On Thursday, November 21, ITIF held an event on Capitol Hill asking Are Advancements in Computing Over? The Future of Moore’s Law. Gordon Moore’s revolutionary observation/prediction in 1965 that the number of transistors on a chip would double every 12-18 months (and thus roughly so would computer processing speeds), has proven prescient. Indeed, over the past forty years, processing speeds have increased over 1 million-fold, unleashing a wave of innovation across industries ranging from ICT and life sciences to energy, aerospace, and services, thus playing a transformational role in driving the global economy and improving quality of life for citizens around the world. Semiconductors (i.e., integrated circuits) constitute the bedrock technology for the entire ICT industry, and annually support an ancillary $1 trillion in electronics-based products—everything from mobile phones and automobiles to medical devices.
Yet—possibly as soon as 2020—the dominant silicon-based CMOS semiconductor architecture will likely hit physical limits (particularly pertaining to heat dissipation) that threaten to compromise Moore’s Law unless a leap can be made to radically new semiconductor chip architectures. This is one of the most critical technology issues the world faces today, because without significant investment in research it’s likely that Moore’s Law will stop before next-generation technologies are available at commercial scale. If so, the negative consequences of a slow-down in Moore’s Law would be enormous, for new needed innovations in robotics, intelligent machines, data analytics, and defense technology all require Moore’s Law’s progress to continue.
Thus, foundational innovation in semiconductor electronics will be needed in both the public and private sector to insure computing power continues to advance and promote our future digital economy. Companies such as IBM, Intel, and Texas Instruments have invested billions in next-generation semiconductor technologies, but they can’t do it all themselves, especially in terms of funding risky, earlier-stage research. U.S. government funding of fundamental scientific research has been instrumental to the United States’ leadership in semiconductor technologies, and continued robust investment is needed now more than ever because of the urgent need to push beyond CMOS with research into new chip architectures leveraging spintronics, other advanced technologies based on nanotechnology, and even quantum computing. Put simply, sustained and expanded federal funding of long-term research will be critical to sustaining Moore’s Law.
Unfortunately, as Robert Colwell, Director of the Microsystems Technology Office at the Defense Advanced Research Projects Agency (DARPA) noted at ITIF’s event, the recent sequestration has compromised funding for a number of semiconductor research teams at universities across the county. Unfortunately, this comes at a time when our global competitors are doubling down on their investments in semiconductor research. As Professor Sanjay Banerjee, Director of the Microelectronics Research Center at the University of Texas, noted, European Union (EU) research initiatives are funding semiconductor R&D “an order of magnitude higher than what we see in the United States,” with the Europeans investing €10 billion ($13.4 billion) over ten years to work on high-tech semiconductor research. Clearly, other nations are putting in the money to win in semiconductor research, and the United States must do more. This should include both increased National Science Foundation funding for basic scientific research in the physical sciences, materials, and engineering fields but also increased co-funding for public-private partnerships such as the Semiconductor Research Corporation, which pools industry and DARPA funding to support research into relevant semiconductor and transistor research at universities, such as the Semiconductor Technology Advanced Research Network, the Nanoelectronics Research Initiative, and the Energy Research Initiative.
We also need to focus on training more semiconductor engineers. As Intel’s Mark Bohr noted, 35 years ago, a team of five or six engineers with a Masters Degrees was sufficient to develop a new generation of semiconductor process technologies, but today it takes several hundred engineers with Ph.D.s to develop each new generation of process technology. Yet fewer and fewer of those engineering Ph.D.s are U.S. students, as other nations are graduating far more engineers than the United States. We must recognize that engineering is vitally important to the future of U.S. economic health and do more to support STEM education. As ITIF writes in 25 Recommendations for the 2013 America COMPETES Act Reauthorization, policymakers should take a number of steps to expand engineering education in the United States, including increasing funding for joint government-industry STEM Ph.D. fellowships; supporting the expansion of specialty Math and Science High Schools; and supporting the designation of 20 U.S. manufacturing universities.
Failing to lead the next-generation of semiconductor technology and carrying Moore’s Law forward would have severe consequences both for the U.S. economy and national security. DARPA’s Colwell noted that integrated circuit advances are critical to defense, citing numerous cases, such as air combat, where U.S. superiority depends on systems and platforms that have 100 times more the computer processing power than our adversaries. Yet winning in semiconductors isn’t important just for defense superiority, but also for the economic benefits from a high-wage industry that both supports direct exports and underpins the broader U.S. ICT sector. In fact, U.S. firms are the world’s largest producers of ICT goods and services, holding a 26 percent share of the global ICT industry. And, from 2005-2009, semiconductors were the number one product export from the United States on an aggregate basis, with exports totaling $48 billion ($10 billion more than automobile exports).
In summary, semiconductor-enabled innovation has been instrumental to U.S. economic and defense leadership over the past four decades, but the dynamic at the heart of the industry, Moore’s Law, is under threat as current technologies meet their physical limits. To sustain Moore’s Law we must not only maintain but increase federal investment in both fundamental basic scientific and applied research while also supporting advanced engineering education and programs.