《布鲁金斯学会:美国的半导体战略(英文版)(47页).pdf》由会员分享,可在线阅读,更多相关《布鲁金斯学会:美国的半导体战略(英文版)(47页).pdf(47页珍藏版)》请在三个皮匠报告上搜索。
1、Christopher A.ThomasOCTOBER 2022A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESFOREIGN POLICY AT BROOKINGS1EXECUTIVE SUMMARY Semiconductors are the lifeblood of the digital economy.The semiconductor industry has moved to the foreground of political discourse both in the United States and other countr
2、ies.The pushes from Americas economic rivals and the challenges faced by its own domestic industry,coupled with supply chain shortages,prompted calls for the U.S.government to“do something”to support the industry.The most visible response is the CHIPS Act,which allocates$39 billion in government fun
3、ding for domestic semiconductor manufacturing facilities and billions more for semiconductor research and development(R&D)and workforce programs.1Many cite Americas declining share of semi-conductor manufacturing as the justification for such measures,with“unfair”subsidies by other countries as the
4、root cause.Much of the debate has centered on evaluating what other countries are doing and matching their programs.Of course,benchmarking is not a strategy.Voluntarily exiting global markets is not a strategy.Fixing short-term product shortages(in ways the industry cannot)is not a strategy.In fact,
5、the argument that the industry needs funding to fix shortages is ringing hollow at the time of this writing,as demand for semiconduc-tors for personal computers and smartphones drops.2 This“downcycle”is not only good for consumers of these products,but it also shifts the policy debate to a more appr
6、opriate objec-tive:How does the United States build a sustain-able,market-centric semiconductor policy that leverages the strengths of the American financial,industry,and academic environments to collectively accelerate the industry and not just for a few years,but in the decade to come?How does the
7、 United States ensure that global competition in semiconductors does not devolve to“zero-sum”negotiations around shifting manufacturing capacity,but rather that competition brings out the best in America:its ability to harness the worlds best scientists and entrepreneurs to solve hard technical issu
8、es,build business around those solutions,and scale those solutions to the world?In short,how does the United States build a government strategy that is open,global,long-term,committed,patient,and successful?To do so,the policy must recognize that competitive advantages do not come from emulating the
9、 approaches of others(for which U.S.capabilities are not a fit)but rather from deepening the existing advantages of the U.S.position.Those U.S.advantages are deep and broad,and not to be underestimated.I recommend that the government policy not solely focus on increasing Americas manufac-turing capa
10、city,but rather holistically strengthen the entire semiconductor industry,enabling it to withstand supply shocks,drive technology tran-sitions,and win future industry control points.Fundamental research and the commercializa-tion of R&D breakthroughs are the ingredients for future success and will d
11、etermine the global semiconductor manufacturing footprint as much as will subsidies.I recommend using the CHIPS Act funding as equity capital in a government fund that can scale via industry and Wall Street co-in-vestment to more than$300 billion and can reduce industry cost of capital by leveraging
12、 the Federal Reserve balance sheet.This fund would be self-replenishing,as it would harness U.S.innovation to fund projects that have market-level rates of return and generate significant returns for the government.These returns would then be reinvested in the next set of challenges the United State
13、s faces three,four,five,and 10 years from now.As opposed to copying policies that place all hope on singular national champions,often saddling them with policy goals that may or 2A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESmay not be achievable,the equity fund would have tiers of financial and ind
14、ustrial partners enabling funding to be provided to both small and large companies and would operate at all levels of the value chain and across the ecosystem.I would urge this U.S.government fund not to compete with incentives from other countries,but would instead encourage it to partner with thos
15、e willing to co-invest transparently in a growing,global,diverse,de-risked,and market-driven semi-conductor industry.The resulting robust global supply chain,populated with more clusters and second supply sources across Europe and Asia,would only help the United States.In concert with creating this
16、fund,the United States could address other barriers to success.The country simply does not have enough engineers to build and ramp the manufacturing facilities in the plan targeted and accelerated immigration must start now.Constructing and ramping fabrication plants(fabs)in the United States takes
17、up to one year longer than it does in Asia.This self-inflicted slow pace,if not solved,will cost billions in lost opportunities and technology leadership for those companies building in the United States thus countering any benefit from the billions financed by the government.The fund would have a p
18、olicy arm that partners with federal and state governments to aggressively simplify permitting requirements and close timing gaps.Too few entrepreneurs,professors and venture capitalists are taking risks on future semiconductor technologies and applications government funding can be a catalyst to re
19、verse this trend,without giving the fund the mandate to pick winners.Finally,to effectively execute a long-term,committed,and patient investment program,the United States needs a new hybrid govern-ment team that can evaluate,structure,and monitor investments at the intersection of semiconductors and
20、 finance.The government needs to rapidly recruit this team from the semiconductor,financial,and policy spheres and insulate the team(via legislative action)from short-term political considerations while maintaining the oversight capability of elected leaders.Empowered as the primary point of contact
21、 for the execution of the U.S.semicon-ductor strategy,this team would ensure speed,consistency,and clarity in its role as the deci-sion-making authority.It would operate through different administrations,through industry cycles,through new generations of technology,and through changes in geopolitica
22、l priorities;and in doing so,it could continuously partner with the global industry to achieve a resilient,winning,global,and market-driven U.S.semi-conductor industry.FOREIGN POLICY AT BROOKINGS3INTRODUCTION AND SUMMARY OF POLICY RECOMMENDATIONSA semiconductor,a substance that has specific electric
23、al properties,serves as the foundational foundation for computers,servers,mobile phones and all other electronic devices.It is typically a solid chemical element or compound that conducts electricity under certain condi-tions but not others.A conductor is a substance that can conduct electricity and
24、 a diode is a substance that cannot conduct electricity.Semiconductors have properties that sit between the conductor and insulator.A diode,transistor and an integrated circuit(IC)are all made from semiconductors.3The global semiconductor industry is critical to all advanced economies.The American s
25、emiconductor industry is also critical to the global leadership and economic well-being of the United States.The industry directly provides nearly 300,000 American jobs and indirectly supports more than 1 million jobs throughout the supply chain and development ecosystem.4 It represents one of the c
26、ountrys leading export industries,with an annual value of roughly$50 billion.It invests nearly$40 billion annually in R&D.The semiconductor industry also helps solidify U.S.national security,ensconcing U.S.leadership in cybersecurity and defense,as the frontier of semiconductors enables the advancem
27、ent of national security tools.And by ensuring that the United States can set the pace of global technology innovation,the semicon-ductor industry confers enormous strategic advantage in foreign and economic policy.5However,as laid out below,there are acute challenges to Americas leadership in semi-
28、conductors and there are imbalances in the U.S.supply chain that create geopolitical risk.U.S.-headquartered companies represent about 47%of global semiconductor sales,but the United States itself has only 10%of installed manufacturing capacity.There has been tangible progress over the last few year
29、s regarding increasing investment in the U.S.industry.Intel has proposed investing$20 billion in(i.e.a newly developed location)semiconductor plants in Ohio.6 The Taiwan Semiconductor Manufacturing Company(TSMC),the worlds largest semiconductor manufacturer,is building a fabrication plant in Arizona
30、.7 Two other large manufacturers,Texas Instruments and GlobalFoundries,have announced plans to build new capacity in Texas.8 But this is only a starting point for a new U.S.government strategy,not the endgame.U.S.government efforts need to accelerate this momentum and make it sustainable not just in
31、 response to the current industry issues but for the decades to come.THE PROBLEM STATEMENTIn the simplest terms,the U.S.semiconductor industry faces five problems:It is simply easier,faster,and cheaper to build scale front-end semiconductor manu-facturing in other countries;for more than two decades
32、,global and American compa-nies have shifted more and more of their investment to Asia.The upstream(e.g.inputs into the manu-facturing process)and downstream(e.g.circuit boards and other systems integrating semiconductors)industries supporting 4A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESsemicondu
33、ctor manufacturing are primarily located outside the United States,increasing the net delivered cost and time to market for the US-based end-to-end supply chain versus those in Asia.9 The U.S.venture capital market spendsnegligible amounts of its capital(1%)on new companies in the semiconductorindus
34、try and has never really been a majorcontributor to the leading-edge chip develop-ment business.10 Many more semiconductorstartups(albeit of widely varying quality)are established in China than in the UnitedStates.The United States overwhelming spendingadvantage in pathfinding R&D across mate-rials
35、and the physical,chemical,and elec-trical sciences has dwindled.While federalfunding of research has“only”dropped froman average of 1%of the gross domesticproduct(GDP)in the 1990s to about 0.7%in2020,actual funding for nondefense generalscience(the category most applicable tosemiconductors)rose by o
36、nly$500 millionor 5%over the last 10 years,a time framewhen the industry doubled in size and U.S.GDP rose by 50%.11The U.S.semiconductor workforce,while ofvery high quality,is aging and is growing at apace far too slow to ensure U.S.leadershipin the future.Geopolitical rivals have a playbook to buil
37、d up their nations semiconductor industries and have been working hard to execute it.Should the United States simply benchmark those efforts and match them,subsidy for subsidy and tax break for tax break?What would that entail?It would likely include establishing a top-down mandate for technology se
38、lf-reli-ance,calling on the industry to build the whole supply chain in America,suppressing wages,pressuring global companies to invest in the United States or lose market access,making the government the(de facto)largest shareholder of unprofitable or politically connected semicon-ductor ventures,f
39、orcing government-subsidized companies to do all of their R&D or manufac-turing inside the United States,pressuring customers to“buy American,”and allowing subsidized firms to maintain negative margins or negative returns on capital through blank check subsidies.This will not work.In an industry wit
40、h a complex,integrated,highly efficient global supply chain that simultaneously feeds the American chip industry and buys its products,any proposed requirements for self-reliance and localization will shrink the U.S.share of the industry.In an industry where success requires constant rein-vention,sp
41、eed,and access to the best talent,money alone is insufficient.And in an economy where capital efficiency is the highest objective,money delivered solely as grants and subsidies will never be enough.In addition,the political will to maintain subsi-dies across administrations and constantly changing c
