Report to the President: Ensuring Long-Term U.S. Leadership in

REPORT TO THE PRESIDENT

Ensuring Long-Term U.S.

Leadership in Semiconductors

Executive Office of the President

President’s Council of Advisors on

Science and Technology

January 2017

REPORT TO THE PRESIDENT

Ensuring Long-Term U.S.

Leadership in Semiconductors

Executive Office of the President

President’s Council of Advisors on

Science and Technology

January 2017

About the President’s Council of Advisors on

Science and Technology

The President’s Council of Advisors on Science and Technology (PCAST) is an advisory group of the

Nation’s leading scientists and engineers, appointed by the President to augment the science and

technology advice available to him from inside the White House and from cabinet departments

and other Federal agencies. PCAST is consulted about, and often makes policy recommendations

concerning, the full range of issues where understandings from the domains of science, technology,

and innovation bear potentially on the policy choices before the President.

For more information about PCAST, see www.whitehouse.gov/ostp/pcast

The President’s Council of Advisors on

Science and Technology

Co-Chairs

John P. Holdren

Assistant to the President for

Science and Technology

Director, Office of Science and Technology

Policy

Eric S. Lander

President

Broad Institute of Harvard and MIT

Vice Chairs

William Press

Raymer Professor in Computer Science and

Integrative Biology

University of Texas at Austin

Maxine Savitz

Honeywell (Ret.)

Members

Wanda M. Austin

President and CEO (Ret.)

The Aerospace Corporation

Christopher Chyba

Professor, Astrophysical Sciences and

International Affairs

Princeton University

Rosina Bierbaum

Rosina Bierbaum,

Professor, School of Natural Resources and

Environment, University of Michigan, and

Roy F. Westin Chair in Natural Economics,

School of Public Policy, University of

Maryland

S. James Gates, Jr.

John S. Toll Professor of Physics

Director, Center for String and

Particle Theory

University of Maryland, College Park

Christine Cassel

Planning Dean

Kaiser Permanente School of Medicine

Mark Gorenberg

Managing Member

Zetta Venture Partners

Susan L. Graham

Pehong Chen Distinguished Professor

Emerita in Electrical Engineering and

Computer Science

University of California, Berkeley

Ed Penhoet

Director

Alta Partners

Professor Emeritus, Biochemistry and

Public Health

University of California, Berkeley

Michael McQuade

Senior Vice President for Science and

Technology

United Technologies Corporation

Barbara Schaal

Dean of the Faculty of Arts and Sciences

Mary-Dell Chilton Distinguished Professor

of Biology

Washington University of St. Louis

Chad Mirkin

George B. Rathmann Professor of

Chemistry

Director, International Institute for

Nanotechnology

Northwestern University

Eric Schmidt

Executive Chairman

Alphabet

Mario Molina

Distinguished Professor, Chemistry and

Biochemistry

University of California, San Diego

Professor, Center for Atmospheric Sciences

Scripps Institution of Oceanography

Daniel Schrag

Sturgis Hooper Professor of Geology

Professor, Environmental Science and

Engineering

Director, Harvard University Center for

Environment

Harvard University

Craig Mundie

President

Mundie & Associates

Staff

Ashley Predith

Executive Director

Jennifer L. Michael

Program Support Specialist

vii

PCAST Ensuring Long-Term U.S. Leadership in

Semiconductors Working Group

Co-Chairs

John Holdren*

Assistant to the President for Science and

Technology

Director, Office of Science and Technology

Policy

Paul Otellini

Former President and CEO

Intel

Working Group Members

Richard Beyer

Former Chairman and CEO

Freescale Semiconductor

Ajit Manocha

Former CEO

GlobalFoundries

Wes Bush

Chairman, CEO, and President

Northrop Grumman

Jami Miscik

Co-CEO and Vice Chairman

Kissinger Associates

Diana Farrell

President and CEO

JP Morgan Chase Institute

Craig Mundie*

President

Mundie & Associates

John Hennessy

President Emeritus

Stanford University

Mike Splinter

Former CEO and Chairman

Applied Materials

Paul Jacobs

Executive Chairman

Qualcomm

Laura Tyson

Distinguished Professor of the Graduate

School UC Berkeley

PCAST Staff

Ashley Predith

Executive Director

President’s Council of Advisors on Science

and Technology

Jennifer Michael

Program Support Specialist

President’s Council of Advisors on Science

and Technology

viii

*denotes PCAST member

National Economic Council Staff

Michael Levi

Special Assistant to the President for Energy

and Economic Policy

National Economic Council

Kavya Shankar

Policy Advisor

National Economic Council

Office of Science and Technology Policy Staff

Megan Brewster

Senior Policy Advisor

Office of Science and Technology Policy

Tim Polk

Assistant Director, Cybersecurity

Office of Science and Technology Policy

Lloyd Whitman

Assistant Director, Nanotechnology and

Advanced Materials

Office of Science and Technology Policy

EXECUTIVE OFFICE OF THE PRESIDENT

PRESIDENT’S COUNCIL OF ADVISORS ON SCIENCE AND TECHNOLOGY

WASHINGTON, D.C. 20502

President Barack Obama

The White House

Washington, DC 20502

Dear Mr. President:

This letter transmits a report entitled Ensuring Long-Term U.S. Leadership in Semiconductors,

prepared by a working group of technology industry leaders, eminent researchers, and former

policymakers. The President’s Council of Advisors on Science and Technology (PCAST) has

reviewed and adopted the report. The report assesses the challenges and opportunities facing

semiconductor innovation, competitiveness, and security, and outlines recommendations for action

to address them.

Semiconductors are essential to modern life. Progress in semiconductors has opened up new

frontiers for devices and services that use them, creating new businesses and industries, and

bringing massive benefits to American workers and consumers as well as to the global economy.

Cutting-edge semiconductor technology is also critical to defense systems and U.S. military

strength, and the pervasiveness of semiconductors makes their integrity important to mitigating

cybersecurity risk.

Today, U.S. semiconductor innovation, competitiveness, and integrity face major challenges.

Semiconductor innovation is already slowing as industry faces fundamental technological limits

and rapidly evolving markets. Now a concerted push by China to reshape the market in its favor,

using industrial policies backed by over one hundred billion dollars in government-directed funds,

threatens the competitiveness of U.S. industry and the national and global benefits it brings. The

report looks at these challenges in greater detail.

The core finding of the report is this: only by continuing to innovate at the cutting edge will the

United States be able to mitigate the threat posed by Chinese industrial policy and strengthen the

U.S. economy. Thus, the report recommends and elaborates on a three pillar strategy to (i) push

back against innovation-inhibiting Chinese industrial policy, (ii) improve the business environment

for U.S.-based semiconductor producers, and (iii) help catalyze transformative semiconductor

innovation over the next decade. Delivering on this strategy will require cooperation among

government, industry, and academia to be maximally effective.

Sincerely,

John P. Holdren Eric S. Lander

Co-Chair Co-Chair

Table of Contents

The President’s Council of Advisors on Science and Technology ....................................... v

PCAST Ensuring Long-Term U.S. Leadership in Semiconductors Working Group..... vii

Table of Contents .................................................................................................................................. 1

Executive Summary ............................................................................................................................. 2

1. Challenges and Opportunities .................................................................................................. 4

Evolving Challenges .................................................................................................................. 6

Technology and Markets .................................................................................................... 6

Chinese Industrial Policies ................................................................................................ 7

Responding Strategically ................................................................................................ 10

2. Influencing Chinese Actions ................................................................................................... 13

3. Creating a More Supportive Business Climate in the United States .................. 15

4. Developing a “Leapfrog” Strategy for Continuing U.S. Leadership ..................... 18

Focus Areas ................................................................................................................................. 19

5. Conclusion ........................................................................................................................................ 25

Appendix A. Moonshots – Methodology and Exemplars ............................................... 26

A.1 Methodology ...................................................................................................................... 26

A.1.1 Overall Approach .................................................................................................... 26

A.1.2 Computer System Architectures ........................................................................ 27

A.1.3 Computing Modalities ........................................................................................... 27

A.2 Sample Moonshots .......................................................................................................... 29

A.2.1 Bioelectronics for sensory replacement and implantable

neuro-stimulation for control of chronic conditions. ..................................... 29

A.2.2 Threat Detection Network ................................................................................... 30

A.2.3 Distributed Electric Grid ...................................................................................... 31

A.2.4 Global Weather Forecasting ............................................................................... 31

Executive Summary

Semiconductors are essential to modern life. Progress in semiconductors has opened up new frontiers

for devices and services that use them, creating new businesses and industries, and bringing massive

benefits to American workers and consumers as well as to the global economy. Cutting-edge

semiconductor technology is also critical to defense systems and U.S. military strength, and the

pervasiveness of semiconductors makes their integrity important to mitigating cybersecurity risk.

U.S. semiconductor innovation, competitiveness, and integrity face major challenges. Semiconductor

innovation is already slowing as industry faces fundamental technological limits and rapidly evolving

markets. Now a concerted push by China to reshape the market in its favor, using industrial policies

backed by over one hundred billion dollars in government-directed funds, threatens the

competitiveness of U.S. industry and the national and global benefits it brings.

The global semiconductor market has never been a completely free market: it is founded on science that

historically has been driven, in substantial part, by government and academia; segments of it are

restricted in various ways as a result of national-security and defense imperatives; and it is frequently

the focus of national industrial policies. Market forces play a central and critical role. But any

presumption by U.S. policymakers that existing market forces alone will yield optimal outcomes –

particularly when faced with substantial industrial policies from other countries – is unwarranted. In

order to realize the opportunities that semiconductors present and to effectively mitigate major risks,

U.S. policy must respond to the challenges now at hand.

Our core finding is this: the United States will only succeed in mitigating the dangers posed by Chinese

industrial policy if it innovates faster. Policy can, in principle, slow the diffusion of technology, but it

cannot stop the spread. And, as U.S. innovators face technological headwinds, other countries’ quest to

catch up will only become easier. The only way to retain leadership is to outpace the competition.

That does not mean that the U.S. government should be silent or passive in the face of Chinese

industrial policies. We found that Chinese policies are distorting markets in ways that undermine

innovation, subtract from U.S. market share, and put U.S. national security at risk. While stepping up

the pace of innovation, the United States should also act in the short term to try to reduce this market-

distorting behavior and its impacts. Our recommendations therefore focus on three approaches. First,

the United States should attempt to influence Chinese behavior by working to improve transparency

around Chinese policy through discussions in bilateral and multilateral forums, joining with allies to

coordinate and strengthen inward investment security and export controls, and responding firmly and

consistently to Chinese violations of international agreements. The United States also should calibrate

its application of national-security controls in response to Chinese industrial policy aimed at

undermining U.S. security.

