America’s lead in advanced computing is almost gone

Abstract

American usage of advanced computing to solve commercial and government issues has a long and august history that ranges from weather and climate simulations to drug discovery and protein folding simulations.  Since computing’s early days, Americans have been at the forefront, outpacing foreign rivals and providing a competitive advantage for U.S. businesses. But American computing leadership is not a birthright; it must constantly be re-earned and, at this moment, it is in jeopardy. In nearly every major aspect of computing, from hardware manufacturing to algorithm design to advanced computing systems, American leadership is waning or already gone.  Instead, other countries, particularly China, are ascendant.  At this point, only large, sustained, and coordinated investments can prevent America’s loss of leadership.

Introduction

America’s lead in advanced computing is nearly gone, putting in jeopardy the many dividends that leadership has historically provided.  In this article on systems and capabilities, and a follow-on article on hardware, we show how this lead has diminished, and discuss what might be done to reverse this trend.

When a competitor has the advantage of cutting-edge microchips, better algorithms, and more computing power, they can out-compete others. Designing and developing new products more quickly and cheaply, developing innovations that would otherwise be impossible due to cost and technical barriers, and making more scientific discoveries in less time translates into greater global competitiveness and impact. 

Historically, U.S. multinationals have leveraged information technology (IT) to become more productive than their European counterparts.  A 2012 analysis showed that about half of the differences in productivity growth between the two from 1994 to 2004 can be explained by this effect (Bloom et al., 2012), which is consistent with the finding that IT accounts for one-third of all productivity gains in the U.S. since the 1970s (Byrne et al., 2013).

Advanced computing is fundamental to national security, from intelligence gathering and military power, to the cybersecurity of financial markets and critical infrastructure. Autonomous systems and widespread use of artificial intelligence across military systems – both underpinned by advanced computing power, leading-edge microchips, data, and algorithms – have the potential to create major discontinuities in national defense. U.S. military advantage will increasingly rely on maintaining the country’s edge in advanced computing (ERI, 2018).

United States’ leadership in advanced computing is now being challenged by other countries – especially China, which claims that it will spend $1.4 trillion on advanced technologies through 2025 (Bloomberg, 2020).  In recent decades, China has risen from an uncompetitive also-ran to parity in many aspects of advanced computing and has the specific goal of toppling the U.S. as the global leader in the hardware-focused technology sector (The Economist, 2021). The Chinese vice-premier Liu He told a group of top scientists that “for our country, technology innovation is not just for growth, […] it is a matter of survival.” (Bloomberg, 2021). 

We conducted a survey of American leaders in advanced computing (details in Appendix A) that underlined the competitive importance of having a computing advantage and shows how America’s leadership is faltering.  Detailed survey responses from 40 of the ~250 biggest computing users in the U.S. reveals that more than 55% of organizations believe that an advantage in computing translates into a disproportionately large competitive advantage (Figure 1). Simply put: out-computing leads to out-competing rivals.

Concurrently, our survey shows that China has closed the computing capabilities gap with America. Of our American respondents, 42% believe that their computational capabilities have been outpaced by their Chinese competitors, about the same level as their Japanese competitors (44%) and considerably higher than those in Europe (31%) (Figure 2). However, this near-parity hides a more alarming outlook for the future: 79% of American respondents believe that their Chinese competitors are improving capabilities faster.

The speed of the Chinese improvement is also expressed in the digital-riser rankings published by the World Economic Forum which scores how countries’ digital competitiveness has developed over the last three years. Across ten dimensions, China gained an average of 21 spots, while the U.S. lost an average of 7 (Meissner, 2021). 

In this article, we characterize how the field of advanced computing is evolving by specifically examining US-China dynamics in three areas central to advanced computing.  In each case, we see America’s lead disappearing, suggesting a broad threat to U.S. competitiveness. 

Pillar I: Advanced Systems 

Since 1993, the TOP 500 index has tracked the most powerful supercomputers. Traditionally, such systems have been employed predominantly for scientific and engineering workloads, but they are increasingly also being used for a wider range of applications (Waters, 2021). Up through 2015, this ranking was a story of U.S. dominance: about half the most powerful supercomputers were in the U.S. and China lagged far behind.  In the past six years, however, China has surged, fielding more supercomputers than the U.S. (Figure 3).  

By other measures, China’s rise has been more contested.  Measured by the total computing power across all of each country’s supercomputers, China did rise to the top of the list in 2016, but the U.S. has since retaken leadership and currently holds a slight lead. Overall, though, the picture is clear: China’s supercomputing is now near parity with the United States.