42、ongressional periods is certain to lapse.It always has in the past.That constant uncertainty would itself undermine any positive lasting effects of the subsidies,as companies would continually have to plan for the cessation of government handouts in the sector.That is why the United States needs a n
43、ational semiconductor policy that is strategic rather than politically convenient.Strategy is about what you do differently than your competitors;it requires building upon core competencies,not wishing you had other ones.We need a policy roadmap that other countries,both allies and rivals,simply can
44、not emulate.The semiconductor strategy I propose builds on the strengths of the United States,which include the following:A large and sophisticated set of capitalproviders with expertise across technolo-gies in every stage of development;Wall Streets unparalleled financial engi-neering capabilities
45、to harness capitalefficiencies that other countries can onlydream about;The venture capital industrys distinct abilityto turn new technologies into profitable andscalable global business models;A body of laws and cases that provide theworlds best investor protections and disclo-sure requirements;FOR
46、EIGN POLICY AT BROOKINGS5The greatest pool of experienced andinnovative engineers,and a culture that isattractive to top global talent;Direct access to the worlds largest purchasersof semiconductors and largest end marketsfor semiconductor-enabled products;Access to semiconductor company leader-ship
47、 teams with a best-in-class ability to selland source globally;A deep and accomplished set of world-classresearch universities with experience inbuilding pathways to commercialize break-throughs;andMost importantly,a trusted relationshipwith the other countries and regions thatalso have semiconducto
48、r leadership andsubstantial end market demand.A strategy that accelerates these strengths and leverages them to shore up weaknesses via market-driven investment for example,by financing fab automation breakthroughs that reduce labor requirements has a much greater chance of success.A strategy that d
49、elivers more than one-off industry benefits will help engineer a long future of renewed growth,innovation,as well as onshore manufacturing leadership.BUILDING ON THE CHIPS ACT:A NEW NATIONAL SEMICONDUCTOR POLICY AND FUND FOR THE UNITED STATES The centerpiece of our proposal allocates a portion of th
50、e CHIPS Act semiconductor funding as equity capital in a domestic govern-ment fund(hereafter referred to as the U.S.Semiconductor Fund or the Fund).12 This equity,scaled via large amounts of industry and Wall Street co-investment,would drive$300 billion or more in total industry investment.The inves
51、t-ment would become self-replenishing,harness U.S.innovation to develop projects with market-level rates of return,generate significant returns and ancillary tax income for the government,and reinforce the global semiconductor supply chains need to have a substantial U.S.presence in both engineering
52、 and manufacturing.In other words,I envision the manufacturing incentives as well as a portion of the R&D funding in the CHIPS Act as permanent capital to be invested by experienced professionals on the basis of strategic and financial principles,not as a collection of subsidies and grants that comp
53、a-nies would bid for via government lobbying.The Fund would pay ever-higher dividends for the nation and the semiconductor industry by growing through low-risk,high-return invest-ments and thus increase impact over time.Co-investment is the first objective of the Fund.The recently approved$52 billio
54、n is not a game changer in an industry that spends nearly$300 billion annually in R&D and Capital Expenditures(TSMC alone has committed to investing more than$100 billion over just the next three years,while Samsung will invest more than$150 billion in leading logic manufacturing capacity over 10 ye
55、ars).13 Meeting Americas goals in even a single part of the supply chain,such as leading-edge logic front-end manufacturing,means adding five or more fabs in the next 10 years with each fab potentially requiring$20 billion in startup investment and then billions more in upgrades on a continuous basi
56、s.This one segment of the value chain alone could require far more than the current proposed funds and great investment is required in other segments as well.Without balance across the segments,all industry participants will question from day one the sustainability of a U.S.-centric global manufactu
57、ring hub.Hence,the scale of any national semiconductor policyneeds to match the scale of the industrysinvestment needs.That effort clearly cannot fallonly on the shoulders of government,but it willif there is no integrated solution that attractsthird-party capital to amplify the Fund.Continuity and
58、consistency is the second objective.Semiconductors are the ultimate long game.Few projects pay back in less than five years;the climb to leadership is hard and arduous;and once you reach the summit,you need to invest incredible amounts every year to stay there.In general,it takes more than 10 years
59、to research,design,and launch a new front-end manufacturing process technology.14 6A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESIn another example,it took Qualcomm the pioneer in code division multiple access(CDMA)and one of the worlds top 10 semicon-ductor companies 10 years to commercialize its t
60、echnology after founding.15 A one-off injec-tion of government capital can spark comple-mentary co-investment in the next few years,but what about the checks that come due in 2025,2030,and 2035?Under a new administration,a new Congress,and a new set of economic pres-sures,will there be political app
61、etite for another large appropriations bill?The United States is a country that took nearly two decades to pass a true,higher spend infrastructure bill.A perma-nent capital fund,designed to be self-sustaining,insulates investment decisions from short-term political winds and shifting national priori
62、ties and appropriation environments.Transparency and an equal playing field is the third objective.An investment fund with clear investment guidelines,decision-making rules,conflict-of-interest guardrails,and standard investor communication procedures would give both industry players and private co-
63、investors greater assurance,thereby reducing their risk calculus and increasing their desire to invest.This means that decisions would need to be well out of the reach of politicians aiming for their states to have an ever-bigger piece of the pie;instead,U.S.capitalism,global semiconductor companies
64、,and leading investors would need to drive the implementation and execution of the Fund.Companies would need to understand that if they put together a proposal that meets Fund guidelines and objectives,it will be accepted rather than rejected in favor of a proposal by a competitor that hires better
65、lobbyists.At the same time,politicians and voters would need to know how this capital is being put to work,that there is constant oversight of the Fund and governance structures,and that there are clear and regularly published measures on how well the Fund is meeting all the objectives.The final obj
66、ective is investment rigor,project due diligence,and execution accountability by enforcing the requirement for market-consistent returns.By bringing in third-party financial inves-tors,the U.S.Semiconductor Fund would benefit from their due diligence and their demands for market-consistent rates of
67、return;this would ensure that the investments are more likely to be successful and self-sustaining.And there is an added side benefit:such an effort would increase the pool of financiers who understand,invest in,and have a passion for semiconductor and related technologies.The leveraging of that kno
68、wledge and passion into new core technologies beyond semiconductors could have a multiplier effect.To fulfill these objectives,a new semiconductor policy(hereafter referred to as the National Semiconductor Policy or the Policy)would need to structure the Fund to draw in and tap the countrys top priv
69、ate equity and venture capital firms to de-risk government investment and provide an additional layer of both diligence and governance.To ensure investment into the broadest set of needed opportunities,not just“leading edge fabs,”the Fund would have a tiered structure that enables smaller investors
70、to write smaller checks to the small-and medi-um-sized businesses that support the overall semiconductor supply chain and ecosystem.To efficiently provide seed capital to the widest possible set of“new ideas and new technolo-gies,”the Fund would leverage existing tools like the Small Business Admini
71、strations Small Business Investment Companies(SBICs)to broaden reach to qualifying researchers and entrepreneurs.To protect government invest-ment,the Fund would have a capital allocation approach that puts government investment“at the top of the capital stack.”Finally,to spur nationwide investment
72、into diverse and inno-vative foundational research,the Fund would invest direct government research funds into a network of labs and universities that can drive fundamental research across all technologies needed to drive future industry transitions,from materials to software.Money,even if smartly a
73、llocated and well spent,is not enough to supercharge the American semiconductor industry.The industry needs to recruit more talent,from top Ph.D.s to tradespeople specializing in fab construction.Targeted interventions in immigration policy are needed to allow skilled manufacturing and R&D personnel
74、 who can help to build up these fabs and R&D facilities,and incentives are needed to ensure that many more U.S.-trained graduates with semiconductor-relevant doctorates remain FOREIGN POLICY AT BROOKINGS7in the United States after their study ends.A coordinated push to streamline the permitting and
75、approval processes for new constructions and expansions across jurisdictions and agen-cies is also needed;this could cut in half the time from initial inquiry to installed semicon-ductor tools on the fab floor.The United States also needs to market this effort via encouraging American talent to stud
76、y the hard sciences that underpin semicon-ductor success(such as electrical engineering,mechanical engineering,optical engineering,and materials science).Essentially,the United States needs to create a series of holistic incen-tives and policies,so that the capital invested in the U.S.Semiconductor
77、Fund goes much further and becomes an intergenerational,transforma-tive industry road map.The new Policy would also need to reimagine the United States technology collaboration with its allies.It is economically infeasible and opera-tionally impossible for the United States to build and maintain a t
78、op-class,end-to-end domestic supply chain that provides every input for every chip consumed by American businesses and consumers.Even a successful U.S.semicon-ductor industry will have to rely on a diverse and broad set of imported products and technolo-gies from a wide range of countries.Therefore,
79、diversifying and strengthening the U.S.supply chain can only be achieved when like-minded countries strengthen their own domestic indus-tries,diversify their supply chains internationally,and partner deeply with the American semi-conductor ecosystem.To drive such actions by global partners,the Polic
80、ys investments and the associated messaging need to center on innovation,new applications and new markets,rather than solely on risk reduction.The Policy needs to provide equal incentives to,and a level playing field for,American and non-American semiconductor companies;it should invite co-investmen
81、t and“club deals”with semiconductor funds driven by U.S.global allies.It also should encourage the United States to offer reciprocal market access and customer prioritization,so that both the United States and its allies know neither party will be shortchanged in times of capacity constraints.The Po
82、licy should also espouse the“global use”of technologies created in the United States.As many technology companies learned the hard way over the decades,proprietary technology that cannot be leveraged by global partners will eventually fail or be worked around.The United States cannot afford to make
83、the same mistake,even if in the short term,it makes for a domestic political win.This last point is controversial.Global collab-oration combined with the global free use of technology has helped the American industry reach its current level of success.But such collaboration will be difficult to achi
