J. Tully, P. Middleton, R. Ball, P. O'Donovan, S. Bruederle Strategic Analysis Report 30 May 2003 Hype Cycle for Semiconductors, 2003 New semiconductor technologies will achieve widespread usage in the next decade. This will fuel new waves of growth in the semiconductor industry. Management Summary The Hype Cycle for Semiconductors, 2003 charts the evolution of 19 technologies from early-stage research to mass-market use. These include a selection of fundamental transistor and logic devices, process technologies, device architectures and design technologies. All of these technologies enable higher-level system functions. In the next decade, semiconductors will extend beyond the horizons of bulk complementary metal-oxide semiconductor (CMOS) technology, as charted by Moore's Law, into molecule-based devices. In the longer term, devices are likely to become unrecognizable from today's chips. Nanotechnologies based on semiconducting molecules and protein-dna structures (perhaps selfpowered) could transform the semiconductor industry. One thing is clear: Semiconductor and related technologies will continue to be the underlying enabler of IT and electronic products for a number of years to come. Gartner Reproduction of this publication in any form without prior written permission is forbidden. The information contained herein has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy, completeness or adequacy of such information. Gartner shall have no liability for errors, omissions or inadequacies in the information contained herein or for interpretations thereof. The reader assumes sole responsibility for the selection of these materials to achieve its intended results. The opinions expressed herein are subject to change without notice.
30 May 2003 2
CONTENTS 1.0 The Hype Cycle...5 2.0 On the Rise...5 2.1 Protein-DNA Logic...5 2.2 Carbon Nanotubes...5 2.3 Optical Packet Switching...6 2.4 Magnetoresistive Random Access Memory...6 2.5 Network on Chip...6 3.0 At the Peak...7 3.1 Micro Fuel Cells...7 3.2 Inkjet Processes...7 3.3 Electronic System Level Design (Tools and Methodology)...7 3.4 Embedded Programmable Logic...8 3.5 Light-Emitting Polymers...8 3.6 Organic Light-Emitting Diodes...8 4.0 Sliding Into the Trough...9 4.1 Reconfigurable Hardware...9 4.2 Radio-Frequency Complementary Metal-Oxide Semiconductor...9 4.3 Ferroelectric RAM...9 4.4 Microelectromechanical Systems...9 5.0 Climbing the Slope...10 5.1 Semiconductor Intellectual Property...10 5.2 Silicon Germanium Devices...10 6.0 Entering the Plateau...10 6.1 Dense Wave Division Multiplexing...10 6.2 System on Chip...11 7.0 Conclusion...11 Appendix A:Hype Cycle Definitions...12 Appendix B:Acronym Key...13 30 May 2003 3
FIGURES Figure 1. Hype Cycle for Semiconductors, 2003...5 30 May 2003 4
1.0 The Hype Cycle Visibility Electronic System- Level Design (Tools and Methodology) Inkjet Processes Micro Fuel Cells Network on Chip Embedded Programmable Logic Light-Emitting Polymers Organic Light-Emitting Diodes Reconfigurable Hardware Key: Time to Plateau Less than two years Twotofiveyears Five to 10 years More than 10 years MRAM RF CMOS Dense Wave Division Multiplexing System on Chip Optical Packet Switching Carbon Nanotubes Ferroelectric RAM Microelectromechanical Systems Silicon Germanium Devices Semiconductor Intellectual Property Protein-DNA Logic As of May 2003 Technology Trigger Peak of Inflated Expectations Trough of Disillusionment Slope of Enlightenment Plateau of Productivity Maturity Acronym Key MRAM Magnetoresistive random access memory RF CMOS Radio frequency complementary metal-oxide semiconductor Source: Gartner Research (May 2003) Figure 1. Hype Cycle for Semiconductors, 2003 2.0 On the Rise 2.1 Protein-DNA Logic Definition: Logic operations based on the properties of DNA molecules. Time to Plateau/Adoption Speed: More than 10 years. Justification for Hype Cycle Position/Adoption Speed: Research is at an early stage. State machines have been shown in theory. Techniques and commercial potential are unclear. Business Impact Areas: Potential for very high computing performance in small physical space with extremely low energy requirements. 2.2 Carbon Nanotubes Definition: Tiny cylinders of carbon atoms, as small as 10 atoms in diameter. These can have insulating, conducting and semiconducting properties. 30 May 2003 5