Failing to grow computing rapidly enough has real consequences for U.S. users. Of our survey respondents, 84% say that they are computationally constrained in running key programs. This result is telling because our respondents are the vanguard of American research enterprises and academic institutions with privileged access to the most advanced national supercomputing resources. This indicates that even those organizations with the most advanced computing systems are experiencing computation bottlenecks that limit them. The good news is that the U.S. is investing to improve our high-end supercomputing via the exascale computing project (ECP, 2022). But, the bad news is that this is unlikely to provide a competitive advantage as China also has been developing exascale systems (Hemsoth, 2021). 

Pillar II: Advanced Algorithms 

Algorithms are step-by-step instructions that computers use to solve problems.  Better algorithms allow users to harness computing power more efficiently and solve more challenging problems. Historically, the U.S. has led in developing new computing algorithms, generating two-thirds of all the major improvements (Thompson et al., 2021). Initially, this was dominated by U.S.-born inventors, but, in recent decades, it has depended on America’s ability to attract talented researchers to American universities and businesses. This ability to dominate top-level talent seems in jeopardy as China outpaces the U.S. and many other countries in producing PhDs in STEM fields since 2007 (Nietzel, 2021). Georgetown’s Center for Security and Emerging Technology (CSET) projects that, by 2025, Chinese research institutions will have nearly twice as many PhDs as the U.S. (Zwetsloot et al., 2021).  

One important measure of the frontier of applied algorithms is the Gordon Bell Prize. Since 1987, this prize has been awarded to recognize outstanding achievement in harnessing supercomputers to innovate in science, engineering, and data analytics. Historically, the prize has been dominated by U.S. winners (Figure 4), but China has made enormous headway, and in the past five years has equaled or surpassed U.S. performance.  

Policy Implications

Our research suggests that the U.S.’s lead in advanced computing is nearly gone.  This begs the question: what steps could be taken to try to re-attain leadership?  The passage of the CHIPS and Science Act of 2022  was a significant step, however, there is more to be done.

We recommend that the Office of Science and Technology Policy (OSTP) define initiatives to:

Re-establish the foundations for success:

1. Democratize access to U.S. supercomputing by:
   • Building more mid-tier systems that push boundaries for many users
   • Building tools so that users scaling up their computations can do so with less up-front resource investment
2. Increase the pool of innovators by funding a large increase in the number of electrical engineers and computer scientists being trained 
3. Compete for the best global students with longer-term U.S. residency incentives and scholarships

Better utilize our existing assets:

4. Significantly expand the private sector’s ability to experiment with high-performance computing by creating more opportunities for them to partner with supercomputing sites in academia and national labs, taking inspiration from the INCITE and HPC4Energy (and related programs) model (Berger et al., 2008)

We will cover semiconductor chip design and manufacturing and additional recommendations in Part 2.

About the Authors

Neil Thompson is the Director of the FutureTech research project at MIT’s Computer Science and Artificial Intelligence Lab and a Principal Investigator at MIT’s Initiative on the Digital Economy.  Previously, he was an Assistant Professor of Innovation and Strategy at the MIT Sloan School of Management and a Visiting Professor at the Laboratory for Innovation Science at Harvard. He has advised businesses and government on the future of Moore’s Law, has been on National Academies panels on transformational technologies and scientific reliability, and is part of the Council on Competitiveness’ National Commission on Innovation & Competitiveness Frontiers.

Chad Evans is the Executive Vice President of the Council on Competitiveness.  He is also both Secretary and Treasurer to the Board of the Council on Competitiveness; Treasurer to the Board of the Global Federation of Competitiveness Councils; a member of the Texas A&M Engineering Experiment Station Advisory Board; 2023 Co-Chair of Science Is US; an ARCS Foundation National Science and Engineering Advisory Council member; a U.S. German Marshall Fund Fellow; and a past member of the Lawrence Livermore National Laboratory Industry Advisory Council and the World Economic Forum Advisory Board on Russian Competitiveness.

Daniel Armbrust is co-founder and director of Silicon Catalyst, founded in 2015, which incubates semiconductor startups.  Its portfolio companies have raised more than $400M in venture funding and are valued at over $1.5B.  Armbrust serves as an advisor, board member, board chairman and angel investor for many semiconductor startups.  Daniel is an affiliate with Lawrence Berkeley National Labs and recently was appointed to the Industrial Advisory Committee, which advises the Department of Commerce on the R&D strategy for the CHIPS Act.  He served as President and CEO of the SEMATECH semiconductor consortium and held various positions in semiconductor manufacturing and development over 25 years at IBM. Daniel’s contributions are made in his personal capacity, and should not be attributed to the Department of Commerce, the Industrial Advisory Committee of the CHIPS Act, or the US Government.