84、eve in the future if companies believe that the intellectual property(IP)created from R&D in the United States will de facto be excluded from global use.Limiting which U.S.-based technologies can be exported or used by companies head-quartered in our economic rivals has a negative impact on U.S.-bas
85、ed research investment by both local and global companies.The indus-tries built on leading-edge semiconductors(mobile communications,cloud,autonomous driving,artificial intelligence,the metaverse,and more)are based on robust ecosystems that demonstrate network effects.Each addi-tional participant st
86、rengthens the ecosystem for every other participant.The removal of a substantial number of(Chinese or other)participants from the American ecosystem weakens the ecosystem and decreases the ecosystems attractiveness as an investment destination.As the removal of any participant reduces the ROI of all
87、 R&D spending across the entire American ecosystem,export controls need to be used in transparent,judicious ways.It is important to thoughtfully execute a U.S.export control strategy that is balanced with the highest possible return on investment for R&D performed in the United States.These actions,
88、in sum,will drive the industry and U.S.allies to view efforts related to the U.S.National Semiconductor Policy and the Fund as open,global,long-term,committed,and patient.Not only is this the positioning required for success,it is the positioning that U.S.economic rivals will have difficulty adoptin
89、g due to the differences in their economic and financial systems.8A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESFinally,along with this new strategy,I propose a new approach to implementation.Without an executable plan,the U.S.Semiconductor Fund and the corresponding investment and policy strategy w
90、ould undoubtedly fail through either negative returns or a lack of focused coordination.There are many ways this could happen:if responsibility for the National Semiconductor Policy is spread across multiple agencies and teams;if there is no centralized decision-making;if Policy objectives are fuzzy
91、 and can be manipulated for political gain;if there is no continuity of investment because the decisionmakers keep changing;if invest-ment decisions are constantly being shifted by short-term legislative horse trading;if invest-ment managers change every time there is a new administration or the con
92、trol of Congress changes;and if investment decisions are made by teams lacking industry expertise.Even if just one of these instances were to become a reality,it is hard to envision the Fund being successful.How can the United States ensure that a good strategy is not undone by poor execution?First,
93、I propose that the U.S.government appoint and empower a single organization to be the point of contact for the semiconductor and financial industries;the organization would coordinate and orchestrate semiconductor policy across all the relevant government agencies.Second,I propose that the governmen
94、t model this orga-nization comprised of financial,industry,and policy experts on the successful Overseas Private Investment Corporation(OPIC).OPIC leveraged market-based investment programs,permanent investment teams,a transparent set of investment guidelines,and impactful and rigorous government ov
95、ersight.It persisted through multiple administrations,invested at scale,returned capital to taxpayers,and effec-tively took on both private sector and public sector roles.The remainder of this report provides further details on the Policy and the Fund.FOREIGN POLICY AT BROOKINGS9SETTING THE BACKGROU
96、NDBuilding a more robust semiconductor industry via government policy will not be easy.Spending money alone is not enough to dramat-ically change the semiconductor industry nor its global manufacturing footprint.China has been funding domestic semiconductor projects at a rate far greater than CHIPS
97、envisions.Yet in the eight years since the launch of Chinas effort to boost its semiconductor industry(the so-called“Guideline for the Promotion of the Development of the National Integrated Circuit Industry”16 and the complementary“Big Fund,”a vehicle for funneling capital to its domestic chip indu
98、stry),Chinese policy has not delivered a major shift in semiconductor market share,product leadership,or manufacturing footprint.For every successful government investment,there are far more failures,and the Chinese government has detained or is investigating many of the fund managers and recipients
99、 of government investment vehicles.17 There is little evidence that the policy attracted substantial foreign talent or capital.In the five years prior to the policy launch,four major foreign manu-facturers built greenfield fabs in China;since the launch of the policy,only one major global semiconduc
100、tor company broke ground on a greenfield fab at technologies three genera-tions behind the leading edge.18 Despite major investments,the share of chips supplied by China-based fabs to China-based customers increased only slightly between 2011 and 2021:from roughly 13%to around 17%.China-based semico
101、nductor production still represents much less than 5%of the global total.19The challenges to any countrys industry are complex and intertwined.SEMICONDUCTORS:A HIGHLY DIVERSE,MULTI-ECOSYSTEM GLOBAL INDUSTRYThe manufacturing chain for any given semicon-ductor is extraordinarily complex and relies on
102、up to 300 different inputs,including raw silicon,commodity chemicals,specialty chemicals,and bulk gases.All of these inputs are processed and analyzed by upwards of 50 different types of processing and testing tools(see below for the taxonomy of the industry value chain).Those tools and materials ar
103、e sourced from around the world and are typically highly engineered.Further,most equipment used in semiconductor manu-facturing,such as lithography and metrology machines,rely on complex supply chains that are also highly optimized and incorporate hundreds of different companies that deliver modules
104、,lasers,mechatronics,control chips,optics,power supplies,and more.Finally,economies of scale and learning efficiencies have consolidated the industry,leading to a high concentration of market share being held by one company at each level of the value chain.The installed base of tools and equipment w
105、ithin a semiconductor factory today represents the cumulation of hundreds of thou-sands of person-years of R&D development.The manufacturing process that integrates these tools into a single manufacturing chain could represent hundreds of thousands more.Each company contributing to that ecosystem is
106、 taking action based on market opportunities,engineering feasibility,and financial returns,not government objectives.20 In addition,different segments comprise a different set of companies,capabilities,key success factors,technological hurdles,and capital needs.10A SEMICONDUCTOR STRATEGY FOR THE UNI
107、TED STATESINDUSTRY VALUE CHAIN LAYERSThe following is a simplified summary of the semiconductor value chain and its seven key layers:1.Manufacturing process technology the expertise and IP that enables the manufac-turing capacity.2.Fabless design the design of complete semiconductors,relying on thir
108、d-partyoutsourced manufacturers to fabricate them.3.Front-end manufacturing the large-scale fabrication of wafers.4.Outsourced assembly and testing(OSAT)the dicing of the manufactured wafers,testing of the chips,and packaging of the chips so that they can be further integratedinto an electronic syst
109、em.5.Equipment and tools the large machines that process and test the wafers.6.Materials and chemicals the highly engineered raw materials that make up the phys-ical structure of the final semiconductor and/or are used in the manufacturing process.7.Electronic design tools the software packages that
110、 help companies design chips andprepares the physical layout of transistors such that the front-end manufacturing facili-ties can physically make the chips.In examining the challenges to the U.S.industry,it is helpful to know the rough global market share of U.S.-headquartered companies in each laye
111、r of the value chain.FIGURE 1U.S.relative market share by value chain layersSource:McKinsey&Company21FOREIGN POLICY AT BROOKINGS11Likewise,it is helpful to narrow the focus to high-level industry segments,as the vast number of product types can be confusing.The dozens of semiconductor product catego
112、ries can be grouped into the following major segments:Leading-edge logic high-end centralprocessing units(CPUs)and graphicsprocessing units(GPUs)designed to powerthe most advanced phones,personalcomputers,and servers.Lagging-end logic lower capability,lessexpensive processing units designed forappli
113、cations with lower requirements(consumer electronics,large householdmachines,automotive sector).Memory semiconductors designed to holddata and usually made with the most lead-ing-edge processes.Analog and power specialty semiconduc-tors for power management,radio transmis-sion,wireline transmission,
114、and other nicheapplications.Discrete nonintegrated chips,such asresistors,capacitors,and diodes.Optoelectronic semiconductors designedto handle both electrical currents and light.Photonic components for creating,manip-ulating,or detecting light,such as laserdiodes,light-emitting diodes,and solar and
115、photovoltaic cells.“BEYOND THE CHIP”ECOSYSTEM:AS IMPORTANT FOR SUCCESS AS THE SEMICONDUCTOR PRODUCT ITSELFA healthy semiconductor industry requires delivering far more than just physical chips.The larger ecosystem that turns a chip into an end-user application is essential to semicon-ductor success.
116、System-level technologies such as operating systems,development frameworks,application software,as well as physical and virtual interface standards,define how a chip becomes a usable product in the hands of an end consumer.These system-level technologies create control points whereby companies or co
117、nsortia can exert influence on the end-to-end supply chain.Semiconductor vendors that individually,or via partners,define control points improve their product competitiveness,increase switching costs for competitors,and enable global scale.Therefore,leading-edge logic semiconductor vendors such as I
118、ntel,NVIDIA and Qualcomm increasingly define,deliver and package system-level technologies as part of their semiconductor offerings.They are some of the largest employers of software engineers in the United States and employ more software engineers than they do hardware engineers.22 In parallel to t
119、his business model expansion by semiconductor vendors,major consumers of semiconductors,such as mobile phone makers,PC vendors,and cloud service companies,have entered or will enter the fabless design segment of the semiconductor industry.These moves are blurring the traditional differences between
120、the“providers”and the“users”of semiconductor technology there is more and more overlap.The global technology industry is always in a process of transitioning to new end-to-end capa-bilities,often on multiple vectors at once(for example,the industry is transitioning currently from 4G to 5G technologi
121、es,from on-premise to cloud computing,and from standard to artificial intelligence workloads in high performance computing systems).Each of these transitions require new system-level technologies,creating new control points for semiconductor vendors to win or to lose.Maintaining American leadership
122、across these control points from generation to generation requires the entire ecosystem that defines,develops and scales technology transi-tions to be as strong as possible.Strengthening the ecosystem requires coordinating fundamental research to define the end-to-end system-level roadmap,accelerati
123、ng that roadmap into commer-cial production,and engaging downstream“beyond the chip”industries to use the system-level technologies espoused by the American semiconductor industry;even if major participants in those downstream industries are located within the borders of Americas economic rivals.12A
124、 SEMICONDUCTOR STRATEGY FOR THE UNITED STATESFIGURE 2Summary of“beyond the chip”ecosystem required to make leading-edge logic semiconductors successfulISSUES TO ADDRESS AND PITFALLS TO AVOID IN GOVERNMENT POLICYBeyond industry complexity,other difficult issues to deal with include the following:Ther
125、e is ample private funding availablefor many industry endeavors,but it is notnecessarily aligned with a national strategy.The industry in the United States does not,ingeneral,have a funding problem.In fact,U.S.semiconductor companies currently havestrong industry fundamentals and high profitmargins.