Second, we find that a competitive domestic industry is critical to innovation and security. We therefore

recommend policies aimed at developing and attracting talent, funding basic research and development

that is critical to innovation, reforming corporate tax laws, and reforming permitting practices.

As noted above, however, a level playing field and a strong business environment are necessary but not

sufficient. Our final set of recommendations focuses on driving transformative innovation. We propose

a series of “moonshots”—such as developing game-changing biodefense systems and cutting-edge

medical technologies—that have independent merit and would, if achieved, also deliver radical

semiconductor advances of much broader applicability.

Delivering on such transformative innovation, as on our other recommendations, will require strong

cooperation among government, industry, and academia to be maximally effective.

Ensuring Long-Term U.S. Leadership in Semiconductors

1. Challenges and Opportunities

Semiconductors are essential to many products used in modern life, from computers, cellular

telephones, and solar panels, to medical diagnostics and self-driving cars. Progress in semiconductors

has opened up new frontiers for devices and services that use them, creating new businesses and

industries and bringing massive benefits to American workers and consumers and to the global

economy. Cutting-edge semiconductor technology is also critical to defense systems and U.S. military

strength, while the pervasiveness of semiconductors makes their integrity an important factor in

shaping cybersecurity risk.

Today, the semiconductor industry faces challenges from technological barriers and rapidly shifting

markets—trends that are now compounded by increasingly active Chinese industrial policy (see Box 1.

The State of U.S. Industry). U.S. policy should aim to sustain and grow the contributions semiconductor

technology makes to the economy and national security by promoting an environment that drives

semiconductor innovation while protecting against specific national-security risks. Delivering on these

goals requires sustaining a vibrant and competitive domestic semiconductor industry, but it will demand

much more than just that.

Industry is a critical part of the semiconductor innovation system. In order to maintain and accelerate

innovation in the semiconductor space, it is essential that companies have an open, market-based

environment in which they have intellectual-property protection, access to affordable capital, access to

leading-edge academic research and pools of well-trained engineers and scientists, and access to large

markets. Industry requires an environment in which market power is not misused. This environment

can be delivered through strong competition or by government-imposed restraints on abuse.

Innovation will benefit the U.S. economy regardless of where it occurs. Sustaining innovation, however,

is far more likely if the United States itself has a robust semiconductor industry (along with a strong

research community). The United States, in contrast with some major competitors, provides companies

the ingredients necessary to innovate rather than simply cutting costs for existing technologies.

Historically, U.S. government-sponsored research and development (R&D) has been essential to driving

semiconductor innovation—but that support will be unsustainable if industry is hollowed out.

The innovation spurred by a robust U.S. semiconductor industry also creates a virtuous cycle: by helping

U.S. producers stay ahead of competitors, it further strengthens U.S.-based industry, which in turn

drives semiconductor innovation. This cycle will be most powerful when U.S.-based industry stays

ahead through genuine comparative advantage, building on strong U.S. capital markets, talent, and

research institutions. If the United States instead tries to keep its industry ahead by shielding it from

legitimate foreign competition, innovation will suffer, ultimately leaving the U.S. industry less

competitive and the U.S. economy worse off.

Economic strength is also fundamental to U.S. national security—increasing the urgency of getting the

economic part of the semiconductor equation right. The United States faces a distinct set of specific,

Ensuring Long-Term U.S. Leadership in Semiconductors

semiconductor-related national-security challenges. To maintain its advantage, the U.S. military needs

access to leading-edge semiconductors that not all potential adversaries have. U.S. government

purchasers of semiconductors, including the U.S. military, also need to be able to mitigate risks to their

supply chains, with regard both to integrity and availability; moreover, mobile computing, automated

vehicles, and the Internet of Things increasingly place similar demands on commercially used

semiconductors, as a much broader civilian cybersecurity imperative.

A strong U.S.-based industry can mitigate some of these security concerns but is not a panacea for them.

Risks to the integrity of the semiconductor supply chain, while lower when critical items are designed

and produced domestically or on the territories of U.S. allies, cannot be assured through domestic

manufacturing and design alone and therefore ultimately need to be mitigated through other means

(such as integrity standards and testing and greater system resilience), regardless of where production is

located. Moreover, if the United States attempted to ensure security by simply restricting the set of

producers that was allowed to sell semiconductors to U.S. firms, it would slow innovation by

fragmenting markets and reducing competition. The U.S. government and U.S. consumers would be

increasingly unable to procure the cutting-edge chips on which the U.S. economy and national security

depend.

Box 1. The State of U.S. Industry

U.S.-headquartered firms have the largest share of the global semiconductor market, as measured by

revenue, but the semiconductor industry has steadily been globalizing over the last 40 years.

For approximately the past two decades, U.S.-headquartered firms have accounted for half of global

semiconductor sales. Other leading firms are based in South Korea, Japan, Taiwan and Europe. No

Chinese-headquartered company is in the top twenty.

How U.S.-based companies do business, however, has been changing. U.S.-headquartered

semiconductor companies have increasingly moved fabrication facilities abroad or focused on design

while contracting out fabrication (the so-called fabless business model). The share of worldwide

fabrication capacity located in the United States fell to about 13 percent in 2015, compared to 30

percent in 1990 and 42 percent in 1980, though nearly half of planned global additions are U.S.-

company-owned.

During the last four decades, the memory business has also largely shifted from

the United States to Asia, with notable exceptions for cutting-edge technologies.

U.S. companies still earn the largest share of revenues in a host of critical areas. The United States

has a majority of the global market for integrated circuits design and fabrication, which makes up

over 80 percent of the global semiconductor market.

Within integrated circuits, the United States

leads in logic and analog.

In particular, the United States has the clear lead in sales of high-end

See: fas.org/sgp/crs/misc/R44544.pdf.

Integrated circuits (ICs) are semiconductor devices (chips) composed of multiple electronic circuits. They are

used for most electronic applications in which semiconductors are needed. They include memory, logic, and

analog chips. The processor chip in a computer, for example, is an IC.

Logic chips are central processing units (CPUs)―sometimes called microprocessors―that control and carry out

computation; analog chips convert continuous signals such as sound or video that are found in the real world into

Ensuring Long-Term U.S. Leadership in Semiconductors

microprocessors, communication chips for smartphones and devices, and networking components for

routers, the Internet, and landline exchanges. The top integrated device manufacturer (combining

design and fabrication), the top three fabless companies, the top three Electronic Design Automation

(EDA) companies, and two of the top three equipment manufacturers (all by revenue) are U.S.-

headquartered.

Evolving Challenges

The global semiconductor market has never been fully free: it is founded on science that historically has

been driven, in substantial part, by government and academia; segments of it are restricted in various

ways as a result of national security and defense imperatives; and it is frequently the target of national

industrial policies. Market forces play a central and critical role. But any presumption by U.S.

policymakers that existing market forces alone will yield optimal outcomes – particularly when faced

with substantial industrial policies from other countries – is unwarranted. In order to realize the

opportunities that semiconductors present and to mitigate major risks effectively, U.S. policy needs to

confront challenges from changing technology at the same time as it faces a new and aggressive set of

Chinese industrial policies designed to shift the competitive dynamics in the global industry in favor of

Chinese production and companies. We now turn to those challenges and threats.

Technology and Markets

The model for semiconductor innovation has long been simple, at least in principle. The semiconductor

industry doubled the number of transistors on a chip, and hence performance, roughly every 18–24

months—the so-called “Moore’s Law”—while maintaining or reducing cost. This trend has been

supported by industry and customers aligned on a simple innovation goal―essentially faster

computing―across the value chain and by a consistent, dominant focus on improving speed and density

in CMOS technology.

Maintaining this pace now faces two fundamental challenges. Because of physical limits, it is now

harder―and will eventually be impossible―to shrink silicon-based transistors further. Progress has

already begun to slow: the doubling time for the number of transistors on a chip has increased to

roughly 30 months and continues to rise. In addition, as the range of major semiconductor applications

has grown from traditional computing to include areas such as mobile devices, automotive advances,

and data centers for the cloud, the appropriate emphasis of innovation is shifting and expanding,

moving beyond mere processing speed to include progress on energy consumption, “system on a chip”

functionality, and other dimensions. In the past, given that companies were largely working towards the

same goal—doubling the processing power of chips while reducing or maintaining costs—private and

public resources were concentrated on a shared goal. Now, companies are working on a more diverse

range of technologies with different goals, making it harder to align strategies and investments across

industry.

digital data used by logic chips. They are used, for example, to convert the sound from a mobile phone into digital

signals.

CMOS stands for Complementary Metal-Oxide Semiconductor and has long been the dominant technology for

constructing integrated circuits used in microprocessors, microcontrollers, memory chips, and other digital logic

circuits.

Ensuring Long-Term U.S. Leadership in Semiconductors

The semiconductor industry is facing these challenges at a time of rising industry concentration. The top

five global semiconductor suppliers now make up nearly 40 percent of the market, up from 32 percent

in 2006.

Costs are driving this concentration. For example, to meet the technical challenges of speed

and miniaturization, new advanced-logic, semiconductor fabrication facilities are expected to cost well

over $12 billion in the United States.

Similar leading-edge fabrication facilities cost under $5 billion just

5 years ago. This escalation is driven in part by higher equipment costs, weaker economies of scale, and

R&D costs associated with fundamental physics challenges. Venture-capital investment in the

semiconductor industry has declined sharply, making it difficult for new companies to enter and

compete.

High industry concentration has some benefits: firms with little competition can, in some cases, afford

to invest more in long-term innovation. But concentration also comes with major risks. High market

concentration means less competition—which reduces incentives for companies to lower costs, raise

quality, and invest in new products and technology, while making collusion more likely. Higher

concentration also raises the stakes for firm location: in the extreme case, if most or all production in an

industry becomes controlled by a single non-U.S.-based company, that situation raises the odds that

production or sales will be influenced by foreign-industrial or national-security policy to the detriment of

U.S. interests. In contrast, in a highly competitive industry with a diverse production and ownership

base, efforts by one country to distort markets—whether for economic or geopolitical purposes—are far

less likely to have wide-ranging consequences.