Part 1 References

Berger, Marsha et al.  2008.  INCITE report.  Advanced Scientific Computing Advisory Committee – Committee of Visitors.  https://science.osti.gov/-/media/ascr/ascac/pdf/reports/Incite_cov_report_aug_08.pdf.

Bloom, N., Sadun, R., Van Reenen, J., 2012. Americans Do IT Better: US Multinationals and the Productivity Miracle. The American Economic Review 102 (1), 167-201. https://doi.org/10.1257/aer.102.1.167.

Bloomberg, 2020. China’s got a new plan to overtake the U.S. in Tech. Bloomberg News. May 20. https://www.bloomberg.com/news/articles/2020-05-20/china-has-a-new-1-4-trillion-plan-to-overtake-the-u-s-in-tech. (Accessed September 11 2021). 

Byrne, D.M., Oliner, S.D., Sichel, D.E., 2013. Is the information technology revolution over? SSRN Electronic Journal, 20-36. https://doi.org/10.2139/ssrn.2303780.

Calhoun, G., 2021. War With China? The Economic Factor That Could Trigger It. Forbes. September 12. https://www.forbes.com/sites/georgecalhoun/2021/09/12/war-with-china-the-economic-factor-that-could-trigger-it/?sh=549f90c45d26. (Accessed September 15 2021). 

Council on Competitiveness, 2020. BUILD for Advanced Computing – The Final Report in the NSCI: Advancing U.S. Competitiveness through Public-Private Partnerships for Advanced Computing Initiative.”

ERI.  2018.  DARPA announces next phase of the electronics resurgence initiative.  https://www.darpa.mil/news-events/2018-11-01a.

Exascale Computing Project.  2022.  Overview of the ECP.  https://www.exascaleproject.org/about/.  Viewed on May 27, 2022.

Ezell, S., 2021. Moore’s Law under attack: The impact of China’s policies on global semiconductor innovation. Information Technology and Innovation Foundation. February 18. https://itif.org/publications/2021/02/18/moores-law-under-attack-impact-chinas-policies-global-semiconductor

Hemsoth, Nicole.  2021.  China has already reached exascale – on two separate systems.  Oct 26.  https://www.nextplatform.com/2021/10/26/china-has-already-reached-exascale-on-two-separate-systems/

Hillman, J.E., 2021, US at risk of losing cloud computing edge to China. Politico. August 26. https://www.politico.com/newsletters/politico-china-watcher/2021/08/26/us-at-risk-of-losing-cloud-computing-edge-to-china-494105. (Accessed September 09 2021).

Meissner, Philip.  2021.  These countries rank highest for digital competitiveness. World Economic Forum.   Sep 2, 2021.  https://www.weforum.org/agenda/2021/09/countries-rank-highest-digital-competitiveness/

Nietzel, M.T., 2021. U.S. Universities Fall Further Behind China In Production Of STEM PhDs. Forbes. August 7. https://www.forbes.com/sites/michaeltnietzel/2021/08/07/us-universities-fall-behind-china-in-production-of-stem-phds/?sh=8cec40046067. (Accessed September 13 2021).

The Economist, 2021. What tech does China want? The Economist. August 14. https://www.economist.com/business/what-tech-does-china-want/21803410. (Accessed September 04 2021).

Thompson, N.C., Ge, S., Sherry, Y.M., 2021. Building the algorithm commons: Who discovered the algorithms that underpin computing in the modern enterprise? Global Strategy Journal 11 (1), 17-33. https://doi.org/10.1002/gsj.1393.

Waters, R., 2021. Supercomputers to reshape tech landscape. Financial Times. May 20. https://www.ft.com/content/db04daea-d893-4d18-b33b-8b09bb7226af. (Accessed August 29 2021).

Zwetsloot, R., Corrigan, J., Weinstein, E., Peterson, D., Gehlhaus, D., Fedasiuk, R.., 2021. China is fast outpacing U.S. STEM PhD growth. Center for Security and Emerging Technology. August. https://doi.org/10.51593/20210018.

Appendix A

The Advanced Computing Users survey was run jointly by the Massachusetts Institute of Technology and the Council on Competitiveness.  The survey was conducted from November 2019 to December 2019.  The sampled group included 120 organizations, including universities, national labs, federal agencies and industry.  We estimate that this group comprises between one-third and one-half of all the biggest computing users in the United States.  

Survey responses were completed by senior leaders in these organizations, often with support from their technical staff.  One-third of the organizations sampled provided detailed responses.