Time to Plateau/Adoption Speed: More than 10 years. Justification for Hype Cycle Position/Adoption Speed: Still at the early research stage. Individual nanotubes can be readily fabricated, but major problems remain in their interconnection and in the fabrication of transistor arrays. Business Impact Areas: Potentially huge impact mainly for the period when silicon devices have reached their minimum size limits in 10 to 15 years. Selected Vendors: IBM. 2.3 Optical Packet Switching Definition: The switching of subwavelength optical signals, including "de-multiplexing" and multiplexing, entirely in the optical domain. Time to Plateau/Adoption Speed: Five to 10 years. Justification for Hype Cycle Position/Adoption Speed: Requires a lot more development. Optical switching between fibers is reasonably established but it requires a conversion to the electrical domain for multiplexing and de-multiplexing. Business Impact Areas: Requires much more development. The technology for optical switching wavelengths between fibers is reasonably established. Multiplexing and de-multiplexing signals, whether on or off a single wavelength, still require processing in the electrical domain. Selected Vendors: JDS Uniphase and Vitesse Semiconductor. Analysis by Peter Middleton 2.4 Magnetoresistive Random Access Memory Definition: Pioneered by IBM and Motorola, magnetoresistive random access memory (MRAM) is a fast, low-power, nonvolatile memory technology. It will be the first integrated circuit based on the magnetic properties of atomic spin. Time to Plateau/Adoption Speed: Five to 10 years. Justification for Hype Cycle Position/Adoption Speed: MRAM will begin breaking into the flash market within the next two years. In the next five to 10 years, it has the potential to displace dynamic RAM in a number of applications. Business Impact Areas: All memory-driven applications. Initially, personal electronic devices later, PCs. Selected Vendors: IBM and Motorola. 2.5 Network on Chip Definition: On-chip distributed computing resources that are configured as a computer network. Time to Plateau/Adoption Speed: Five to 10 years. 30 May 2003 6
Justification for Hype Cycle Position/Adoption Speed: At sub-60 nanometer (nm) process dimensions, increased electrical noise and cross talk will render on-chip bus structures useless. Computing resources will need to be interconnected by an Internet-style network with an appropriate error-correction protocol. Business Impact Areas: High impact across all applications. Selected Vendors: IMEC and Intel. 3.0 At the Peak 3.1 Micro Fuel Cells Definition: An alternative power source to batteries for mobile devices. They have the potential to provide 10 times the energy of lithium ion batteries (see "Micro Fuel Cells Power Mobile Devices," T-18-6478). Time to Plateau/Adoption Speed: Five to 10 years. Justification for Hype Cycle Position/Adoption Speed: Working prototypes are available. Large manufacturers are aiming for commercialization by 2005; others are partnering with device manufacturers. Business Impact Areas: Equipment usage life. Facilitates more-powerful mobile devices and applications. Selected Vendors: Casio, NEC, Sony and Toshiba. Analysis by Rafe Ball and Jim Tully 3.2 Inkjet Processes Definition: Depositing transistors onto a flexible substrate using an inkjet-style process. Time to Plateau/Adoption Speed: Five to 10 years. Justification for Hype Cycle Position/Adoption Speed: Inkjet processes were initially used in the assembly of traditional semiconductor dies for large-screen display and solar cell applications. Later developments will involve inkjet-type processes in which polymer transistors are directly printed onto substrates. Business Impact Areas: Applications in woven clothing and radio frequency identification tag applications. Selected Vendors: Alien Technology, CTD Technologies and Philips Semiconductors. 3.3 Electronic System Level Design (Tools and Methodology) Definition: Software tools and methods for the high-level definition of a complete system (hardware and software). The tools should generate individual hardware and software descriptions for implementation by lower-level tools. Time to Plateau/Adoption Speed: Twotofiveyears. Justification for Hype Cycle Position/Adoption Speed: Many attempts have been made to provide electronic-system-level design tools. However, these attempts have largely failed to generate software descriptions at the appropriate levels. Business Impact Areas: All applications. 30 May 2003 7