Chinese Industrial Policies

Slowing innovation, changing markets, and rising concentration would be significant challenges by

themselves. But Chinese industrial policies aimed at achieving a global leadership position in

semiconductor design and manufacturing through non-market means, together with the steady growth

in Chinese domestic semiconductor consumption, are now compounding those challenges. Chinese

competition could, in principle, benefit semiconductor producers and consumers alike. But Chinese

industrial policies in this sector, as they are unfolding in practice, pose real threats to semiconductor

innovation and U.S. national security.

China’s starting position in its quest for semiconductor prowess is well behind that of the United States.

Chinese manufacturing of advanced-logic chips is significantly behind the state of the art in the United

States, Taiwan, and elsewhere. China has many semiconductor foundry companies, but all are at least

one-and-a-half generations behind the state of the art in volume production. In addition, there are

currently no domestically-owned memory companies producing at commercial volume in China; all

advanced-memory manufacturers in China are foreign-owned. Since the foreign companies have

chosen to have no Chinese companies involved as joint venture partners, China is spending significant

amounts of capital to develop its own indigenous memory industry. Similarly, China lacks a tier-one

semiconductor equipment firm.

There is one tier-two semiconductor equipment company in China—

See: technology.ihs.com/553230/preliminary-2015-semiconductor-market-shares and

i.cmpnet.com/eetimes/eedesign/2007/chart1_031507.gif.

See: http://semiengineering.com/10nm-fab-watch.

Tier one companies are direct suppliers to equipment manufacturers (OEMs). Tier two companies are the key

suppliers to tier one suppliers, without supplying a product directly to OEM companies.

Ensuring Long-Term U.S. Leadership in Semiconductors

AMEC—that makes manufacturing tools involved in the fabrication of semiconductors. The most likely

avenue for Chinese growth will be acquisition of global players (or divisions of them) in the United

States, Europe, or Japan; Chinese firms have been increasingly active in the acquisition space. China’s

brightest spot is its fabless semiconductor industry, which is booming. There are close to 400

companies, many of which are growing. But there is a technological gap between China’s fabless

semiconductor industry and those of other countries; at this time, most of China’s fabless companies are

focused on the low-end and mid-range parts of the market. Assuming that these companies continue to

grow quickly—as they are making tailored products for the China market—that will provide motivation

for foreign-owned fabrication companies to have a presence in China, which in turn may draw foreign-

owned equipment makers.

The Chinese government, motivated by economic and national-security goals, has publicly asserted its

desire to build a semiconductor industry that is far more advanced than today and less reliant on the

rest of the world. After more than a decade of failed attempts to promote its semiconductor industry,

in 2014 China promulgated “IC Promotion Guidelines” putting forth a new plan, including revenue

targets and technology goals. This plan has been backed by the senior Chinese leaders (including

President Xi Jinping according to public reports). One stated aim of Chinese policy is for China to be at an

“advanced world-level [semiconductor capability] in all-major segments of the industry by 2030.”

China’s strategy relies in particular on large-scale spending, including $150 billion in public and state-

influenced private funds over a 10-year period, aimed at subsidizing investment and acquisitions as well

as purchasing technology. This figure is slightly smaller than the average of $23 billion spent annually on

semiconductor mergers and acquisitions (M&A) by all U.S. companies over the past 5 years. Already

multiple Chinese-government investments executed by investment firms are enabling this government-

directed strategy. Consistent with its industrial-policy tactics in other industries, China also places

conditions on access to its market to drive localization and technology transfer, according to public

reports.

Chinese policy exploits headwinds currently facing semiconductor innovation: if the leading

edge is advancing slowly, that makes it easier for China to use industrial policies to get technologically

close enough to supplant the innovation leaders economically.

Chinese industrial policy can usefully be divided into two categories: subsidies and zero-sum tactics.

Subsidies

The Chinese government provides a range of subsidies to strengthen domestic production. These

subsidies are driven in part by a desire to decrease reliance on foreign suppliers for technologies

deemed critical to Chinese national security, and in part by a desire to capture market share for

economic reasons. China’s subsidies to the semiconductor industry include capital subsidies that

encourage foreign companies to locate facilities in China as well as subsidized capital to domestic

companies and investment firms to use in the acquisition of foreign companies and technologies. While

China’s subsidies are largely zero-sum in their impact on foreign semiconductor producers (companies,

workers, or both) in the same market segment, they may not be zero-sum to other market participants

See: www.minneapolisfed.org/research/sr/sr486.pdf.

Ensuring Long-Term U.S. Leadership in Semiconductors

(whether to participants in other parts of the semiconductor supply chain or to users of

semiconductors).

In the short run, Chinese subsidies can benefit U.S. consumers and firms that use semiconductors by

reducing costs and lowering product prices. In the long run, however, subsidies to incumbent

technologies tend to reduce innovation.

Depending on the initial state of the market, subsidies can

also increase market concentration in China. This can increase national-security risks for the United

States and other countries and, to the extent that Chinese policy allows firms to sell below cost of

production, raises risks of overcapacity, which threatens direct competitors. Subsidies also, more

directly, can erode U.S. market share, damaging industry employment as well as innovation.

Zero-Sum Tactics

China also employs a variety of tactics that are more broadly and unequivocally zero-sum. These are

policies that shift business to China while raising, not lowering, costs. These policies are harmful

because they hurt otherwise sound businesses without bringing countervailing economy-wide benefits,

raise prices for consumers and other businesses that use semiconductors, and can deter innovation.

Such policies also can create defense-related national-security risks by accelerating the spread of

sensitive technologies. Chinese zero-sum activities include:

• Forcing or encouraging domestic customers to buy only from Chinese semiconductor suppliers.

China is doing this both explicitly (e.g., in government contracts) and indirectly (e.g., through its

proposals to implement “secure and controllable” requirements relating to cybersecurity

concerns). Such practices reduce incentives for innovation across the board: non-Chinese

companies see smaller markets, while Chinese companies face less competition. Given the size

of the domestic Chinese market, these practices could also result in a high concentration of the

global market in China over the longer run.

• Forcing transfer of technology in exchange for access to the Chinese market. This practice

affects innovation by reducing incentives for R&D, including in the United States (since U.S

companies sell into and compete with China), and by quickly turning leading-edge technologies

into commodities that anyone can produce. It can also increase market concentration in China;

conversely, as Chinese market concentration increases, so does China’s ability to force

technology transfer, creating a vicious cycle.

• Theft of intellectual property. This activity affects innovation in a similar manner to forced

technology transfers. According to media reports, China steals intellectual property both

covertly and overtly. Overt means using inspections for “secure and controllable” technology to

gain access to detailed knowledge of semiconductor technologies.

Zero-sum tactics are those where one entity’s gain is equivalent to another’s loss, where instead of growing

overall value, entities are taking from each other.

See: economics.mit.edu/files/8790.

See: www.nytimes.com/2016/05/17/technology/china-quietly-targets-us-tech-companies-in-security-

reviews.html.

Ensuring Long-Term U.S. Leadership in Semiconductors

• Collusion. According to media reports, Chinese companies have colluded to lower the value of

takeover targets before purchasing them in distressed situations.

These behaviors sometimes violate international agreements but in other cases do not. Actions taken

by other countries in response to Chinese industrial policy are described in Box 2.

Responding Strategically

In responding to these challenges, U.S. policymakers should be guided by six principles.

1. Win the race by running faster. There is often a strong temptation to respond to the challenges

outlined above by focusing centrally on attempting to slow China down. But, even if that were

possible, it would not be desirable. If the United States stays ahead in manufacturing but does not

innovate, progress in semiconductors will cease to drive increased living standards or military

strength; moreover, China will inevitably advance in semiconductors, even if the United States slows

its progress. The most effective path to continued semiconductor leadership would focus centrally

on sustaining U.S. innovation.

2. Focus principally on leading-edge semiconductor technology. Leading-edge semiconductor

technology is central to economically transformative innovation. It is also critical to sustaining a

national-security edge. Away from the leading edge, policymakers should focus sharply on major

security risks, including continuity of supply and large violations of international trade and

investment rules, rather than on broad leadership across the board in semiconductors.

3. Focus on making the most of U.S. strengths rather than trying to mirror China. The United States

and China have very different views about the appropriate relationship between government and

the private sector. In the United States, the role of the Federal government in the economy is

largely to create the right environment for the private sector to succeed, including by funding pre-

competitive innovation and acting as a trusted convener; the U.S. government role is not to allocate

capital to particular firms or sectors. China has been far more willing to provide subsidies to mature

firms and industries and to sustain that support even if it results in overcapacity and economic

losses. The United States has also advocated for open global trade and investment, with some

exceptions for national security. This stance has benefited consumers and the global economy.

China has gained from global openness but has been less committed to sustaining it—and, in some

cases, has worked against it, for example, by limiting or conditioning access to China’s market. Now,

globally, more countries are questioning the benefits of economic openness—a trend that will

shape, and be shaped by, how the United States responds to challenges in the semiconductor arena.

4. Anticipate Chinese responses to U.S. actions. China will not stand still—and, in particular, will likely

adjust what it does in response to U.S. policy. This is a particularly acute challenge for those areas in

which the U.S. government had its greatest leverage over industry, e.g., U.S. investment review of

Chinese companies seeking to buy U.S. companies and U.S. export controls, which are also those

See: www.nytimes.com/2016/09/17/business/dealbook/china-germany-takeover-merger-technology.html.

Ensuring Long-Term U.S. Leadership in Semiconductors

areas in which U.S. policy changes would be most likely to influence Chinese policy (and Chinese

corporate behavior) in turn.

5. Do not reflexively oppose Chinese advances. The U.S interest in promoting an open, competitive

global economy, with its attendant economic and security benefits, will often outweigh the benefits

of attempting to stop a specific undesirable market-based development, so long as that

development is not driven by Chinese policies that breach international rules or accepted standards

of fair behavior. At the same time, the U.S. government will need to identify areas in which the

diffusion of particular semiconductor technologies, or control of particular companies, poses

intolerable national-security risks that cannot be mitigated through steps short of stopping their

acquisition and, therefore, should be stopped to the extent possible.