126、U.S.-headquartered semiconductorcompanies invest heavily already,roughly$40 billion dollars in R&D and$35 billion inCapex per year.In the latest annual reportingperiod,the ten largest U.S.-headquarteredsemiconductor companies had over$110billion in cash and generated over$40billion in cash.23 There
127、is also a mature andefficient private equity,growth equity,andventure capital industry that can adequatelyfund any commercial efforts with market-level returns on investment(ROIs).It makes little sense for the government to invest in areas where there is alternative private financing.Yet,while priva
128、te financing optimizes outcomes for individual compa-nies,it is failing to close industry-level or country-level gaps that an individual company is unwilling to close by itself.For instance,since private capital companies can generally make higher ROIs by scaling U.S.-developed semiconductor innovat
129、ionsin other countries,they have been doing soover the last two decades.Publicly driven research efforts andconsortia have less and less impact overtime.Over the lifespan of the industry,governments and government-relatedindustry groups(for example,theInteruniversity Microelectronics Centre,Sematech
130、,and ITRS)have attempted manytimes to establish common industry roadmaps.However,these industry groupsefforts have become less relevant totechnology advancement than the indi-vidual efforts of major manufacturers whocoordinate their suppliers and customersaccording to a privately driven road map.FOR
131、EIGN POLICY AT BROOKINGS13The U.S.industry has big challengesbeyond capital.Even with equivalent inputcosts,U.S.-based manufacturing will likelyend up with higher finished product coststhan their Asian competitors.Too often,U.S.-based facilities lag behind Asian-based facilities in speed and efficie
132、ncy(inlaunch,in ramp to scale,and in continuingoperations)due to more stringent regu-latory regimes,Asian governments longexperience in“making manufacturingeasier,”and the outstanding manufacturingculture demonstrated by Asian competi-tors.24 The United States,at both the policyand company levels,ne
133、eds to accelerateevery part of the end-to-end manufacturingapproach and pursue innovations(e.g.,viaautomation)to overcome inherent disad-vantages to U.S.-based manufacturing.Thegovernment can help most readily in areassuch as permitting/regulatory approvalsand industrial site setup(e.g.,clearingland
134、,delivering gas and water lines,andconstructing access roads).These effortscan make a real difference in the return oninvestment to manufacturing investment.Studies show that over the last 20 years,ithas taken about 25%longer to build fabs inthe United States versus in South Korea or Japan.The Unite
135、d States is doing little,if anything,to increase the speed,while Asian countries are doing everything possible to accelerate it.In fact,the speed has been decreasing in the United States;the average U.S.-based semiconductor facility took 38%longer to construct in the years 2010-2020versus 1990-2000.
136、25 For leading-edge semi-conductors,a nine-to-12 months head startis of enormous financial and strategic value;a more slowly built U.S.facility is a muchless competitive U.S.facility.FIGURE 3Comparison of the time from construction start to production1,0009008007006005004003002001000Days
137、0-20102010-2020USAMainland ChinaTaiwanSource:World Fab Watch26Another problem is that the United States has a very talented but too small,aging workforce.Roughly doubling the U.S.semi-conductor manufacturing base and R&D capacity in the next decade would require tens or even hundreds of thousands of
138、 highly skilled,highly trained(and highly sought-after globally)engineering talents to join the industry.A foreign company looking at investing in the U.S.semiconductor industry today may doubt that the United States can provide such talent and would scale back their aspirations accordingly.14A SEMI
139、CONDUCTOR STRATEGY FOR THE UNITED STATESThe United States is competing with arising bar,as other countries executetheir own semiconductor strategies.Othercountries also aiming to scale their next-gen-eration semiconductor manufacturingcapacity are offering lower wages,few laborprotections,less strin
140、gent environmentalreview processes,and looser subsidy rules.Particularly notable,Chinese policies enticeboth American and third-party country semi-conductor suppliers to invest in an emergingChinese ecosystem.Many U.S.,European,South Korean,and Taiwanese companies(and American-trained engineers)will
141、continue to take up Chinas offers becauseof financial and market share reasons.U.S.policy may slow but will not stop this trend companies and engineers will not respondwell to messages such as“do not serve theChinese market”or“turn down a big paypackage to work in Shanghai.”U.S.policy willneed to ma
142、ke investing and working in theUnited States a more attractive proposition.This means accelerating innovation,makingit easier and cheaper to get things built inAmerica,and burnishing the United Statesreputation for transparency,IP protection,and rule of law.Aptly playing to Americasstrengths and cre
143、ating a favorable andcompetitive ecosystem at home are goingto be far more successful approaches thanblunt nationalist or protectionist policies.Previous U.S.industrial policies haverepeatedly failed to avoid common pitfalls.The United States needs to avoid commonmissteps that have plagued industria
144、l policy.These include the following:“Protect”over“promote”thinking protectionist and/or punitive policystances advanced by well-intentionedU.S.stakeholders without the sufficientinvolvement of U.S.industry championsand executives.Half-hearted investments limited scaleinvestments in an extremely cap
145、ital-in-tensive industry,with a focus on currentnot future technologies cannot lead totransformation impact.Blunt subsidies and industry handouts subsidy-led policies that lead to importduties on U.S.chip exports,which hurt,rather than help the industry.Subsidy-ledpolicies are not sustainable.Ineffe
146、ctive bureaucratic team structuresand decision-making processes thosethat do not reflect industry expertise anddo not move at the pace of the industry.FOREIGN POLICY AT BROOKINGS15A TOP 10 OBJECTIVES LIST TO DEFINE SUCCESSA new U.S.National Semiconductor Policy must have a tangible impact,so I propo
147、se these top 10 objectives for the nations industry push a“North Star”to drive concrete policies:1.The global and American semiconductor industries are financially healthy and continue to drive global innovation,with posi-tive annual revenue growth across industry cycles with stable or increasing R&
148、D intensity(R&D spending as a%of sales revenue).2.The United States maintains or improves its global semiconductor revenue share(47%today)and fabless revenue share(about 60%),while increasing its share of global semiconductor manufacturing from about 10%today to about 15%within 5-10 years,with a goa
149、l of reaching about 20%within 10-15 years.3.The top-five leading-edge semiconductor manufacturers(Intel,Samsung,TSMC,Hynix,and Micron)all have or are planning scale-level,best-in-class fabs in the United States,each with the capability to ramp leading-edge nodes as the alpha or beta ramp site.4.The
150、supply chain for these U.S.-based manufacturing facilities(including materials,chemicals,and equipment)is resilient,with capacity via domestic suppliers,foreign companies manufacturing in the United States,or U.S.company-owned assets in low-risk foreign locations.In addition,the overall global semic
151、onductor industry has no single point of failure with global leaders(e.g.,leading-edge foundry providers)possessing distributed networks to provide backup capacity even during unexpected,disruptive geopolitical or economic events.5.Each major,non-U.S.player in the global ecosystem(e.g.,TSMC,Infineon
152、,ASML,Nikon,Tokyo Electron,Merck,ARM,MediaTek,and Hynix)and their much broader set of ecosystem partners are focusing their road maps on the require-ments of the U.S.semiconductor market and prioritizing U.S.customers(i.e.,the United States is their“must-win”market).6.Across emerging semiconductor s
153、egments and the fundamental technologies(e.g.,new materials,neuromorphic computing)under-lying those segments,the United States has a clear leadership position in research,development,and commercialization.This means that American companies,universi-ties,and labs are spending more,developing more br
154、eakthroughs,and retaining more of the absolute best talent in all key semicon-ductor research areas(e.g.,the creation of new semiconductor elements,transistor structures,lithography,wide bandgap materials,silicon photonics,and quantum computing).7.U.S.companies(both in semiconductors and in downstre
155、am industries)lead the global industry in delivering the critical system-level technologies that define the control points for technology transitions in semiconductors(e.g.mobile,PC and cloud operating systems;machine learning frame-works;software developer tools,interface standards,communication st
156、andards,etc.).8.U.S.investors double the amount of invest-ment in emerging semiconductor compa-nies(e.g.,startups,spinouts,business model expansion from the system or soft-ware industries),with efforts spread across all key emerging segments.16A SEMICONDUCTOR STRATEGY FOR THE UNITED STATES9.The U.S.
157、industry doubles the annual numberof new Ph.D.hires and increases its U.S.jobbase by more than 25%while increasing theparticipation and representation of underrep-resented groups in the industry.10.The U.S.industry establishes at least threenew semiconductor research clusters thathave the scale and
158、expertise to thrive longterm.This can create a network effectthat goes well beyond the semiconductorindustry,reducing regional economic dispar-ities and developing economic multipliers.I recognize no set of objectives is perfect;industry experts could debate and refine these objectives ad infinitum,
159、leading eventually to a different list.The key idea is the Policy(1)has defined quantitative objectives to help measure progress,rather than ambiguous objectives such as“supply chain security”that have no concrete meaning;(2)is aggressive,because the United States should be playing to win;and(3)comp
160、rises a holistic set of objectives thatmakes the entire global industry stronger,ratherthan objectives focusing on a single topic likeleading-edge fabs.FOREIGN POLICY AT BROOKINGS17A DURABLE APPROACH TO FINANCING THE TOP 10 OBJECTIVESSo,how can the U.S.government achieve the best results?What type o
161、f government funding program would avoid waste,loss of capital,and negative geopolitical implications?A good approach would be to think about the ecosystem more broadly,focus resources on the big problems,encourage other“smart capital”to co-invest,be open and global,and operate with transparency.KEY
162、 INVESTMENT CRITERIA FOR ACHIEVING SUCCESSThe United States should develop clear invest-ment criteria that allocate public and private capital dollars to the initiatives that will have the most impact and that can organically compound success in perpetuity.These criteria include the following:Advanc
163、ing Americas interest in the semi-conductor industry(the top 10 objectives)while returning taxpayer money and makinga return on that money;Investing in areas that are intriguing toprivate capital and companies but for whichthere are limited lead investors;Taking a full ecosystem view by building ahe
164、althy,durable,and secure supply chainsupported by a robust software and systemarchitecture ecosystem;Supporting both American companiesand foreign companies critical to the U.S.ecosystem in a cohesive manner thatpromotes the unavoidable global intercon-nectedness of the industry;andGenerating co-inv
165、estment from third-partyinvestors to effectively cap the overall govern-ment investment required,de-risk theseefforts,and simultaneously harness Americancapitalistic ingenuity and government dollarsto gain the massive scale necessary.SUPPORTING BOTH AMERICAN AND FOREIGN COMPANIESAs discussed previou
166、sly,the goal of the proposed funding program is to leverage global industry investment to sustain American competitiveness in the semiconductor industry.In the current fraught geopolitical environment,the external positioning of this effort would be vital.The United States should avoid positioning t
167、his promotion effort as a competitive response to,or retaliation for,the industry-building efforts of other countries.The proposed U.S.effort is about building a strong American ecosystem that can provide inno-vation and growth for the technology industry around the world.If foreign companies meet t
168、he United States investment criteria,allow full due diligence,and operate with the required transparency,they should be allowed to partic-ipate no matter where their headquarters are located.American companies would have little basis to complain in taking this level-playing-field approach;since 2000
169、,U.S.companies have located 85%of their greenfield fab construction outside the United States,taking advantage of the incentives and favorable oper-ating environment in those locales.27 18A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESA BROAD SET OF OPPORTUNITIES ALL REQUIRING INVESTMENTLogically,wit
170、h a complex value chain,many types of investment are needed to ensure success.Each type carries a different level of return potential as well as different technical,execution,and other business risks.The Fund should look for the following opportunities:Strategic efficiency and capital stack invest-m
171、ents Level the playing field for Americanmanufacturers competing with overseascompanies through fundamentally lowercost capital and/or labor.Optimization and automation investments Invest in the companies and technologiesthat automate manufacturing and reducelabor intensity and cost disadvantages.Sc
172、ale-up and commercialization investments Accelerate the link between advanced,early-stage research and the companiesand investors that could commercialize theresulting breakthroughs.Business expansion or new market entryinvestments Support semiconductorcompanies to make disciplined and well-man-aged
173、 efforts to enter adjacent semiconductormarkets or downstream(e.g.software orsystem-level technologies)integration;andpartner with system companies as theyinvest to enter the semiconductor designmarket specific to their product offering(e.g.automobile manufacturers designing autono-mous driving logi
174、c semiconductors)Innovative foundational research Providefunding for truly advanced research effortsthat benefit the whole industry,that arebeyond the investment time frame of mostcompanies,or that one single individualcompany would not have an incentive toinvest in,as well as for long-shot but stra
175、te-gically critical projects to remove industrysingle point of failures.Government procurement acceleratorprograms Build concrete links(e.g.,seedcapital programs)between U.S.-government-driven infrastructure(e.g.,the spaceprogram)and emerging U.S.semiconductorinnovations and companies.Future industr
176、y workforce foundations Strengthen the semiconductor-focusedtechnical workforce(e.g.,physics Ph.D.s)by investing in research programs at U.S.universities,providing incentives for U.S.and foreign-born talent to attend theseprograms,and encouraging the graduatesto remain in the United States and remai
177、n inthe semiconductors field after graduation.THE INVESTMENT OPPORTUNITIES AND CORRESPONDING PROGRAMSI summarize here the various proposed invest-ment programs reviewed in this section.Generating co-investment from partners:Turning$52 billion into$200-250 billion-plusAmerica absolutely needs more th
178、an$52 billion to meet the proposed top 10 objectives.28 The Policy leaders need to figure out how to use third-party capital to sustainably grow the funding to$200-250 billion-plus,the amount necessary to make a truly significant and durable impact.The approach to partnering with third-party investo
179、rs must be flexible,depending on the type of investment to be made,the time frame,and the underlying business being financed.The approach needs to meet the return and risk requirements of sophisticated investors or the government fund will bear all the risk.Over the past decade,the U.S.government ha
180、s success-fully implemented and executed several finan-cial structures that can be utilized to achieve this goal.Co-investment by private investors and semi-conductor companies will help stretch the$52 billion(provided by American taxpayers)to FOREIGN POLICY AT BROOKINGS19approximately$200-250 billi
181、on,enabling the U.S.semiconductor industry to gain enough capital,at the right cost,to drive change.Third-party capital will also provide a second layer of governance and human resources to ensure that these investments are a success.Likewise,over time,as government incentives help bring top-tier in
182、vestment firms into the semiconductor sector,the firms increased knowledge and sophistication around such a complex industry and set of technologies should lead to a more robust set of third-party capital providers for decades to come.Again,extending the invest-ment multiplier well beyond the envisi
183、oned taxpayer capital makes this a self-sustaining,capitalistic solution that will build a more robust and thriving industry.Attracting different types of investors to maximize capitalThe U.S.financial system and the associated asset management industry are nearly as complex as the U.S.semiconductor
184、 industry.The key to raising the maximum amount of capital from third parties,at the right cost of capital,is to understand the different types of capital providers and their different return requirements and risk tolerances.The three main groups I propose focusing on are private equity,debt lenders
185、,and venture capital.(Of course,the Fund can also work 1:1 with oper-ating companies as co-investors,as there is no mandate that an investment must have a third-party source of capital.)Private equity$70 billion of additional capitalPrivate equity investors are logical partners for manufacturing cap
186、acity investment.These firms generally search for investment returns in the high teens,but with government support and the asset-heavy nature of the business,a suffi-cient number of firms might be willing to accept 14-16%gross returns to invest alongside industry.While this is below the targeted gro
187、ss returns for many private equity funds,the lower risk profile and predictable returns could justify the lower returns,especially as private equity firms can partner with industry-leading companies.The typical return on equity for a big semicon-ductor fabrication plant,a different financial metric,
188、is in the high single digits.To close the gap,the government fund could invest in debt FIGURE 4Summary of proposed investment programs and capital allocation targetsFoundational Research“Scale Up”and CommercializationStrategic Efficiency,Capital Stack,Optimizationand EfficiencyPartnersEnterprisesUni
189、versitiesNational LabsVenture CapitalEnterprisesPrivate EquityDebt LendersGovernment Fund ComponentNational Semiconductor Technology CenterPartnership for Revitalizing American Semiconductor ManufacturingCapital AllocationGovernment$39BPrivate Sector$9B$200-250B$2B$10-20BNational Semiconductor Ventu
190、re Capital Pool Government Provided CapitalSubsidies,Preferred Equity and Debt via STARP and TALFVenture Capital FundingSubsidies and Grants20A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESand/or preferred equity that is more senior in the capital structure;this would solve two big problems for priva
191、te equity investors and the government:taxpayer risk and the cost of capital.First,U.S.taxpayers will far less likely to lose money by making third-party capital subordinate to taxpayers,as is the case with investments by the U.S.Development Finance Corporation(DFC).Second,investing in debt and/or p
192、referred equity is lower risk and can involve a fair but lower cost of capital,enabling common equity invested by private equity funds and semiconductor companies to generate the higher returns demanded by their investors,who are mostly U.S.pensions,endowments,and high net worth individuals.This sim
193、ilar structure was used during the financial crisis to recapitalize banks through the Troubled Asset Relief Program(TARP).Not only was TARP successful in recapitalizing some of the most challenged parts of the U.S.banking system and in flowing credit back into the economy,but taxpayers also made mon
194、ey.If the United States could combine the best of what worked with TARP and the DFC,it could create an entity with a track record of success modeled on public-private collaboration.To this end,I propose a new Semiconductor Technology Asset Revival Program(STARP),which when blended into the capital s
195、tack of a particular investment,would enable$70 billion of private equity capital to be economic and invested alongside STARP,but critically,on a subordinated basis to U.S.taxpayers.Debt lenders$120 billion of additional capitalA significant driver of ROIs is the considerable time it takes to constr
196、uct a fab and the neces-sary capacity in the ecosystem to feed it.In addition,cost is a factor;the equipment for a fab plant is extremely expensive,especially leading-edge semiconductor equipment.The Fund could reduce the total capital cost of asset-heavy investments by leveraging a lending program
197、like the Federal Reserves Term Asset-Backed Securities Loan Facility(TALF),which was used in the administrations of both Barack Obama and Donald Trump to great success(TALF and TALF 2.0,respectively).29 In this case,instead of the Federal Reserve lending money to purchasers of certain asset-backed s
198、ecurities,it would lend money to any qualified fund that made loans or leases to companies building new semiconductor plants in the United States.This new TALF-like facility,or a Term Asset Lending Facility(TALF 3.0),borrowers of TALF 3.0 provide low-cost loans for property and equipment investments
199、.This program could attract massive amounts of low-cost capital,leveling the playing field for new fab construc-tion in the United States with highly subsidized fab construction overseas.This effort would require the Fund to provide$5 billion in capital to the U.S.Treasury.The Treasury would use thi
200、s capital to“equitize”the Federal Reserves TALF 3.0 13.3 Facility.With first lost capital from the Treasury and significant private equity capital in the form of subordinated,first loss common equity,TALF 3.0 could provide borrowers in the program 12.5-to-1 leverage,as was done for the many of the a
201、sset-backed security types in TALF 2.0(some asset-back securities categories were leveraged even higher than this).With 12.5x leverage,and a guarantee from Treasury,the Federal Reserve could lend short-to medi-um-term capital at a 1.5%interest rate to quali-fied TALF 3.0 borrowers to create a compel
202、ling opportunity to invest in these borrowers asset management products.The borrowers would raise additional capital,representing approxi-mately one-twelfth of the total loans and leases made.Investors of the TALF 3.0 borrowers would,after accounting for asset management fees,likely receive returns
203、in the low teens,with the semiconductor companies receiving medium-term financing at just 2.50%.As the fabs enter the construction phase,the TALF 3.0 equipment loans could be rolled into new,senior-term debt loans at the senior most part of the capital stack,matching the semicon-ductor companies and
204、 private equitys common equity investments into these massive projects.The initial cash flow,after accounting for TALF 3.0 debt service and STARP returns,would go to pay back STARP.Depending on how high the returns of the new fab were,this payback could be achieved in seven to nine years.However,it
205、is likely that the capital structure could be FOREIGN POLICY AT BROOKINGS21refinanced and STARP could be paid back from this refinancing much sooner than this time frame.Altogether,the Funds$5 billion to Treasury could attract more than$10 billion in first lost investor debt capital that the Federal
206、 Reserve provides$110 billion of loans against through TALF 3.0 thereby amounting to an additional$120 billion at the top of the capital structure.Obviously,more capital could be attracted if Treasury and the Federal Reserve were to deem a higher level of leverage appro-priate,which they have in the
207、 past for mort-gage-backed securities,for instance.The Fund,STARP,and TALF enabling a partnership for revitalizing American semiconductor manufacturingThe Fund,STARP,and TALF programs would be synergistic and thus need a common gover-nance umbrella to ensure the best fit for invest-ment opportunitie
208、s.With increased government and third-party support,the United States would need a structure that validates investors and enterprises but,at the same time,allows the market to set valuation,structuring terms,and perform investment due diligence.To do so,the government would set the terms for receivi
209、ng co-investment,based on a sliding scale of government support in relation to private sector support and on the amount of third-party private equity capital provided as a percentage of the total common equity pool for a project.The sliding scale would enable more government funding and greater leve
210、rage of TALF 3.0,as third-party capital increased.This package would be collectively known as the Partnership for Revitalizing American Semiconductor Manufacturing(hereafter referred to as the Partnership).This Partnership would be the fundamental mechanism for managing the Funds private equity and
211、manu-facturing capacity investments.Semiconductor companies and third-party investors would be welcome to bring any investment proposal to the Partnership.The proposal would be approved if it supports the semiconductor strategy goals,meets the governance terms,and has committed investor capital.When
212、ever a semiconductor company and private equity investor start to draw on the approved,initial government capital(i.e.,invested in the project or joint venture),they would be legally bound to invest their proposed amount of capital,according to the Partnership.Through the Partnership and the Fund,th
213、e U.S.government would become a minority equity investor at either the project or corporate level.After making the investment,the governments sole role would be to protect it as is the role of any investor in any venture by ensuring that the project meets the investment commitments made during the p
214、roposal negotiations.This financially driven governance role is preferable to a more nebulous and idiosyncratic governance role the U.S.government would have to play as a provider of“strings attached”grants or subsidies.Research grants and venture capital$10-20 billion of additional capitalMajor ind
215、ustry players make limited invest-ments in advanced research because the payoff time frame is so far in the future.Venture capitalists(VCs)have shifted their investments to asset-light/digital business models that can scale up and provide much higher returns much faster.In addition,with the cost to
216、design just one semiconductor on a leading-edge process semiconductor exceeding$100 million,the cost of entry for new design companies is simply too high and uncertain for most VCs.As a result,semiconductors today account for less than 5%of U.S.investment in venture capital(as opposed to more than 2
217、0%of Chinese invest-ment in venture capital in 2021).30This is a problem.Startup companies provide benefits that established giants cannot:dyna-mism,technology experimentation,and a pathway for new talent into the industry.The United States needs two separate approaches for stimulating venture capit
218、al.The first would be very early-stage investments and the second would be later-stage commercializa-tion investments.Very early-stage/“lab to world”tech transfer In this area,the government wouldwant a broad set of investors looking atthousands or tens of thousands of ideas.But the government would
219、 not be able to22A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESmicromanage individual research invest-ments;it would need to create a matching grant program that engages a wide-ranging set of organizations,such as SBICs,to administer a simple grant application in return for a fee.Applicants would ne
220、ed to work for an accredited U.S.higher education institution,as is often the case with National Institutes of Health grants.As the potential for loss is higher,this would be a relatively small pool of investment.Later-stage scale-up or commercialization investments I propose a government-led Nation
221、al Semiconductor Venture Capital Pool in which the Fund and private venture firms would be limited partner investors.Government-matching funds would have preference in the capital structure over third-party capital investors.In return,third-party investors would receive nearly all the upside(above a
222、nd beyond a preferred return of about 5%to the Fund),as opposed to the standard venture capital 80/20 split.This setup would induce VCs to deploy much larger portions of their third-party capital funds in the semiconductor industry,as the VCs would gain additional upside poten-tial from the governme
223、nt-invested dollars,thereby lowering the cost of capital in these riskier projects.The government investment team would have a simple veto right up front before a VC spends significant time,money,and effort but only to make sure that the potential investment meets the broad strategic parameters of t
224、he National Semiconductor Policy.VCs would then use their standard investment toolkit and gover-nance approach to build winning companies and take them public,without government interference.This approach enhances the scale and scope of VC funding flowing into the industry and does not place the Fun
225、d in the role of picking winners.Government Procurement Accelerator Programs(GPAP)The U.S.federal and state governments purchase billions of dollars of semiconductors(directly or indirectly)annually.While many programs are in place to promote the purchase of U.S.-sourced technologies,the complexity
226、and fragmentation of these programs makes it difficult for all but the largest companies to identify and pursue opportunities.The United States needs to simplify the interface and process to expand the links between industry and government.All companies and programs supported by the Fund should be e
227、ligible for inclusion across all government procurement efforts related to semiconductors.The United States could also design a“matchmaking”mechanism that communicates all procurement oppor-tunities to registered participants.Similar to the philosophy of the rest of the program,this matchmaking supp
228、ort would open to U.S.-headquartered companies and global companies that are making qualifying investments in the United States.In addition to increasing the financing available for R&D activities,the United States should use tax credits to encourage more R&D investment.These credits should be exten
229、ded on an even playing field to both U.S.and global companies investing in the United States.Funding pure foundational research The last investment focus of the Fund would be foundational research.In many ways,this is more important than the manufacturing investments.A semiconductor fab is the physi
230、cal instantiation of R&D investments that were initiated up to a decade earlier.The United States cannot lead in manufacturing if it does not lead in early-stage research.The semiconductor industry(and all the industries that depend on semiconductor innovation)cannot grow if it cannot continuously i
231、dentify,improve and industrialize new materials,new transistor architectures and manufacturing techniques that increase the performance and lower the cost of computing.As current mate-rials and manufacturing techniques are hitting physical limits,we cannot assume that the industry will simply“figure
232、 it out”.There is an immense amount of basic research that must be completed much of which is not underway in the United States today.In prioritizing pure research topics,the focus should be on end-to-end ecosystem leadership across technology transitions.In other words,FOREIGN POLICY AT BROOKINGS23
233、the research themes should comprehend both“winning in transitions”and building technology leadership across system-level technologies:investing in the full spectrum of standards,system architecture,processor architecture,operating system and applica-tions that enable future technologies to be launch
234、ed.For example,the future move to quantum computing will require fundamental breakthroughs across transistor architecture,materials,design,manufacturing,supply chain;as well as entirely new approaches to software and system design.The research agenda needs to comprehend all of these components.It wi
235、ll be the holistic strength of the U.S.end-to-end ecosystem that ensures Americas future leader-ship in semiconductors,not just the strength of the U.S.manufacturing base.Therefore,focusing research investment on an R&D pilot line for leading-edge semiconductor manufac-turing is useful,but too narro
236、w to be the sole focus of the research effort.The government team,in conjunction with the Industry Advisory Council created by the CHIPS Act,needs to develop a comprehensive R&D approach,one that allocates funding at scale across a series of R&D themes required to meet the top 10 objectives,and gene
237、rates co-in-vestment from both private and public sectors sources of research capability.A National Semiconductor Research Center would be the centerpiece of this advanced research effort.The research center would operate as(1)a steering committee to drive nationwide research programs;(2)a connector
238、 across government,research,and industry;(3)a designer of programs to encourage talented students to pursue advanced degrees in the disciplines critical for semiconductors;(4)an evaluator of Americas competitiveness across different technology areas,and(5)the direct driver of research execution effo
239、rts on“moon-shot”technology breakthroughs or government priority projects that would be untenable for private industry to finance.The Fraunhofer-Gesellschaft research organization in Germany is a compelling example to explore and potentially emulate as a model.This group comprises 76 affiliated inst
240、itutes across Germany,employs nearly 30,000 researchers and is the biggest applied research organization in Europe.31 The Institutes ability to self-finance 70%of its expen-ditures via contract work,royalties and other means demonstrates that government funding can be multiplied(in this case by 2.5x
241、)through an effective industry engagement model.In addition,the United States could accelerate the connection between the government and emerging technology providers via the GPAP program.This would allow the U.S.government to become a far more common alpha customer of new technologies,thereby de-ri
242、sking the efforts of third-party investors.By sharing startup risk,the United States can vastly accel-erate the flow of capital and innovation in this critical research area of the industry.24A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESACCELERATING DEPLOYMENT AND ACCUMULATING TALENTFUTURE INDUSTRY