Selected Vendors: Cadence Design Systems, Mentor Graphics and Synopsys. 3.4 Embedded Programmable Logic Definition: Programmable logic blocks that are embedded within a semiconductor device such as an application-specific integrated circuit or an application-specific standard product. Time to Plateau/Adoption Speed: Twotofiveyears. Justification for Hype Cycle Position/Adoption Speed: Very few vendors are developing embedded programmable logic. The adoption is constrained by the availability of architectural exploration and similar tools to make use of this technology. Business Impact Areas: High impact. Will potentially enable devices to be designed for use in a wider range of applications, due to the flexibility of the programmable logic. 3.5 Light-Emitting Polymers Definition: Light-emitting polymers (LEPs) are based on long-chain polymers that fluoresce when a current is applied. Using inkjet technologies, LEPs can be "printed" onto practically any substrate to form a display of light-emitting pixels. Time to Plateau/Adoption Speed: Twotofiveyears. Justification for Hype Cycle Position/Adoption Speed: The technology of LEPs is still in development. The life of some colors is limited, making large-screen applications impractical. Business Impact Areas: The potential for applications using ultra-thin, low-voltage and potentially flexible displays is enormous. Selected Vendors: CDT Technologies and Philips Semiconductors. Analysis by Paul O'Donovan and Jim Tully 3.6 Organic Light-Emitting Diodes Definition: Organic light-emitting diodes (OLEDs) use short-chain molecules that are vacuum deposited onto a transparent substrate, such as glass. Applying low voltage across the matrix stimulates each pixel to transmit light. OLEDs require no backlight (unlike liquid crystal displays), so power consumption is low. Time to Plateau/Adoption Speed: Twotofiveyears. Justification for Hype Cycle Position/Adoption Speed: Further development is required for longer-lasting colors and full-motion video. Business Impact Areas: Low-voltage, full-color displays will have a significant impact in mobile electronics. Selected Vendors: CTD Technologies and Philips Semiconductors. Analysis by Paul O'Donovan and Jim Tully 30 May 2003 8
4.0 Sliding Into the Trough 4.1 Reconfigurable Hardware Definition: The ability of a semiconductor device to perform different functions at different times not through software changes, but rather by reconfiguring the hardware. Time to Plateau/Adoption Speed: Five to 10 years. Justification for Hype Cycle Position/Adoption Speed: A lack of tools is the main reason for the long time to adoption. May use embedded programmable logic. Business Impact Areas: High impact. However, users don't know how they would handle reconfigurability at this time. 4.2 Radio-Frequency Complementary Metal-Oxide Semiconductor Definition: Using standard complementary metal-oxide semiconductors (CMOSs) for microwave radio frequency (RF) applications of 5GHz and above. Justification for Hype Cycle Position/Adoption Speed: Some vendors are already offering CMOS-receive functions at these frequencies. Business Impact Areas: High impact. Will eliminate the need for different processes, such as bipolar complementary metal-oxide semiconductor (BiCMOS) and gallium arsenide, thereby reducing costs and chip counts. 4.3 Ferroelectric RAM Definition: A nonvolatile memory with high write speed that is based on arrays of ferroelectric crystals. Justification for Hype Cycle Position/Adoption Speed: It is already available. Business Impact Areas: Moderate impact across a wide range of applications. 4.4 Microelectromechanical Systems Definition: Semiconductor devices incorporating structures that can physically move, in addition to electronic circuits. Justification for Hype Cycle Position/Adoption Speed: Some applications are already in widespread use (for example, airbag accelerometers, inkjet nozzles). Business Impact Areas: High impact across a wide range of applications. Significant long-term impact for biomedical applications. 30 May 2003 9
