in this relentless, repetitive assembly manufacturing process). Foxconn is now a &130B company looking to expand into its own product technologies in displays and other products - and a move into India. They are also sidling up to US manufacturing and have a $10M robotics program with Carnegie-Mellon University.
Product life cycles also come into question: They are short in electronics, perhaps 1-2 years before a major re-tooling is required. Flexible manufacturing techniques can overcome some of this, and future robotic assembly processes may be more-easily reprogrammed. This is an area where many contract suppliers may not have the expertise or financial capability to encompass the coming robotic paradigm shift, and many OEMs may be caught off guard after having abandoned manufacturing years before. The big EMS firms would get bigger.
Going back in history, IBM may have been the one to start it all. Being the visionaries they were at that time, IBM was the first to outsource manufacturing on a large scale. They did it with their PC business, using SCI in Atlanta GA first. Later they spun off their Toronto manufacturing facility to what is now the EMS firm Celestica. More recently, IBM sold off the PC business to Lenovo in China, and this past year did the same with their ‘X86’ Server business. After 64 years in Louisville, KY, GE finally unloaded its appliance business to Haier of China. Before that a deal struck with Electrolux AB fell through due to antitrust concerns. There are many other less published incidents like this, where US companies have given up the ghost on manufacturing a product. The US TV industry was decimated by Japan, but is now making a nascent recovery with merchant LCD systems outsourced to Asian manufacturers.
Are these incidents and trends just examples of needed globalization and offloading of unprofitable businesses - or do they signify the possibility that something is amiss with our domestic commitment to manufacture and sustain products, industries, et al? And while in the GE case, Haier says they will continue the GE site, can you imagine how many suppliers will be affected, from sheet metal to electric motors to tool and die makers? In other areas of the industry, this has been going on for a long time now - and doesn’t seem to be turning around.....
However, there is talk and some action. Some companies are attempting to ‘Insource’ their manufacturing due to logistics and quality issues. There is no major influx yet, but there are signs: Apple has made some moves involving its Mac Pro and potentially other products. Lenovo, with its IBM roots in N.C. is opening a plant there. The domestic automotive industry uses domestic and Mexico-based manufacturing, but they do use a lot of international components. Foreign implants: Toyota, Honda et al use mostly foreign parts, e.g. Japan.
Starting with the Semiconductor Industry:
This industry has globalized, yet has its Roots and Technology Foundation here
in the U.S. It has become more Concentrated, particularly with Intel’s Dominant
Market Share in Logic. In Memory, Manufacturing is Dominated by Asian
Suppliers plus US Micron and SanDisk. In the Developing: ‘More than Moore’
or ‘Beyond Moore’ world there is a Major New U.S. Opportunity to innovate
at the chip, package and process areas.
The Top 20 Semiconductor makers are analyzed below. China has been making moves to acquire US companies, but has been slowed by US government oversight. 11 of the Top 20 were U.S.-Based, which makes the semiconductor industry the crown jewel of US manufacturing. Why? The process produces hundreds to thousands of exact replicated circuits on a single silicon wafer, with feature sizes approaching the atomic domain.
The semiconductor deposition process is primarily planar. As feature sizes reach molecular levels, the industry is moving to 3D circuits - primarily stacked die with wire bonds or via structures. This is occurring in logic and memory and will become more pervasive until a solutions are perfected to do multilayer microcircuits, or a new 'Beyond CMOS' process is developed to bridge the gap signifying the end of 'Moore's Law'. It is imperative that US manufacturers are at the forefront of these developments which will require billions in research funding.
At the same time there is a technology movement toward Silicon Photonicscircuitry - first hybrid and then monolithic to support high performance computing, communications, and Big Data installations. The IPSR consortium mentioned on our home page is roadmapping this effort for: a) System Interconnects, b) Packaging, c) Assembly, d) Monolithic Integration, e) Optical Sensors, f) The Internet Of Things (IoT). The goal is 128TB data communications capability within 5 years, utilizing single mode in-system fiber, polymer waveguides and advanced SiPhotonics chip technology.
The semiconductor industry may hold one key to the future of US manufacturing dominance. It will provide complete system-level circuits on a silicon die within a decade, using its photolithography, diffusion and packaging capabilities to create whole new applications on a chip, i.e. SiP or SoC. Domestic employment here will be high tech - and limited. The bulk of this new employment will be upstream from manufacturing: designers, firmware engineers, application engineers, supply chain and marketing.
More, soon.....including Robotics and Lights Out Manufacturing
Intel CPU Socket 1155 Wikipedia.org Single Mode Optical Fiber: Coming to a Computer System near You
Proposition: An Electronics Manufacturing Renaissance in the U.S
This from EE Times. It confirms meeting I was present recently with the National Director of Intelligence:
SAN FRANCISCO — Junko Yoshida, 12/13/16: "There has been much buzz in the global semiconductor industry about the accelerated consolidation of chip vendors. But the biggest untold story this year is the presence at the negotiating table — in almost every M&A deal — of Chinese investors, or U.S.-based private equity funds whose money can be traced back to China. In deals involving Marvell Technology, Micron, Atmel, Anadigics, Micrel, Pericom Semiconductor, PMC-Sierra, Lattice Semiconductor, Western Digital, and more, Chinese bidders have lurked behind practically every attempted — or speculated — negotiation in the last two years".
For the full report go here:
Electronics is certainly not the only industry with domestic manufacturing issues, but it is one of the highest tech. We DO need to compete in a global marketplace, penetrate foreign markets, be respected global citizens in our industries - but not at the expense of hollowing out our own manufacturing infrastructure, and exposing the U.S. to potential disaster in the event of a prolonged international supply disruption. That is exactly what we have done! This is a complicated subject with various parameters moving in different directions, powered by individual companies' differing business strategies and parochial concerns. Currently, with China's slowdown, Imported goods have flattened. Whether this is a trend or blip, we don't yet know. Some companies are domestic and/or under the radar screen in market niches, others are global with significant (and growing) percentages of their business over seas. A major challenge is how you steer this massive heterogeneous vessel more toward domestic manufacturing without disrupting the global positions of hundreds of companies. A recent short but hopefully continuing dialog with Professor David Autor, Associate dean of MIT Economics Dept. resulted in the following: I said we need to regain high volume manufacturing as a society. Dr Autor observed in a dispassionate non-editorial position, that a rebound in domestic manufacturing mat not result in a large increase in the workforce. He observes that future manufacturing will require both very high skill positions and capital intensive automation, and that 'Rehoring' is not likely to "rejuvenate labor-intensive production such as we formerly saw" the flight "in textiles, furniture, leather goods", and now my input: Electronic assembly can be more highly automated, and one US company is developing that capability now. However - it IS capital intensive and will need millions in investment to bring it to market. And what if, like in many other cases, the market is still in China or other Asia-Pacific, and they are still leveraging low cost labor?
IN ADDITION - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -