Browse by year:
The changing landscape of semiconductors
Vinod Dham
Wednesday, October 14, 2020
The Pune-born Vinod Dham grew up to be renowned as ‘Father of the Pentium Chip’. Vinod graduated engineering at the age of 21 in 1971, and he joined Teradyne Semiconductor, the only semiconductor company existed then in India. The company was located in New Delhi. Past four years in 1975, he migrated to the U.S with just $8 in his pocket. Post which there was no looking back. Vinod became a chip engineer and contributed to inventing Intel’s first flash memory chip. He further continued to manage Intel’s microprocessor projects that include the Pentium chip which elevated the company to be the world’s biggest chipmaker. Later, Vinod joined NexGen and Advanced Micro Devices. Later, he transformed into a venture capitalist initially at NewPath Ventures and later at NEA-IndoUs Ventures.

The semiconductor industry is not only among the largest manufacturing industries in the world, but also, arguably, the most important. Having enabled personal computing, mobile communications and Internet revolution over the last three decades, it is now ready to penetrate every fabric of our lives. In addition, there has been a stepwise increase in the silicon content of electronic systems with every passing generation.

The semiconductor industry, however, is going through fundamental changes. After proceeding down the same path for the past 35 years, we're on the threshold of a new semiconductor era. In the past two decades, the CMOS, scaling limits have been projected and then been defeated several times. However, going forward as we migrate to 90 nanometer and beyond, the technological issues that start cropping up are so huge that, even though innovation will continue to find a way to solve those issues, the economics of the cost to scale below 90 nanometer is beginning to exceed Moore's law* potential benefit. This will limit the number of suppliers of leading edge chips at 65nm and beyond to merely a handful.

In the future, the increasing cost and risk of leading edge design will be so costly that only highest volume application will take advantage of it. Chip designers will avoid the leading edge unless performance is absolutely necessary. The bulk of semiconductor consumption, especially in the cost- sensitive emerging Asian economies of China and India, will be centered on the optimum design node of 90 nanometer for the next decade.

The migration of wafer size from 200 mm to 300 mm and the consequent reduction in die size with migration from 130 nm to 90 nanometer will result in dramatic increases in chip count. In addition, chip production in China is slated to increase by 40 percent annually, much faster than the world average of 10 percent. Overall, this capacity increase coupled with a slowing of the technology migration treadmill will result in a perfect storm.

The low capital cost of manufacturing mature semiconductor technologies will spur growth of specialty designs to create inexpensive differentiated consumer products to meet the growing needs of a billion Asian consumers. With the rapid growth of Asian economies, semiconductor consumption is forecasted to continue the migration from the Americas to Asia-Pacific. These consumers, with varying affordability, will demand a broad range of functionality and price performance, thus creating a need for more designs.

The challenge for the semiconductor industry is to understand the dynamics of this mass market, which will require shifting the mindset from selling chips to selling applications. A different approach may be needed to create the perceived value in the consumer business, where customers pay for value and not technology. The consumer business is characterized by short life cycles, along with highly segmented and personalized products.

Although the devices that combine entertainment, mobile communications and computing are emerging as likely engines for the industry's near-term growth, healthcare, energy and security are some emerging areas ripe to benefit from the maturing semiconductor life cycle. Bio technology development's accelerated pace, led by use of powerful and mature IT technologies which favor demographics of the aging European and American population, will create further opportunities for designing shrinking semiconductor chips- the 'Bio Chips' for the Healthcare industry. Our oil and gas reserves are quickly dwindling. With the growth of Asian economies, our energy needs will surpass new applications and usage to develop alternative energy sources at affordable costs. New designs and improvements in 'solar cells' can especially benefit India due to its high sunlight availability. Security is now a higher priority with the threats of terrorism. An enormous amount of new designs and applications are needed in areas, ranging from Biometrics which develop safeguards for protecting our water and food supplies, to embedding low cost semiconductor sensors in intelligent dust. This trend will accelerate and explode further in the new millennium as the semiconductor devices become ubiquitous and get embedded in every fabric of our lives to make us more productive, safer and longer.

These trends will have a profound impact on semiconductor chip design sourcing and services. India and China, with their low cost structure, will be the biggest beneficiary of sourcing these needs. The trend will be to integrate intellectual property from various sources, like customers, vendors and third parties, to create system level solutions on a piece of silicon. Thus, the industry will increasingly need to focus more on moving away from pure components to platforms. Platforms are application-specific standard products that are programmable for a sub set of high-volume applications. This is already beginning to happen with the cell phone platform in China, which is delivered by the ODMs and uses standard off the shelf, high volume standard semiconductors customized through software.

In last five years, chip design activity has skyrocketed in India, and over the next decade a core competency in chip design will be developed, like it has today in IT software. India needs to be prepared to take advantage of emerging chip design needs of the next decade. We should include substantial preparatory material on VLSI technology and design in our engineering programs at the Bachelor's degree level. There is opportunity to train engineers in design tools and methodologies by training institutes similar to how NIIT did for the IT software industry. We should encourage development of intellectual property and cross licensing, because Indian engineers are strong in architecture and software. In this mature semiconductor environment, software will be more important than the silicon. India is well positioned to use its dominant software prominence as a strategic weapon to build a huge chip design franchise.

With increased focus on mobility and connectivity, there is a huge need for delivering very low power chip designs for emerging consumer needs. This is especially challenging at 90 nanometers and beyond. The transistors' leakage has increased by several orders of magnitude requiring innovation of new low power architectures and power management techniques to address this critical issue.

The cost of a new, leading-edge fab is soaring out of control. Today it costs $3-4 billion to build a 90-nanometer 300 mm fab with capacity to process a thousand wafers per day. At 90 nanometer and beyond, the technologies are also becoming more expensive as they require huge innovations in materials and equipment to achieve the desired scaling. New improvements in techniques such as ’straining’ the lattice silicon to improve the mobility of charge carriers, high-k gate dielectrics, metal gate electrodes and fully depleted silicon-on-insulators are being undertaken to wring incremental performance. ’Many of the exponential trends are approaching limits that require new means for circumvention if we are to continue the historic rate of progress,’ said Gordon Moore in a famous keynote speech delivered at the 2003 International Solid-State Circuits Conference. We have reached the point where very few companies can build their own factories. The foundries may have an inherent advantage over captive fabs, because Foundries can fill their fabs with products from a number of designers. Captive fabs require increasingly scarce high-volume products for more effective operations. This will force more radical disintegration of the semiconductor supply chain. In this new paradigm, Wafer Foundries and their customers are forming closer collaborative partnerships. India’s chip design industry must establish closer relationship with these mega foundries in Asia. These partnerships are necessary because of leading edge technologies, where the design and manufacturing will need to work more tightly for end product delivery.

The landscape of the semiconductor industry in the next decade will be shaped by a number of converging scenarios. In the maturing semiconductor environment, where revenue growth seems to be slowing while new, more fundamental technological barriers and financial challenges of developing shrinking geometries are rapidly escalating, we will face formidable obstacles.
Father of the Pentium, Vinod Dham has spent the last three decades in the semiconductor industry. Following successful stints at Intel, AMD, he went on to head a startup’Silicon Spice’that was acquired by Broadcom. Dham is a co-founder and Managing Partner of New Path Ventures, which aims to create leading-edge technology companies by leveraging cost effective development and implementation strengths of India.

Share on LinkedIn