5.0 Climbing the Slope 5.1 Semiconductor Intellectual Property Definition: Pre-designed functional blocks that are licensed for use in a chip. Could be microprocessors, memory or many other functions. Justification for Hype Cycle Position/Adoption Speed: Already widely used in system on chip (SOC) devices. The industry still has trouble finding the optimum business model for intellectual property. Business Impact Areas: High impact. Complex SOC devices can't be designed without a large-scale use of intellectual property blocks. Selected Vendors: ARM, Rambus and Synopsys. 5.2 Silicon Germanium Devices Definition: Heterjunction BiCMOS technology made by adding germanium to the base of the bipolar transistor. Justification for Hype Cycle Position/Adoption Speed: Becoming widely used in mobile phones, fiber-optic communications and test equipment. Business Impact Areas: Moderate impact, mostly in wireless and fiber-optic transceiver functions in the previously listed applications. Selected Vendors: Atmel, IBM, Infineon Technologies, Maxim Integrated Products, Motorola and Texas Instruments. Analysis by Stanley Bruederle 6.0 Entering the Plateau 6.1 Dense Wave Division Multiplexing Definition: An optical-fiber-encoding technology that enables communication signals on several wavelengths to be transmitted simultaneously on a single fiber. Justification for Hype Cycle Position/Adoption Speed: Established use in optical transport networks worldwide. Business Impact Areas: Significantly increases bandwidth over time division multiplexing systems. Selected Vendors: Alcatel, Cisco Systems, Lucent and Nortel Networks. Analysis by Peter Middleton 30 May 2003 10
6.2 System on Chip Definition: A device containing compute engine, memory and user logic on a single chip. Justification for Hype Cycle Position/Adoption Speed: Widely used in many applications such as mobile phone handsets, video game consoles, digital video, graphics and storage. Business Impact Areas: High impact. Product cost and performance requirements could not be achieved without SOC. Selected Vendors: IBM, Intel, STMicroelectronics and Texas Instruments. 7.0 Conclusion Semiconductor technology is changing fast. Many new technologies will become feasible and achieve widespread usage in the next several years. Enterprises should be aware of the opportunities and risks that these changes present. 30 May 2003 11
Appendix A: Hype Cycle Definitions Technology Trigger: A breakthrough, public demonstration, product launch or other event generates significant press and industry interest. Peak of Inflated Expectations: During this phase of overenthusiasm and unrealistic projections, a flurry of well-publicized activity by technology leaders results in some successes, but more failures, as the technology is pushed to its limits. The only enterprises making money are conference organizers and magazine publishers. Trough of Disillusionment: Because the technology does not live up to its overinflated expectations, it rapidly becomes unfashionable. Media interest wanes, except for a few cautionary tales. Slope of Enlightenment: Focused experimentation and solid hard work by an increasingly diverse range of organizations lead to a true understanding of the technology's applicability, risks and benefits. Commercial, off-the-shelf methodologies and tools ease the development process. Plateau of Productivity: The real-world benefits of the technology are demonstrated and accepted. Tools and methodologies are increasingly stable as they enter their second and third generations. The final height of the plateau varies according to whether the technology is broadly applicable or benefits only a niche market. Approximately 30 percent of the technology's target audience has or is adopting the technology as it enters the Plateau. Time to Plateau/Adoption Speed: The time required for the technology to reach the Plateau of Productivity. 30 May 2003 12
Appendix B: Acronym Key BiCMOS CMOS LEP MRAM OLED RF SOC bipolar CMOS complementary metal-oxide semiconductor light-emitting polymers magnetoresistive RAM organic light-emitting diode radio frequency system on chip 30 May 2003 13