6. Enforce trade and investment rules. The U.S. government should oppose Chinese actions that

violate rules of open trade and investment, even if those actions might appear to narrowly benefit

the U.S. economy. This opposition should be active, drawing on the full range of tools available to

the United States under international agreements, rather than merely rhetorical. It should also be

done, in so far as possible, in coordination with other countries to increase effectiveness and reduce

the risk of retaliation (see “Responses by Other Countries”).

The semiconductor space is complex. As U.S. policymakers craft solutions, they should draw on industry

expertise, in order to increase the odds that their policies have the desired impact.

Recommendation 1.1: Create new mechanisms to bring industry expertise to bear on semiconductor

policy challenges. As the U.S. government further develops and executes a strategy to sustain

semiconductor leadership and drive innovation, it will be important to be able to draw on cutting edge

technological and markets expertise from within semiconductor producing and consuming industries.

PCAST recommends that the Administration explore several (potentially complementary) options.

The U.S. government could benefit from a standing committee of industry experts who are

engaged on a continuing basis. This could be structured as a Department of Commerce

Technical Advisory Committee (TAC), an arrangement that already exists for eight industry

areas. It could also be structured as a PCAST subcommittee. It would be critical for committee

membership to be broad, including industries that use semiconductors, in order to provide

broad insight into how changes in the semiconductor space might affect the broader economy

and national security. Regardless of where an advisory committee is housed, its expertise would

need to be made available to the broader U.S. government.

The U.S. government could also take steps to make it easier to access industry expertise on an

ad hoc basis. This could include creating alternative processes to provide temporary security

clearances to industry experts, as has been done for some aspects of cybersecurity. It could also

include creating legal arrangements that make it possible for the U.S. government to consult

responsibly with industry experts on commercially sensitive matters.

Promoting U.S. interests will ultimately require a strong focus on advancing semiconductor innovation.

This demands a three-part strategy that pushes back against innovation-inhibiting Chinese industrial

Ensuring Long-Term U.S. Leadership in Semiconductors

policy, improves the business environment for U.S.-based semiconductor producers, and helps catalyze

transformative semiconductor innovation over the next decade.

Box 2. Responses by Other Countries

Japan, Korea, Taiwan, and Europe all have prominent roles in the global supply chain for

semiconductors. Industries and governments across the world have reacted to Chinese industrial

policy (in the semiconductor domain and others). Concern around the semiconductor industry is

most pronounced in South Korea and Taiwan, but is also rising in Europe. In response, governments

have used their trade rules, tightened their investment rules, and increased funding for

semiconductor research and design. Specific responses include:

• Taiwan has taken steps to prohibit or heavily restrict Chinese acquisition of Taiwanese

semiconductor technology by not approving any mainland Chinese acquisitions of, or

investments into, any domestic semiconductor firms. The Taiwanese government has also

launched a public-private partnership with industry to co-invest in R&D.

• South Korea has tightened its rules that restrict flow of critical semiconductor intellectual

property to China. In addition, the South Korean Government and industry have partnered to

launch a $175 million semiconductor start-up fund. The government has also created

incentive programs to encourage its semiconductor engineering talent not to work for

Chinese semiconductor firms.

Ensuring Long-Term U.S. Leadership in Semiconductors

2. Influencing Chinese Actions

The United States has a range of tools available to respond directly to Chinese activities. These include

formal trade agreements, informal trade and investment norms agreed to with foreign countries, and

unilateral tools such as scrutiny of acquisitions of potential national security concern by the interagency

Committee on Foreign Investment in the United States (CFIUS). The current set of strategies pursued by

the Chinese government through its policies brings the effectiveness of these tools, as currently applied,

into question, however. The U.S. government should revisit its tools to ensure that they are sufficiently

able to protect against actions that may unacceptably harm the country’s economic and security

interests. This section focuses on three foreign policy efforts that U.S. policymakers should pursue.

Recommendation 2.1: Boost the transparency of global advanced technology policy. Ideally, all

countries should pursue fair, market-oriented policies toward the semiconductor industry, with

reasonable exceptions for national security. Finding agreement with China around that goal, though, is

likely to be difficult if not impossible. The United States therefore should aim to boost transparency

around Chinese policy in advanced technology, including semiconductors.

China is obligated under the World Trade Organization (WTO) Subsidies Agreement to notify

other countries of all of its subsidy programs. While China has notified other countries of some

semiconductor programs, the U.S. government believes that its subsidy notification is not

complete. China also has obligations under the WTO to publish all trade-related legal measures

in a single place and make translations available; it does not appear as if China has fully adhered

to this obligation; moreover, in November 2016, in the context of the U.S.-China Joint

Commission on Commerce and Trade (JCCT), the two countries asserted a joint commitment to

a semiconductor industry that “operates in fair, open, and transparent legal and regulatory

environments.” China stated that its government has never asked for its investment funds to

“require compulsory technology or IPR transfer,” and the two countries agreed to continue to

exchange information related to the sector.

Building on these foundations, the U.S. government, in consultation with industry, should

identify specific areas where greater transparency could be mutually beneficial, and press for

progress there. In doing so, it should be open to increasing transparency around its own

activities. The United States should also seek to broaden support for increased transparency,

whether by reinforcing it in other U.S.-China dialogues such as the Strategic and Economic

Dialogue, multilateral forums such as the G-20 and Asian-Pacific Economic Cooperation (APEC),

or new trade agreements that seek to set new, higher standards.

Recommendation 2.2: Reshape the application of national security tools, as appropriate, to deter and

respond forcefully to Chinese industrial policies. The U.S. government has focused its semiconductor-

related inward investment restrictions, export controls, and government procurement that requires

domestic content on national security, rather than narrow economic objectives. It should continue to

maintain that focus rather than expand these tools to pursue explicitly economic goals, a step that could

Ensuring Long-Term U.S. Leadership in Semiconductors

easily backfire. That does not mean, though, that the United States should not respond forcefully to

Chinese policies that distort the global market by limiting access of U.S. companies and U.S. exports to

China’s large and growing semiconductor market.

The U.S. government should start by clarifying the types of measures that it believes are

acceptable for protecting national security. (For example, restricting certain defense

procurement to firms based in particular countries may be appropriate.) To the extent possible,

it should do that in dialogue with China, with both governments agreeing in principle on what is

and is not reasonable, much as the two countries have done in the area of cyberattacks. So long

as China adheres to these norms, the United States should continue to implement its policies

much as it has in the past. If, however, China acts (or continues to act) in ways well outside

these bounds—for example through its “secure and controllable” rules for information

technology—those actions should affect U.S. policy. One way to respond would be to tie U.S.

assessments of the national-security threats posed by particular technology exports,

investments, and contracts to Chinese policy. (For example, if China pursues a policy of

undermining cutting-edge, defense-critical U.S.-based companies by flooding markets using

government support, that should alter U.S. assessments of whether Chinese acquisitions of the

capabilities required to do so are acceptable.) The main goal here should be to deter dangerous

Chinese actions; this means that the United States will need to be more open and clear about

how its investment and export restrictions actually work. If, however, this effort to deter fails,

changes in U.S. national security threat assessments will presumably lead to changes in the

specific exports and acquisitions allowed by the U.S. government.

Recommendation 2.3: Work with allies to strengthen global export controls and inward investment

security. The United States dedicates substantial resources to export controls and inward investment

security. The United States should work with like-minded partners to develop common principles

(insofar as possible) for acceptable and unacceptable market behavior, and to help build their

administrative capacity to effectively implement appropriate controls and pursue needed investigations,

since many countries are currently far less capable than the United States in this regard. Led by the

Departments of Treasury and Commerce, and by the Intelligence Community, this effort should provide

global partners with technical assistance, training, and diplomatic support to identify risks and

vulnerabilities and to remediate them.

On export controls, the Departments of Commerce, Defense, Homeland Security, State, and

Treasury work together to target export controls on national security priorities while facilitating

legitimate commerce. To oversee investment security, CFIUS, chaired by the Treasury

Department, plays a vital role in reviewing certain transactions with important national security

implications involving potential foreign control of U.S. businesses.

The United States should not, however, carry this burden alone. Unilateral action is increasingly

ineffective in a world where the semiconductor industry is globalized (though keeping the

United States at the forefront of innovation allows it to retain some ability to be effective

unilaterally). At the same time, when it acts unilaterally, the United States often raises

suspicions (however ill-founded) among allies that it is motivated by economic-competitiveness

concerns, rather than by national security.

Ensuring Long-Term U.S. Leadership in Semiconductors

3. Creating a More Supportive Business Climate in

the United States

A U.S. semiconductor strategy needs to foster a supportive business environment in the United States.

We focus here on sustaining a world-class workforce, boosting general-purpose scientific research,

enacting prudent business tax reform, and responsibly speeding facility permitting. These are all areas

in which U.S. companies, and many policymakers, have long called for action, and the ideas we outline

below are largely not new. These policy areas are, however, particularly important to sustaining

semiconductor innovation.

Recommendation 3.1: Secure the talent pipeline. Global businesses invest where the talent exists.

With technological advances transforming existing facilities and with the growth of the so-called fabless

model, the talent needs that are required to continue to drive semiconductor innovation and production

are changing. The United States needs to strengthen its home-grown talent and attract talent from

around the world.

Thanks to the efforts of the government, the private sector, and academia, the share of science,

technology, engineering, and mathematics (STEM) bachelor’s degrees of total bachelor’s

degrees awarded has grown 12.3 percent since the 2008-2009 academic year. The U.S.

government needs to double down on these collaborative efforts, continuing and expanding its

investments in STEM education in partnership with the private sector and academia. As PCAST

has noted in its 2012 report Engage to Excel, the first 2 years of college are the most critical to

the recruitment and retention of STEM majors.

Ways to lift the Nation’s game in this critical

phase of STEM education are detailed in that PCAST report.

The United States also needs to attract talented scientists and engineers from abroad to live and

work in the United States: immigrants play a critical role in U.S. high-tech businesses, 25

percent of which were founded by immigrants and provide jobs to immigrants and U.S.-born

residents alike. The U.S. government should pursue steps that attract talented immigrants while

expanding opportunity for all Americans. Some examples include giving STEM graduates from

accredited U.S. universities fast-tracked, long-term visas; increasing the number of H-1B visas;

and/or allowing existing visas to cover an employee’s spouse and children.

Recommendation 3.2: Invest in pre-competitive research. Pre-competitive research is essential to

continuing to drive innovation in the semiconductor industry and broader economy. For example,

recent advances in wide bandgap semiconductors—semiconductors that can withstand certain extreme

environments—were originally driven by government-funded basic research, but have recently found

widespread, important use in applications including electric vehicle charging and solar power. The U.S.