243、 WORKFORCE FOUNDATIONS:RECRUITING,TRAINING,AND TARGETED IMMIGRATIONThe U.S.semiconductor workforce is strong and world-class;but as one would expect,it is concentrated in the areas where the United States has greater share,such as fabless design.However,for the country to achieve the proposed top 10
244、 objectives over the next decade,it would need to launch 10 or more manufacturing facilities(in the leading-edge,lagging-edge,analog and power,memory segment,and front-end and assembly/testing areas).With a simplistic assumption that each facility would require 5,000 highly skilled employees,such an
245、 expansion in facilities would require 50,000 new workers.32 And this is just workers for fabs to be built the industry needs new researchers,architects,designers,validation,verification,and software engineers.The demand for new personnel could reach more than 100,000 highly trained employees in the
246、 next five years a growth of 40%.33 The need for new talent is increased by an aging workforce.Roughly 60%of semiconductor engineers in the United States are over 40 years old,representing a much higher proportion than in the software industry.34 With every major country in the world now aiming to b
247、uild up its semiconductor industry,the United States is in a talent war.U.S.companies are,of course,working hard to recruit and train talent.However,their efforts are not enough.While hiring for commercial activities is best performed by labs,universities,and companies,the govern-ment should do more
248、 to increase the size of the pool and remove roadblocks to accessing it.Forty percent of the semiconductor workforce in the United States is foreign-born;the country is great at attracting foreign talent.35 But with the war for global talent accelerating,with everyone chasing the same limited supply
249、,the U.S.needs to up its game.America needs to incentivize large-scale immigration of the top South Korean,Taiwanese,and Japanese semiconductor workers(and,of course,talent from other leading semiconductor countries).These workers are best positioned to accelerate and de-risk the investment of South
250、 Korean,Taiwanese,and Japanese companies in the United States.Second,these workers run the best fabs and OSAT facilities in the world;the experience and tacit knowledge they could bring to new U.S.efforts(and could impart to U.S.workers)would be invaluable.Another type of talent will be urgently nee
251、ded as multiple new fabs are built in the United States:specialized construction workers with specific talents required for high-end manufacturing facilities.Typically,up to 6,000 total workers are required for each greenfield fab construction.36 The“cleanroom”in a fab is the most advanced building
252、in the world and therefore requires specialized and experienced construction capa-bilities,including translating 3D models to phys-ical spaces,making the worlds tightest welds,designing lethal gas protection systems,and inspecting airflow to sub-parts-per-million accu-racy(and hundreds more).These s
253、kill sets are in short supply in the United States and will be in increasing demand.The gaps offer opportuni-ties to generate high-quality blue-collar jobs that can later become advanced construction jobs(e.g.,to design and build facilities for lithium batteries and hydrogen capture systems).FOREIGN
254、 POLICY AT BROOKINGS25While the industry itself is making efforts to close these gaps,the National Semiconductor Policy can and should partner with the industry to accelerate the efforts.The United States needs to adopt a holistic set of bold initiatives to attract and grow more talent.These could i
255、nclude the following:Employee training and development funds asside-by-side Fund investments The UnitedStates could ensure that each Fund invest-ment includes a small amount of money,matched by financial and corporate inves-tors,for training and recruitment efforts.University research hiring program
256、s TheUnited States could ensure that universitygrant programs funded by the proposedNational Semiconductor Research Center haveexplicit targets and money for bringing Ph.D.and postdoctoral talent into the program.Targeted immigration relief programs TheUnited States could remove the hard capson H1-B
257、 visas for workers seeking to joinsemiconductor manufacturing and researchprograms.In particular,to attract talent fromleading semiconductor manufacturing coun-tries in Asia,the U.S.government could setup country-specific programs,leveraging thetemplate set by the 10 existing country-spe-cific progr
258、ams setup for other purposes.37Targeted attraction of global semiconductortalent The U.S.government,in collabo-ration with research universities and topcompanies,could identify the top severalhundred to a thousand research personnelglobally;co-develop incentive programs;andprovide an expedited,low b
259、ureaucracy pathfor these personnel to take on academic orcorporate research roles in the United States.ACCELERATING THE INDUSTRY:STREAMLINING PERMITTING AND REGULATORY PROCESSESAs mentioned,the United States is one of the most expensive and slowest countries in which to build manufacturing facilitie
260、s issues that constrain investment in both leading-edge semiconductor fabrication facilities and the many smaller facilities that make the tools,gases,chemicals,and construction mate-rials needed for giant cleanrooms.There are numerous underlying causes of these issues,and many of them bring great b
261、enefits(such as citizens demand for clean air and water and the respect for private property rights)or are products of the United States federalist system.It is too far a stretch to propose refashioning the national regulatory system in the pursuit of greater success for the semiconductor industry.I
262、t is unlikely the U.S.system could ever match the decision-making efficiency of a monolithic city-state or a single-party political system.However,the United States should not accept the status quo;the Policy team must direct a set of partners to reduce the time-to-market gap from 40%-plus to less t
263、han 20%.I propose a set of common-sense,feasible steps to accelerate the process:Create a central mechanism to coordinateand streamline regulatory engagement bysemiconductor companies.Staff an empow-ered organization to be the single windowthrough which semiconductor companyinvestors can understand,
264、review,discuss,and monitor permitting and regulatoryprocesses.Similar,successful setups canbe found in other countries and regions.38Delete duplicative reviews and procedures.Create a mechanism to review federalpermitting procedures for technologymanufacturing facilities,with two purposesin mind:rem
265、oving those procedures that areoutdated or no longer needed and removingthose are duplicative of state requirements.Build best practices and a problem-solvingculture.Staff a national expert team todevelop a sustainable process to sharebest practices with industry players andstate and local officials
266、 and to collaboratetogether to remove bottlenecks that slowdown launches.Prioritize speed in the investment process.Incorporate into team training,investmentevaluation,and decision-making the under-standing that speed is a core competitive26A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESfactor and th
267、at locations and companies that do not have best-in-class speed are less attractive investing targets.Benchmark the best.Evaluate the“speedysemiconductor starts”of global leadingcountries such as Japan and Singapore,identify their best practices,and bring thosepractices back to the United States.One
268、reason to create a stand-alone team for theFund is so that the Fund can replicate globalbest practices in a way that establishedgovernment agencies cannot.FOREIGN POLICY AT BROOKINGS27WIN-WINS FOR THE UNITED STATES AND ITS GLOBAL PARTNERSGovernments around the world understand that the semiconductor
269、 industry is the underpinning platform for innovation-led economies.These governments are rolling out policies either to promote investment in semiconductors,to protect their industry from overseas compet-itors,or to stop their national champions from investing outside their home market.For instance
270、,in December 2020,19 European Union member states signed the European Initiative on Processors and Semiconductor Technologies in a bid to reestablish the EU as a global power-house in the semiconductor industry.39 The joint declaration aims to enhance cooperation and increase investment in equipment
271、 and materials,design,and advanced manufacturing and packaging and includes a total pledge of 145 billion euro in support(of which little is new funding and therefore readily available for European companies).In February 2022,the European Commission announced the European Chips Act,which includes a
272、target to more than double Europes share of global manufacturing output to 20%.40 In May 2021,South Korea announced a 50%reduction in tax rates for semiconductor companies and that,together,South Korean companies and the government will invest more than$400 billion in the industry over the next 10 y
273、ears.41 In December 2021,India announced a new policy to build up the semiconductor industry in the country,with a goal becoming a global hub for lagging-edge manufacturing facilities.42U.S.policy will need to be successful within this context.But success does not depend on these other country effor
274、ts failing.In fact,the growth and success of semiconductor industries in similarly aligned countries could be a great boon to the U.S.industry;diversified geographic footprints increase supply chain resilience,provide market opportunities for U.S.firms,and supply more“smart brains”to solve the harde
275、st industry-related problems.The endgame needs to be a stronger and more resilient order in the global semiconductor ecosystem.First,success means that the globally leading semiconductor companies,such as those headquartered in Taiwan,South Korea,and Japan,are manufacturing at scale in the United St
276、ates,are performing R&D in the United States,and consider the United States a second home.Non-American partners are essential to achieving the top 10 objectives,as many are important suppliers,customers,or intellectual property owners.These non-American compa-nies will need tacit or explicit approva
277、l from their home governments to invest in the United States.Therefore,a common understanding and agreement with governments in key semi-conductor countries will be required.Second,success means engagement with Europe on technology has been renewed,whereby the American and European indus-tries espec
278、ially in the areas of automotive,industrial,medical,and telecommunications semiconductors consider themselves part of the same ecosystem that leverages common technologies,manufacturing capacity,and business rules.Renewing this engagement is of particular importance to the European tech-nology indus
279、try,and specifically Germany.As noted later in this paper,semiconductors are the foundation of Germanys most important 28A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESindustries,and that foundation has been progressively weakening.Partnering to turn that around offers the United States a rare openin
280、g to renew the critical US-German relationship.Third,success means careful diplomacy with Asian nations to define the“win-wins”and global industry rules with the Asian nations that lead in many semiconductor segments and now represent more than 70%of global installed manufacturing capacity.43 The Un
281、ited States needs to identify ways in which Asian countries,companies,and capital benefit from,rather than bear costs from,U.S.policy.There is only one path to accomplish this win-win scenario:expanding the pie by creating new technologies,markets and customers,so that all countries can increase emp
282、loyment and deploy more capital.Innovation through R&D is a far more compelling centerpiece of Americas foreign policy regarding semiconductors than supply chain realignment.The former creates opportu-nities for everyone to grow and benefit,while the latter drives countries to defend their turf and
283、see the policies of the United States as a threat.ENGAGING FOREIGN COMPANIES TO INVEST IN THE UNITED STATESFor the United States to achieve the top 10 objectives,non-American leading companies need to invest in U.S.-based manufacturing capacity and U.S.-based advanced R&D activi-ties.In fact,the obj
284、ectives cannot be achieved without the substantial contribution of many of the most important non-American semicon-ductor companies.In the long term,the United States needs to have a simple but aggressive goal:for leading foreign companies to consider the United States a second home,a base from whic
285、h they can serve U.S.customers and scale their business globally.Leading companies are constantly making investment decisions to boost production and enhance their capabilities and innova-tion leadership.In all cases,investment and research decisions made years ago are still being implemented today.