See: www.whitehouse.gov/sites/default/files/microsites/ostp/pcast-engage-to-excel-final_2-25-12.pdf.

See: www.whitehouse.gov/sites/default/files/microsites/ostp/pcast_future_research_enterprise_20121130.pdf.

Ensuring Long-Term U.S. Leadership in Semiconductors

government should increase its R&D spending, focusing in particular on collaborative pre-competitive

R&D in science and technology areas outside the health sciences (which have received large increases in

recent years). We recommend targeting a doubling of Federal non-health R&D spending.

The semiconductor industry spends roughly 20 percent of sales on R&D; however, this

investment is focused on competitive research and product development. A critical role of the

U.S. government is to support pre-competitive research, where accelerated innovation would

benefit the government, industry, and academia, but no one firm has sufficient incentive to

invest alone.

Pre-competitive basic research will often not be specific to semiconductors, but rather build a

broader foundation for innovation. For example, support for basic science related to carbon

nanotubes was never aimed at any one particular application but has now led to advances that

applied researchers are translating into cutting edge commercial applications to semiconductors

and other technologies. Unfortunately, the U.S. government now spends less on R&D than in

the past: as of 2012, it spent approximately 0.7 percent of GDP on R&D, down from 0.8 percent

in 2008 and 1.8 percent in the 1960s. In particular, this pre-competitive basic research would

fall into non-defense discretionary spending that is at a historic low as a share of Federal budget

and continues on a downward trend.

Recommendation 3.3: Enact corporate tax reform. The U.S. tax system penalizes asset-heavy

industries by discouraging capital investment. The semiconductor industry—particularly fabrication—is

highly capital intensive. To encourage market-driven investment in the United States in the

semiconductor industry, the corporate tax system should be reformed to create a more attractive

environment for businesses to compete globally, while ensuring that the U.S. tax system as a whole is

fair and responsible to all Americans.

The United States has the highest statutory corporate tax rate, including Federal and State

taxes, among the 34 members of the Organization for Economic Co-operation and Development

(OECD), yet loopholes, tax expenditures, and tax planning strategies narrow this tax base. As a

result, many U.S.-headquartered businesses are disadvantaged relative to foreign and domestic

competitors. Comprehensive U.S. tax reform is particularly important for an industry like

semiconductor manufacturing, given its capital intensity. The U.S. government should

implement the actions that PCAST has previously recommended: (i) recognizing the importance

of manufacturing through the tax code; (ii) strengthening R&D tax credits; and (iii) creating an

internationally competitive corporate tax system. These recommendations are presented in and

elaborated further in the 2012 report on the Advanced Manufacturing Partnership (AMP).

Recommendation 3.4: Responsibly speed facility permitting. Continuing to lead in semiconductors

and other research-intensive, high-capital industries requires being constantly able to move beyond the

current generation of technology. That, in turn, often requires building next-generation facilities in

short time frames. Improving permitting could yield a step change in the pace of U.S. semiconductor

foundry innovation.

The permitting process provides important public benefits, examining projects to determine

whether they meet public objectives, including minimizing environmental and community

Ensuring Long-Term U.S. Leadership in Semiconductors

impact, which companies are not always economically incentivized to do. The combination of

the current Federal and state permitting and review processes, however, can be slow,

unpredictable, and lacking in transparency.

Within the semiconductor industry, some executives identify current implementation of the

Federal Clean Air Act (CAA) as the primary barrier to responsible and timely facility permitting.

Under the CAA, two main permitting programs are likely to be relevant to the semiconductor

industry: a preconstruction permit and an operating permit.

Each of these programs has

specific applicability thresholds; however, generally, these permits are issued by State and local

air agencies. In most circumstances, for both types of permits, the U.S. Environmental

Protection Agency (EPA) has the opportunity to review a draft permit and provide any

comments to the State or local permitting authority. For some large projects, the permitting

process can take 12–18 months. This can create significant barrier to planning and building new

manufacturing capacity in an industry where speed to market can be very important, given the

pace of competition and innovation.

To responsibly accelerate facility permitting, we recommend simplifying the existing permitting

process for high-technology facilities and carefully providing opportunities for speedier review:

1. The Federal government should review Federal permitting for high-technology facilities to

identify important areas where regulations or procedures are redundant with state rules

and might therefore be modified or removed. A first step would be for Congress to request

that the Government Accountability Office provide an assessment of areas with significant

overlap and identify any that have large impacts on semiconductor foundry investment.

2. EPA should create additional “fast track” permitting options and review existing ones to

ensure that they are operating as intended. While the Federal government may need to

conduct a more thorough review for brand new facilities, “fast track” options could focus on

companies building new fabrication facilities at existing sites. States are leading the way

here—for example, the State of Oregon has developed a Plant Site Emissions Limit (PSEL)

program, which assigns to all facilities in the state an emissions cap that they must operate

under for all CAA-covered pollutants. As long as a facility remains under its cap, it can make

many types of operational changes, including expansions, without significant oversight by

the State or EPA. Indeed the EPA has already adopted provisions similar to the Oregon PSEL

program in the Federal New Source Review (NSR) rules as part of a 2002 reform. EPA should

consult with industry to verify that these reforms are functioning as intended, make

modifications as appropriate, and identify additional fast track approaches that might be

appropriate.

3. Congress should provide funds to increase staff capacity at EPA and other relevant agencies

to handle the permitting process. While some applications need significant review, a major

bottleneck in the permitting review process can be insufficient capacity.

See: www.epa.gov/sites/production/files/2015-12/documents/20090925fs.pdf, www.gpo.gov/fdsys/pkg/FR-

2009-10-06/pdf/E9-23794.pdf, and www.epa.gov/sites/production/files/2015-09/documents/eval-

implementation-experiences-innovative-air-permits.pdf.

Ensuring Long-Term U.S. Leadership in Semiconductors

4. Developing a “Leapfrog” Strategy for Continuing

U.S. Leadership

The United States will not remain a semiconductor leader if it confines its efforts to making it cheaper

and easier to build today’s semiconductors and opposing damaging Chinese industrial policy.

Ultimately, to maintain a strong and globally competitive semiconductor industry, the United States

needs an economic and policy environment that fosters innovation and keeps the U.S. industry at the

technological frontier.

The United States last faced a major challenge to semiconductor competitiveness and innovation in the

1980s. The U.S. government responded through technology policies that focused on continuous

improvement in computing speed based on existing technology fundamentals (which required

significant technological innovation). A successful strategy today must be different for the two reasons

described in Chapter 1—looming limits to CMOS technology and fundamental shifts in the

semiconductor market—that now emphasize other performance dimensions beyond processing speed.

Additionally, the diversity of computing systems has blossomed dramatically in the last 30 years,

revealing a greater variety of solutions beyond today’s CMOS technologies that are capable of whole

new paradigms of computing.

These circumstances mean that the role of government will need to be supporting rather than central.

Government procurement is only a small part of the semiconductor market—not enough for

government to cause a wholesale change to how semiconductor technology is pursued. While total U.S.

government spending on all non-defense R&D was $65.9 billion in 2015, the semiconductor industry

alone nearly matched this level of R&D spending at $55.4 billion. But U.S. policymakers can help a

diffuse set of players in academia, industry, and government laboratories organize around important

common goals and support catalytic activities that remove obstacles to fundamental technological and

industry progress. This approach lies somewhere between “top-down” and “bottom-up:” government

should set ambitious and clear goals, rather than assuming that all progress is equally useful and support

only key activities, rather than trying to comprehensively dictate all activities. In short, semiconductor

innovation should not be viewed as an independent goal—rather, it must be part of broader innovation

in the ways semiconductors are used.

Ensuring Long-Term U.S. Leadership in Semiconductors

Focus Areas

We recommend that U.S. policymakers carefully select ambitious challenges, which we call

“moonshots,” as focal points for industry, government, and academic efforts to drive computing and

semiconductor innovation forward together. These moonshots are aspirational goals with society-wide

benefits—like developing affordable desktop semiconductor fabrication capabilities that could take the

place of a billion dollar fabrication facility and allow the production of small batches of structures; using

3D printing at the nanoscale to connect “hard” electronic materials with “soft” biological materials,

which could be the foundation of a zero-day bio-threat detection network; or a commercial, gate-based

quantum computer to work on large-scale problems.

The purpose of selecting a goal is that it will

catalyze activity that will accelerate innovation and create new technologies and systems that can then

be used more widely. The Apollo program itself—the original moonshot—did just this: it captured the

imagination by setting a big, important goal that ultimately drove fundamental technological advances

of much broader value.

Our recommended approach to designing the moonshots is driven by the fact that the future of

semiconductors and computing lies in innovating along multiple dimensions: new ways of performing

calculations (such as non-von Neumann and approximate computing), utilization of materials other than

silicon (such as carbon nanotubes and DNA for computation and storage), and novel approaches to

integrating semiconductors into the devices we use (such as embedding into fabrics and the Internet of

Things). (For further discussion of opportunities for innovation, see Appendix A, particularly Table A1.)

This contrasts with the traditional approach associated with Moore’s Law of focusing most innovation

into regularly doubling the number of transistors on a chip. A focus on innovating along multiple

dimensions, many of which are novel, also plays to capabilities where the U.S.-based innovators are

particularly strong. The specific moonshots are therefore crafted with an eye toward capturing the

imagination and driving innovation along multiple dimensions, which should attract a wide range of

players.

Government should loosely coordinate industry, government, and academic efforts around solving these

moonshots, with an aim to drive innovation with broader payoffs. Government will also almost certainly

need to back these efforts with significant, catalytic funding to overcome the risks associated with

radical innovation. Four principles should guide the design and selection of moonshots to strengthen

semiconductor competitiveness and innovation:

1. Applications-driven approach. Policymakers should take an application-driven approach to

innovation. This means that each moonshot that policymakers select should be chosen with the

goal of motivating progress in one or more semiconductor-enabled applications of significant

economic or strategic importance that currently lack adequate technological solutions. Put another

A zero-day bio-threat detection network is infrastructure to catch undisclosed software vulnerabilities that may

enable biological or chemical weapons. A zero-day bio-threat detection network is infrastructure to detect

previously unseen biological threats that may result from chemical or biological weapons. A quantum computer is

one that runs on qubits, and is the quantum analog of classical computer that runs on classical bits. Similarly,

quantum gates are analogous to classical logic gates, which implement Boolean functions, but in the case of

quantum computers these quantum gates form quantum circuits which perform the manipulations of the

quantum information state.

Ensuring Long-Term U.S. Leadership in Semiconductors

way, these moonshots should be selected so that the process of attempting to solve them can be

expected to yield major general-purpose advances in semiconductor technology and related

computing innovation. The R&D should integrate efforts from top-level applications through

component technologies. This approach can provide coherence to a sprawling space, is more likely

to attract sustained funding and buy-in from the agencies that will ultimately be responsible for

carrying the work out, and can motivate individual scientists and engineers to develop new

advances. It also reflects the fact that developing such applications can be broadly beneficial to

society in ways that are not reflected in profits they generate for those who develop them.

2. Ten-year time horizon. Policymakers should focus on projects where the right technological

approach could, in principle, yield a breakthrough solution in less than ten years. This means that

this effort will need to focus on strategies to move existing research to market.

3. Compensate for weak industry investment. Key application domains (clusters of applications) can be

broken down into those with stronger and weaker interest from industry. Certain areas—like big

data analytics, artificial intelligence, and autonomous systems—have strong industry interest, as

they are more easily monetized in conjunction with their existing business interests. These are likely

to receive adequate funding and industry-led support by companies to develop leapfrog

technologies for their own business reasons within ten years. For these domains, the government

should simply help accelerate where appropriate, on an ad hoc basis, primarily as a funder of

foundational research and as an early adopter of these new systems. Other domains—such as

advanced materials science, advanced manufacturing, and modeling and simulation—are critical to

the U.S. economy and would be widely utilized by industry, but the consuming industries lack the

capabilities to build these leading-edge systems as part of their own business strategies. In these

cases, in addition to funding research, the government should play a more active role coordinating

early purchases and facilitating industry collaborations on moonshots that accelerate progress in

these domains to ensure that promising technologies are ultimately commercialized. A table with

more detailed examples is included below.

4. Reduce design costs. The cost to design complex integrated circuits is increasing rapidly, stifling the

ability to design systems that focus on tailored applications. The Federal government should invest

in R&D that makes it as easy to design hardware as software. For example, the government could

support the development of tools akin to the modern computer-aided design tools that emerged in

the 1980s, with the goal of reducing design costs by one to two orders of magnitude. These tools

will also enable the same design to be used for a range of technologies.

Ensuring Long-Term U.S. Leadership in Semiconductors

Table 1. Examples of key application domains that will benefit from advances in semiconductors.

Strong Industry Interest

(Government Support)

Weak Industry Interest

(Government Leadership)

• Big Data Analytics

• Artificial Intelligence and Machine Learning

• Biotechnologies, Human Health Technologies

• Robotics, Autonomous Systems

• Telepresence, Virtual Reality, Mixed Reality

• Machine Vision

• Speech Recognition and Synthesis

• Nanoscale Systems and Manufacturing

• Ultra-High Performance Wireless

• Holistic Secure Systems

• Computational Chemistry

• Advanced Materials Science and

Manufacturing

• Modeling and Simulation

• Space Technologies

Recommendation 4.1: Execute moonshot challenges. The National Science and Technology Council

(NSTC) should form a Subcommittee on Semiconductor Moonshots under its Committee on Technology

to coordinate the selection, development, and execution of moonshot challenges. The membership

should include officials from the Executive Office of the President (National Economic Council, Office of

Big Data Analytics: Local real-time data analysis and visualization enabled by advances in security, low-power

computation, and processor specialization.

Artificial Intelligence and Machine Learning: Supervised and unsupervised machine learning enabled by new

processors, including low-power processers, graphics processing units, and quantum computers.

Biotechnologies, Human Health Technologies: Medical implants that are capable of ultra-low power processing,

communications, and wireless charging.

Robotics, Autonomous Systems: Speech and image recognition for mobile computing.

Telepresence, Virtual Reality, Mixed Reality: Local real-time sensory input, such as video and graphics.

Machine Vision: Imaging-based automatic inspection and analysis for applications such as process control and

robot guidance.

Speech Recognition and Synthesis: Portable systems enabling recognition and artificial production of human

speech.

Nanoscale Systems and Manufacturing: Democratized, small-batch fabrication structures at the nanoscale using

a variety of material classes. For example, point-of-use nanoscale 3D Printers may provide desktop-sized

fabrication capabilities for rapid prototyping novel interfaces between traditional “hard” electronic materials and

“soft” biological materials.

Ultra-High Performance Wireless: Wireless systems with very low latency and extremely reliable

communications, for example, between autonomous vehicles.

Holistic Secure Systems: hardware-based defense in-depth, such as tamper resistant hardware that

electronically authenticates software integrity.

Computational Chemistry: Design of novel solutions for catalysis, low-temperature nitrogen fixation, etc.

Advanced Materials Science and Manufacturing: Simulation of solid state materials, etc.

Modeling and Simulation: Efficient exascale computing to enable advanced earthquake prediction (CMOS-based

high-performance computing capable of 1-10 exaflops), high-fidelity weather modeling (superconducting-based

hyperscale computing capable of 10-100 exaflops), and optimization problems (quantum computing).

Space Technologies: radiation hardness through circuit design and technologies (e.g., wide-bandgap electronics)

rather than special manufacturing processes (e.g., insulating substrates or shielding).

Ensuring Long-Term U.S. Leadership in Semiconductors

Science and Technology Policy, National Security Council) and relevant Agencies (Department of

Defense, Department of Commerce, Department of Energy, Department of Health and Human Services,

Department of Transportation, Department of Homeland Security, National Science Foundation).

This interagency group would coordinate the selection and prioritization of the moonshots

across the U.S. government. For each moonshot, the group would identify: (i) a lead agency to

host the moonshot, (ii) the resources required for achieving the moonshot, and (iii) the timeline

for execution. The group should focus on moonshots that have common commercial and U.S.

government interests. In each case, the moonshot should be chosen remembering that the goal

is to drive advances in systems and component technologies that not only solve the particular

problem, but also provide a general capability that can be subsequently used to address other

problems without completely new breakthroughs.

In addition, for each moonshot, an advisory group across industry, government, and academia

should be established. The purpose of this group would be to help identify the right people and

resources to address the moonshot. The group would help identify companies and people to

engage, determine sources of funding and other resources, and provide additional feedback and

support to the lead agency throughout the process. A similar approach was used by the

National Nanotechnology Initiative to launch a Grand Challenge for Future Computing.

The Subcommittee, along with the lead agency, would determine the appropriate government

tools for achieving the moonshot. A standard approach is government acquisition, where

agencies contract mission-relevant products or services, and the contractor performs applied

research and development to bridge the gap between off-the-shelf products and the mission

requirements. In many cases, the technologies developed to meet the mission requirements

are also valued by other agencies and the commercial marketplace. While the development of

new technologies is often incidental to product acquisition, new methods are emerging that

explicitly recognize technology development. For example, the “progress payment” model

specifies incremental payments when technical milestones are reached, which may be beneficial

for semiconductor-related technology development. For several additional ideas, see Box 3.

Tools to Reach Moonshots.

The following three examples of moonshots show further how such an effort could catalyze

semiconductor innovation. These examples illustrate the integration and application of the guiding

principles, and have been developed with the goals of capturing the imagination, motivating ambition,

and catalyzing innovation along multiple dimensions. They are offered as examples and starting points

for discussion and debate—including over the feasibility of executing the moonshots within the next

decade—and for creative thinking about additional moonshots.

1. Development of a zero day bio-threat detection network, which would identify previously unknown

threats (for example, some unexpected types of biological or chemical agents). This would require

innovation in the following areas: (i) design of advanced and low-cost bio-sensors, (ii) real-time data

The goal of the Grand Challenge is to “create a new type of computer that can proactively interpret and learn

from data, solve unfamiliar problems using what it has learned, and operate with the energy efficiency of the

human brain.”

Ensuring Long-Term U.S. Leadership in Semiconductors

analytics based on classic and machine learning algorithms to detect threats and anomalies, (iii)

ultra-secure and encrypted communications including methods such as post-quantum cryptography,

and (iv) real-time communications through this sensor network for rapid dissemination of threat

information.

This effort could be sponsored by Department of Homeland Security (DHS) (lead

agency), National Institutes of Health (NIH), and Defense Advanced Research Projects Agency

(DARPA). The challenge could involve engaging with established and startup companies through

prizes and procurements of products for both domestic and battlefield monitoring of bio-threats.

While the ability to identify all zero-day bio-threats is unlikely to be achievable in the next ten years,

that should not deter innovators from driving toward transformative progress on that problem. The

technologies developed here could ultimately be applied more broadly in healthcare, battlefield

sensing and communications, and population health surveillance activities.

2. Development of a high efficiency, domain-focused architecture to produce a system for 1-kilometer-

resolution Global Weather Forecasting that consumes less than 5 megawatts (i.e., has unusually low

power needs); is not CMOS-based; includes non-Von Neumann computing elements that allow for

parallel processing (non-von Neumann architectures) and/or approximate computing methods; and

can be designed and programmed using high-level programming models.

(Prescribing a set of

technological solutions here would help ensure that this effort yields more broadly valuable

advances. Recall that the goal is to yield innovative, generally reusable component technologies and

systems, and not a hard-coded machine to do weather forecasting.) This could be sponsored by the

National Oceanic and Atmospheric Administration (NOAA) (lead agency), Department of Energy

(DOE), and Department of Defense (DOD). The challenge could involve engaging with Federally

Funded Research and Development Centers (FFRDCs) and National Labs along with university and

corporate teams to design competitions with prizes and subsequent procurement with progress

payments to actually deploy the winning concept. This project would have many potential benefits,

both as a result of fine-grained weather event forecasting globally, but also the ability to deploy this

regionally and perhaps even aboard ships in a reduced capability way. New programming

architectures developed here could also be applied to other important problems, potentially

reducing cost, power and time-to-market for those solutions.

3. Development of commercial quantum computers capable of handling computational chemistry and

materials science problems needed to develop and deploy a pilot version of a high-scale, zero-

carbon, cost-competitive energy system to power a large government base or National Lab.

This

work could be sponsored by DOE (lead agency) and DOD and build on existing interagency efforts,

Post-quantum cryptography covers methods that are secure against attack by a quantum computer.

Non-Von Neumann computing elements are computing elements that allow for data-driven parallel processing;

Existing efforts should build on prior work, including: www.dhs.gov/health-threats-resilience-division

www.dhs.gov/biowatch-program; www.dtra.mil/Research/Chemical-Biological-Technologies; and

www.dtra.mil/Portals/61/Documents/CB/BSVE%20Fact%20Sheet_04282015_PA%20Cleared.pdf.

Such a quantum computer should be gate-based (i.e., have a quantum analog for the classical logic gate in a

conventional digital circuit), providing for the general purpose programming of quantum circuits to implement

arbitrary algorithms. The physical qubits (the quantum analog of the classical bit, which tend to have impractically

short lifetimes) could have various constructions, but after error correction should provide a minimum of 100

logical qubits and grow exponentially larger as the technologies mature.

Ensuring Long-Term U.S. Leadership in Semiconductors

including the National Strategic Computation Initiative. The challenge could involve engaging with

FFRDC and National Labs along with university and corporate teams to design competitions with

prizes and subsequent procurement with progress payments to actually deploy the winning concept.

Relatively small, commercial quantum computers would be able to solve modeling problems in

physical systems that cannot be solved by even the largest traditional computers. The underlying

technologies would lead to advances in chemistry and materials that could have broad application

across the economy, including in energy storage, generation, transmission, and pollution-related

issues This exemplifies the concept of a moonshot in that it is driven by a big problem that matters,

and fundamental semiconductor innovations, given that, to the best of our knowledge, these

problems cannot readily be solved by current technologies or envisioned extensions of current

technologies, . This would also allow the beginning of the process of developing a cadre of

technologists and scientists who will lead the world in the challenges associated with using these

radically different forms of computation. Quantum computers are also likely to have application in

machine learning and artificial intelligence applications.

Box 3. Tools to Reach Moonshots

There are a number of best-practice models that have already been tested within the government

and can be used to reach moonshots. While a typical tool used is government procurement, here are

four additional tools that could be drawn on:

1. Incentive Prize. An agency could host an incentive prize to encourage teams across industry

and academia to solve a moonshot that agency is hosting within a timeframe. The winning

team of the incentive prize should be rewarded in a way that allows them to actually deploy

the winning concept – such as through funding or a government procurement agreement.

This model has been used by many U.S. government agencies to address a wide variety of

technology challenges, including to accelerate the development of technologies for self-

driving cars, automated cybersecurity, and overcoming spectrum scarcity.

2. U.S. Government Fellowship. Once moonshots are selected, the Subcommittee on

Semiconductor Moonshots could bring on two fellows per project—one from industry and

one from academia—for 1-2 years to work together on assembling and running a team to

tackle the moonshot and on coordinating external resources (such as funding and research).

3. Collaborative Institute. The government could create a collaborative institute that brings

together resources from industry, academia, and government. These shared facilities could

support a range of activities, including early-stage research, small-batch production and

implementation, technology transfer, and loosely coordinated R&D. An example model is the

Manufacturing USA Institutes, which the U.S. government developed based on a PCAST

recommendation.

4. Industry-Led Venture Capital Consortium. The number of Series A deals for chip-only ventures

has been declining over the last 15 years due to the long time to revenue, the relatively large

investment required, few IPO successes, and less investment from traditional sources.

Industry could increase venture funding by creating a venture consortium that invests

collectively in next-generation technologies. If necessary, the U.S. government could make

anti-trust exceptions as it did in the 1980s for SEMATECH. The U.S. government could also

consider co-investing through organizations such as DARPA and the Advanced Research

Projects Agency-Energy (ARPA-E).

Ensuring Long-Term U.S. Leadership in Semiconductors

5. Conclusion

Semiconductor innovation long has been an engine of U.S. economic prosperity and national security.

Today it faces major technological, market, and geopolitical challenges. A concerted effort in

partnership among the Federal government, State and local governments, industry, and academia is

critical to fully meeting those challenges.

This report has focused mainly on recommendations for the U.S. Federal government; however,

important parts of what it proposes—particularly around long-term innovation—can be pursued by

industry, non-Federal government, and academia without Federal government involvement, albeit with

less success. Our proposal for loose coordination around moonshots, for example, could be led by non-

profit consortia that include industry and academic participants if the Federal government chooses not

to lead. Federal money, organization, and diplomatic influence can often fill gaps that private efforts

will never meet—but, in other cases, leadership from outside government can be just as effective.

That said, we strongly recommend a coordinated Federal effort to influence and respond to Chinese

industrial policy, strengthen the U.S. business environment for semiconductor investment, and lead

partnerships with industry and academia to advance the boundaries of semiconductor innovation.

Doing is essential to sustaining U.S. leadership, advancing the U.S. and global economies, and keeping

the Nation secure.

Ensuring Long-Term U.S. Leadership in Semiconductors

Appendix A. Moonshots – Methodology and Exemplars

This appendix expands the moonshot concept introduced in Recommendation 4.1 in the body of this report.

The methodology section provides context for how the moonshots were developed and a roadmap for creating

additional moonshots, either industry-led or government-led. The remainder of the appendix provides

additional example moonshots.

A.1 Methodology

Each moonshot vertically integrates technologies across the entire compute stack, from the top-level application

down to the component technologies, in order to sufficiently nurture transformative ideas and create a more

sustainable competitive advantage for the United States. The technology stack has the following traditional

layers (from top to bottom):

1. Ultimate Software Application

2. Application Programming Model

3. Platform Software Services

4. Platform Programming Model

5. Operating Systems Services

6. Computer System Architectures (processing, storage, and interconnect at every scale)

7. Component Technologies

Traditional CMOS semiconductor technologies and von Neumann system architectures have dominated system

development. This has been driven by the economic and performance implications of Moore’s Law, and a

robust ecosystem of consistent tools and a large workforce which has developed in support. As performance

gains in CMOS semiconductor technologies become more elusive, newer technology options are more

compelling than in the past, but the supporting ecosystem is often lacking. Moonshots based on combinations

of new technologies can jump start the development of the corresponding ecosystem.

A.1.1 Overall Approach

PCAST’s methodology assumes an application-driven approach to fostering innovation. Desired progress in a

particular application domain provides concrete goals for a particular moonshot, but the resulting advances in

the underlying technologies are almost universally more broadly applicable. Application domains considered by

PCAST are summarized in Table 1 in the body of the document. Moonshots are chosen so that at least some

experts in the field believe that achieving them within ten years is feasible.

After identifying a promising application domain, PCAST considered potential options for realizing the

technology stack described above. They considered options on three distinct dimensions: computing systems

architectures, computing modalities, and component technologies (Table A1). Ideally, a moonshot envisions

solution which requires a technology stack requiring interdependent innovations along one or more of these

axes, and which helps spur innovation beyond the current technology base. The goal is for these moonshots to

Ensuring Long-Term U.S. Leadership in Semiconductors

result in innovations that can be used for other general scientific, engineering, or economically valuable

advancements.

A.1.2 Computer System Architectures

• Von Neumann: Changes in technology to accommodate post-Moore’s Law realities, such as multi-core CPUs

with different, complex memory hierarchies, will demand new engineering paradigms across the existing

range of traditional Von Neumann architectures for digital computation.

• Quantum: Quantum Computing has the potential to substantially advance our compute capabilities and

solve currently intractable problems. There are several quantum architectural approaches which may

support different strategic domains, and along different timelines. These approaches, in rough order of

likely deployment, are: analog quantum simulation; adiabatic quantum annealing; and circuit-based

quantum computing.

• Bio/neuro-inspired (neuromorphic computing): Biologically-inspired power consumption and “topology” of

the circuitry (using three dimensions, more like the brain), analogous to how radio networks are now

designed in the post-Shannon Limit era.

• Analog computing: Analog computing approaches predate digital computing and in theory can solve some

problems that are intractable on digital computers. In practice, digital computing techniques have

overtaken analog computing, but advances in noise minimization could allow solutions in some areas.

• Special purpose architectures: Field-programmable gate arrays, graphics processing units, and deep

learning/machine learning accelerators, including for edge computing.

• Approximate Computing: Performing bounded approximation instead of exact calculations for error-

tolerant tasks (such as multimedia processing, machine learning, and signal processing), significantly

increasing efficiency and reducing energy consumption.

A.1.3 Computing Modalities

• Embedded systems: Specialized semiconductors, ranging from high-volume/low-cost for applications like

Internet of Things (IoT) devices to low-volume/high-cost semiconductors for robotics or defense systems.

Power efficiency requirements will vary by application (harvesting energy from the ambient environment

versus dedicated power sources, respectively). Flexibility and agility in fabrication and design will be needed

to maintain profitability.

• Personal/Portable systems: Desktop, mobile, and wearable computing devices. These are frequently

battery-powered computational devices, which will be optimized for performance, price, and power

efficiency. General purpose computing will be augmented by accelerators, sensor add-ons, and other

function augmenting ICTs.

• Hyperscale systems: Supercomputing devices for “remote” computation that will be aggregated to form the

most powerful systems that can be produced in each architectural class. These systems are expected to

solve otherwise intractable problems; or, for classical architectures, to maximize performance within

practical power constraints. Emerging architectures providing new capabilities and domain-specific

optimizations will become increasingly important as performance increases lag and practical power limits

are reached in traditional computing architectures.

Ensuring Long-Term U.S. Leadership in Semiconductors

Table A1. Selected component technology vectors that have a high probability of deployment in ten years

(* denotes more speculative deployment within this timeframe)

Component

technology

vector

Time-frame to

first commercial

products

Approach to achieving and retaining competitive

advantage

Neuromorphic

Computing

Available now

Continued R&D into new architectures coupled with 3D

technologies and new materials, Deep Learning accelerators (for

mobile and data center applications), and applications for true

brain-inspired computing

Photonics

Available now

Foundries for tools and materials R&D; integrate photonics with

CMOS and other materials

Sensors

Available now

Foundries for tools and materials R&D; integrate new

types/classes of sensors with CMOS and other materials

CMOS (sub 7nm

node size or new

3D structures)*

Advances in thermal

management available

with new process nodes

Deep understanding of transistor physics and chipset architecture

and related design know-how; foundries and labs for transistor

and materials R&D

Magnetics

1-2 years (MRAM as

eFlash), 3 years (as DRAM),

5-7 years (as SRAM)

Foundries for tools and materials R&D; integrate magnetics with

CMOS and other materials

2-3 years (wafer-to-wafer

stacking), 4-5 years (die-to-

wafer stacking), 5-7

(Monolithic 3D)

Deep understanding of applications space and benefits associated

use of 3D technologies and design know-how; foundries for tools

and materials R&D; design automation tool R&D

Data-flow based

architectures

3-4 years

Continued architecture R&D, coupled with materials, integration,

and manufacturing; build an ecosystem for solutions using data-

flow based architectures

Ultra-high

performance

wireless systems

3 years (5G), 10-12 years

(6G)

Continued R&D in new materials and processes, antenna design

advances, chipset manufacturing, and integration

Advanced non-

volatile memory as

SRAM

5+ years

Deep understanding of applications space and chipset

architectures

Carbon nanotubes

and phase change

materials*

5-7 years

Foundries/labs for materials R&D for hardware architectures;

chipset designs to leverage these technologies

Biotech/human

health

5-10 years

R&D towards low power, highly integrated, high performance

processing, high-data rate communications, wireless charging;

couple R&D with clinical research to create, build, and evaluate

on new materials and interfaces

Quantum

Computing

< 10 years

Pre-competitive R&D labs for new materials; foundries for new

materials and hardware architectures; tools for quantum

algorithms and software programming with various architectural

paradigms

Point-of-Use

Nanoscale 3D

printing

Available now

Desktop fab capabilities for rapid prototyping, additive

manufacturing, moving beyond silicon and interfacing with soft

matter, and small batch production

DNA for compute

and storage*

10+ years

Multi-disciplinary basic research in efficiently and reliably reading

and writing and retrieving DNA strands

Ensuring Long-Term U.S. Leadership in Semiconductors

A.2 Sample Moonshots

A.2.1 Bioelectronics for sensory replacement and implantable neuro-stimulation for control

of chronic conditions.

Context: Implantable medical devices are currently used for treatment of a limited set of conditions, such as

Parkinson’s disease, but their application is limited by size, power consumption, and inefficient electronic-

biological interfaces.

Goal: With a concerted effort by the electronics and semiconductor industry, clinicians, surgeons, and

neuroscientists, we can develop implantable medical devices for important new classes of conditions, including:

• Restoring sight or hearing to someone with sight or hearing loss by connecting implantable devices for

vision and sound to the visual or auditory cortex

• Alleviating chronic conditions, including pain and all auto-immune conditions, with tiny implants and

non-invasive surgery

Challenges/Innovation: Achieving these goals would demand simultaneous advances in semiconductor

technologies:

• Ultra-miniaturization to enable non-invasive surgery and enable implanting in the brain or on the

peripheral nerves

• Enhanced processor performance to enable executing complex algorithms such as those based on

machine learning or sophisticated signal processing in the implant

• Ultra-efficient electronics to enable implants powered by energy harvesting or fast and extremely

convenient wireless charging (such as through ultrasound)

These advances would need to be paired with application-specific innovations, including:

• Materials for improved neuron/nerve interfaces and power transfer between the electronics and the

brain/nervous system

• Advanced robotic surgery and surgical tools, to handle the small implant reliably

Key Government Stakeholders: The National Science Foundation (NSF) and Health and Human Services (HHS).

Potential supporting roles for Defense Advanced Research Projects Agency (DARPA), Intelligence Advanced

Research Projects Activity (IARPA), and National Institute of Standards and Technology (NIST).

Other Potential Applications: Advances in sensors may support “smart dust” technologies, where microscopic

sensors that float in the air record and transmit basic measurements for environmental control or enhanced

weather prediction. Enhanced processor performance and efficiency would likely benefit hyperscale systems,

which maximize performance within practical power constraints. Advanced materials may also apply to neuro-

inspired computing platforms. Advances in robotics will support a broad range of national needs, including

Ensuring Long-Term U.S. Leadership in Semiconductors

advanced manufacturing, logistics, services, transportation, homeland security, defense, medicine, healthcare,

space exploration, environmental monitoring, and agriculture.

A.2.2 Threat Detection Network

Context: The technology to develop and deploy biological, chemical, and nuclear threats has become

increasingly accessible, but our ability to detect these threats has not kept pace. Smart phones and IoT devices,

in combination with cellular service and the Internet, form a vast sensing and communication network that

could offer early detection and efficient warning for these threats.

Goal: Develop a high-speed biological, chemical, and/or nuclear threat detection network through deployment

of electronic devices that incorporate chemical, bio-chemical, spectral imaging, and radiation sensors in addition

to sensors for primary functions, that would cut detection times by an order of magnitude. Benefits would

include:

• Early detection of toxic biological, chemical, and/or nuclear materials could facilitate intervention by

first responders before deployment and ensure medical personnel are appropriately equipped for an

evolving event

• Ubiquitous sensing and robust communications after a nuclear event would facilitate appropriate

guidance to the public based on location (e.g., whether to shelter in place, or safest escape routes)

reducing casualties from fallout

Challenges/Innovation: Achieving these goals would demand simultaneous advances in semiconductor

technologies:

• Advances in special purpose processor design to support real-time data analytics based on classic and

machine learning algorithms in edge devices to detect biological threats

• Ultra-Secure cryptographic algorithms to simultaneously authenticate sensor data and protect privacy of

device owners

These advances would need to be paired with application-specific innovations, including:

• Design of advanced and low-cost sensors (chemical, biochemical, multi-spectral imaging, and others)

that are also easily integrated into devices such as smartphones, automobiles, buildings, security

cameras, and other IoT devices

• Real-time communication protocols for rapid dissemination of threat information

Key Government Stakeholders: The Department of Homeland Security (DHS) and NSF. Potential supporting

roles for DARPA, IARPA, and NIST.

Other Potential Applications: The sensor advances developed in this moonshot would also contribute to the

realization of “smart dust” technologies (see moonshot A.2.1 for details). Advances in special purpose processor

design to support real-time data analytics and machine learning in edge devices would offer significant benefits

in autonomous systems such as self-driving vehicles.

See: www.whitehouse.gov/blog/2011/06/24/developing-next-generation-robots.

See: www.darpa.mil/news-events/2016-08-23.

Ensuring Long-Term U.S. Leadership in Semiconductors

A.2.3 Distributed Electric Grid

Context: Renewable energy offers significant benefits to society, but these sources amplify long standing

electrical grid challenges such as balancing demand and supply. Traditional power plants generate power

efficiently in a steady state, but are costly to bring on- and off-line. A decentralized power system, where

distributed generation meets local demand, or distributed electric grid has the potential to better leverage the

complete portfolio of power technologies.

Goal: Through a suite of advances in novel power generation methods and energy storage (e.g., batteries),

application of artificial intelligence and machine learning to energy management, standalone energy systems,

local energy distribution, and real-time communications protocols, technology advances would accelerate the

realization of distributed electrical grid that would:

• Efficiently leverage renewable energy resources, including cost-effective storage when supply exceeds

demand

• Rapidly and efficiently add power from storage traditional technologies when renewable energy sources

are not available

Challenges/Innovation: Achieving these goals would demand simultaneous advances in semiconductor

technologies:

• Advances in special purpose processor design to support real-time data analytics based on classic and

machine learning algorithms in edge devices to optimize electrical usage and minimize cost

• Design of advanced power electronics, utilizing wide bandgap semiconductors, for inverters, frequency

control and voltage stability

These advances would need to be paired with application-specific innovations, including:

• Energy storage technologies with greater energy density and cost efficiency

• Standalone energy systems with a mix of sources to counter unpredictability of energy generation (and

integration of energy generation and consumption with as co-generation and electric vehicle-to-local

grid)

• Local energy distribution grid and related communication network

Key Government Stakeholders: The Department of Energy (DOE) and Nuclear Regulatory Commission (NRC).

Potential supporting roles in energy management for the Federal Energy Regulatory Commission (FERC) and in

semiconductor technology research for NSF, DARPA, IARPA, and NIST.

Other Potential Applications: Advances in special purpose processor design to support real-time data analytics

and machine learning in edge devices would offer significant benefits in autonomous systems such as self-

driving vehicles. Advanced electronics for inverters, frequency control and voltage stability would likely benefit

advanced manufacturing facilities. Advances in energy storage capacity would benefit electric vehicles and a

myriad of consumer products.

A.2.4 Global Weather Forecasting

Context: Due to the longstanding success of Moore’s Law, the computing and semiconductor industry are

optimized for a world where general purpose microprocessors provide dependable performance increases and

Ensuring Long-Term U.S. Leadership in Semiconductors

cost reductions to von Neumann based applications. As maintaining Moore’s Law becomes more difficult,

industry must look to new sources for performance increases and cost reduction.

Goal: Develop a global weather forecasting system with a fidelity of 1 km and power consumption less than 5

MW using an innovative, high-efficiency, domain-focused architecture. This level of fidelity would support

explicit inclusion of cumulus clouds in weather simulation and modeling. Advances in systems architecture,

along with new design methods and tools for hardware and software development, required to field such a

system would also offer:

• Alternative component technologies to achieve the performance increases but with much lower power

than historically provided by Moore’s Law-advances.

• Support innovation through integration of radical computer architecture, and attendant software

technologies while retaining the potential for re-use in other applications.

Challenges/Innovation: Achieving these goals would demand simultaneous advances in the following

semiconductor technologies:

• Advances in theory and tools for system architects to create domain-focused architectures

• Advances in design tools that reduce the cost and time for hardware design (e.g., special purpose

processor design) to parity with software development

• Advances in hardware design tools that enhance portability so that existing weather mode designs and

related software algorithms can move to radically new underlying technologies, greatly accelerating our

ability to bring new integrated circuit technologies to the market

These advances would need new programming approaches that are less dependent on architecture or hardware

but fully leverage the performance of the underlying system

Key Government Stakeholders: The National Oceanographic and Atmospheric Administration (NOAA) and

Department of Defense (DOD).

Other Potential Applications: Tools for rapid and portable hardware design will support innovative hardware

design for all strategic domains and will accelerate the adoption of emerging integrated circuit technologies.

New programming approaches that ease programming of special purpose devices will accelerate their adoption

in low cost (e.g., IoT) devices, accelerating a shift to processing in edge devices. New programming approaches

that are less dependent upon system architectures would also benefit traditional supercomputing, alleviating

workforce shortages and reducing the cost of porting software to successive generations of systems.

President’s Council of Advisors on Science and

Technology (PCAST)