286、The challenge for the United States will be to convince global leaders to invest outside their home market,which is usually where they acquire the technical talent,engineering talent,suppliers,and services needed to build their business.Fitting into the ongoing long-term capital expenditure programs
287、 of these global leaders will be no easy feat,even for companies willing to work closely with the United States.Key plan imperatives include the following:Encouraging and incentivizing global leadersin leading-edge logic,advanced packaging,memory,and other industry segments tobuild and invest in the
288、 United States and tobring along their partners.Besides offeringfinancial incentives(e.g.,low-cost capitaland tax incentives),the United States mustunequivocally demonstrate a long-termcommitment,provide access to othersuppliers and their IP,and espouse certaintyto ensure a proper return for the lea
289、derstime,money,and effort.The effort willinvolve complex and difficult decisions by allparties.Compelling global companies to build R&Dcenters in the United States,or to add staffto their existing centers.Again,the resultingIP and know-how must be allowed to crossborders freely,or they will not inve
290、st.Global leaders will decide on manufacturing and R&D locations from a business standpoint;first and foremost,investing in America must make business sense.For manufacturing,they will need inducements similar to those required by U.S.companies:adequate returns on capital invested,the meeting of cus
291、tomer needs,and the avoidance of stranded capital.They will want to be part of a flourishing industry cluster that can both reduce risk and attract talent for their opera-tions.Therefore,they will follow the lead of big players in their industry segment.This is why it is so important for the United
292、States to first gain the commitment of a few strategic partners,in each corresponding segment,to invest in capacity in America.FOREIGN POLICY AT BROOKINGS29R&D investment:A more challenging taskThe strategy must incentivize global R&D investment into the United States with as much vigor as it does f
293、actory investment.Without a constant stream of new process technologies based on continuous R&D breakthroughs,a leading edge fab will soon fall behind and become economically unviable.Promoting a global shift in R&D will be more difficult(or at least less straightforward)than in manufacturing.Foreig
294、n companies will find it hard to separate out R&D functions and place them in different locations;they tend to cluster their core R&D functions at headquarters and around the“alpha fab”where new process technologies are rolled out.Companies make R&D decisions based on these factors:Strategic conside
295、rations;Proximity to the market and customers;Talent,at both the senior and junior levels;Associated costs and overall financialreturns;andPolitical,regulatory,and operational risks.In general,U.S.-based R&D will score well on strategic considerations and market proximity.However,global companies wi
296、ll have substan-tial concerns about access to talent,the level of returns,and the comparatively higher regulatory and political risks.Some major concerns and issues to be addressed include the following:Organizational costs While the UnitedStates has clear leadership and significantsenior research t
297、alent,it,unlike other coun-tries,has a smaller and more expensive poolof junior engineers.For instance,China isgraduating more than 10 times the numberof bachelors-level electrical engineers thanthe United States.In addition,the averageengineering salary in the city of Hsinchu,Taiwan site of the wor
298、lds largest semi-conductor cluster is only about$30,000,versus an average of roughly$75,000in Hillsboro,Oregon,a large U.S.-basedmanufacturing cluster.Add in insurance,other benefits,and taxes and the U.S.labor costs are roughly four times higher.44Other financial return drivers Beyond laborcosts,th
299、ere are many other issues thataffect an R&D centers financial returns.These include the centers ability to receiveU.S.government grants to help supportdirect R&D expenses,its access to equiva-lent R&D tax credits as U.S.-headquarteredcompanies,and the existence of a largeenough U.S.-based ecosystem
300、to warrant aseparate well-funded R&D operation.Open access to major U.S.customers The government should not push U.S.-headquartered companies or governmentagencies to“buy American,”as this will mini-mize a major reason to invest in the UnitedStates proximity to the worlds largestcustomer base.Free g
301、lobal use Non-U.S.companies will beconcerned that if they develop IP in the UnitedStates,they might not be allowed to export itto their overseas affiliates or will be subjectto export controls.The United States needsto establish substantial coordination betweenthe National Semiconductor Policy team
302、anda separate,independent government teamthat sets related export control policies.Open access for home-country capital providers Global leading semiconductor compa-nies have home-based capital providersthat will want to access investment in newU.S.-based facilities;non-U.S.investors needcomparable
303、access to the U.S.Fund program.ENGAGING FOREIGN GOVERNMENTS:TAILORED APPROACHES TO HOME MARKET ISSUESThe home countries of leading semiconductor companies face broader issues,including a more complicated calculus beyond near-term financial returns(e.g.,national security,foreign policy,domestic polit
304、ics,and industrial relations).Their semiconductor strategy is one piece of a 30A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESSegmentsRelative capabilities indexed to Market LeadersIndustry Segment CapabilitiesLeading-Edge LogicMemoryAnalog and PowerOptoelectronicsand PhotonicsValue Chain Capabilitie
305、sProcess technologyFoundry front-end manufacturing capacityEquipment and ToolsElectronic Design Tools(0%)(100%)USEuropeNon-PRC Asia Mainland ChinaFIGURE 5High-level comparison of capabilities by region45 much larger plan to support multistakeholder national interests.A U.S.policy approach must take
306、into account these specific hot button issues,which will vary in each country.As part of this broader engagement,the United States must make building in the country an easy decision.The following considerations are important:Foreign government buy-in Winning foreigngovernment buy-in will require hig
307、hlightingthe economic rationale for the companiesto produce in the United States,as well asdemonstrating how U.S.semiconductorpolicy is not meant to choke off local produc-tion and negatively impact that countrysindustry employment.Just like the UnitedStates does,foreign countries benefit whentheir
308、companies enter new markets and havediverse global supply chain footprints.Local regulatory changes In certain cases,it will make sense to request foreign govern-ments to change laws to enable local compa-nies to invest in the United States(e.g.,exportcontrols or technology transfer restrictions).Ap
309、propriate countermeasure deterrents TheUnited States must avoid bidding wars(e.g.,country X promises its home players“wewill match or beat U.S.economics”for semi-conductor investments)not only by building friendly,collaborative relationships with semiconductor authorities in other countries,but also
310、 by emphasizing that transparency in government support is essential and that the United States will use proper channels,such as the World Trade Organization,to stop or fully offset new subsidies created in a race to the bottom.While there are no tools that can fully stop“secret”or excessive subsidi
311、es,global cooperation can deter the worst excesses.DIFFERENT COUNTRIES NEED DIFFERENT ENGAGEMENT STRATEGIESWhile the United States has enough internally generated demand to provide a baseline scale for nearly any investment in semiconductors,other leading countries generally do not.Therefore,these c
312、ountries specialize in one or more industry segments or layers of the value chain.The table below,which by necessity is simplified and high level,lays out the compara-tive capabilities of major regions.FOREIGN POLICY AT BROOKINGS31There are several simple takeaways from this chart:(1)Asian countries
313、 and regions grade tops in numerous categories;(2)no country has the ability to be self-sufficient,as they all rely on a global value chain;(3)different countries have different problems to solve and therefore will have different engagement models vis-vis the United States;and(4)other countries will
314、 be more concerned with protecting their own vulnerabilities than with supporting the United States domestic efforts.Engaging an important partner:A deeper dive into the German situationThe automotive and industrial automation industries are Germanys economic mainstays.46 In total,advanced manufactu
315、ring drives more than 25%of German GDP and more than 75%of its exports.Leading-edge semiconductors,new materials,and software are critical inputs into these German industries.To protect its world-leading position,the German automotive industry needs access to new technology innovation,stability in s
316、upplies,and ecosystem support to put together the full solution for leading electric vehicles(e.g.,batteries,battery management systems,sensor systems,power electronics,CPUs for automotive control and autonomous driving,and underlying software stacks).Roughly 75%of the EUs semiconductor firms are he
317、adquartered in Germany.47 The countrys companies have strengths in certain analog,power,and sensing technologies and select equipment markets,but Germany on the whole lacks the semiconductor manufacturing capacity(e.g.,in leading-edge logic processing,software capability,and materials)to be a leader
318、 in the future automotive and“industrial internet”industries.As a result,German auto-motive and industrial automation industries rely on foreign suppliers and do not control their own future.In addition,Germanys current semiconductor strengths are being eroded by technology transitions.For instance,
319、the coun-trys traditional,silicon-based,power inverter semiconductors for automotive applications are now less relevant than emerging materials for future electric vehicle powertrains,such as silicon carbide and gallium nitride.This is a significant and growing issue for German car companies and the
320、 government.American companies have strong market share,IP position,and manufacturing know-how in many of the key technology areas where Germany has gaps.They see opportunities in the German and European markets.This is why Intel and GlobalFoundries,major U.S.semicon-ductor companies,have already an
321、nounced plans to invest in Germany.48The respective national funds of both countries could jointly facilitate the build-out of the supply chain in both the United States and Germany to leverage economies of scale and foster a more resilient global supply chain.The United States could extend the fron
322、tiers of a National Semiconductor Research Center umbrella to include Germany so that the countrys research teams can join either broad or individual,specific U.S.efforts.The framework could encourage U.S.and German companies to work together on key semiconductor solutions for the electric vehicle,a
323、utonomous driving,and indus-trial Internet of Things markets.U.S.semicon-ductor companies would get new customers,new insights into customer needs,and more access to German technology,while German companies(or German subsidiaries of U.S.companies)would gain experience in designing and building out a
324、dvanced semiconductor manufacturing and more access to better tech-nologies and U.S.markets and customers.Of course,Germany would need to step up its efforts for this collaboration to work.They would need to commit real capital,invest in partnerships with the Fund,remove roadblocks to allow German c
325、ompanies to invest in the United States,establish common rules on global access to critical semiconductor tech-nologies and supplies,and give U.S.semicon-ductor companies open access to the German customer base.32A SEMICONDUCTOR STRATEGY FOR THE UNITED STATESA BIGGER OPPORTUNITY:SEMICONDUCTORS AS A
326、FOUNDATION FOR A TECH PARTNERSHIP WITH EUROPEThe engagement with Germany could be both an example of and the foundation for a renewed technology partnership with Europe.A win-win partnership would help align the European tech-nology and business communities closer with the corresponding American com
327、munities.The United States could provide more market access,proximity to customers and capital,and joint investment in future technology break-throughs.Europe holds less than 10%of global installed semiconductor capacity,less than 4%of annual capital expenditures,and less than 5%of global annual ven
328、ture capital investment into semiconductors.And its major industries(e.g.,auto,industrial,telecoms,and medical devices)are highly dependent on the innovations that semiconductors bring.In return for U.S.support,Europe could provide joint funding and help define a competitive,global,transparent and r
329、obust semiconductor ecosystem.A more robust European semiconductor manu-facturing cluster would not only benefit the U.S.semiconductor industry,but also boost the busi-ness case for Asian companies to build their capacity to serve the European market which will be necessary to meet the top 10 object
330、ives for global supply chain resiliency.This type of global partnership does entail tradeoffs;American companies supported by the Fund would build some of their global manu-facturing capacity in Europe rather than solely in the United States.However,these tradeoffs are inevitable if the United State
331、s wants global support for its own manufacturing revitalization efforts.In fact,the United States should see the investment by other countries as a way to achieve the Policys top 10 objectives,because only the United States has the ability and foun-dational elements to lift all boats in the global s
332、emiconductor industry,if policy is coordinated and messaged correctly.Formalizing the global approach:Launching a global semiconductor partnership frameworkTo win the hearts and open the wallets of the global semiconductor community,the U.S.National Semiconductor Policy needs to be open,global,long-
333、term,committed,and patient.Just like leading U.S.companies,top foreign companies invest many years,across many cycles,to develop true breakthroughs.The Dutch company ASML spent 17 years developing its extreme ultraviolet technology,while the UK company ARM spent more than six years building its first 64-bit micropro-cessor.49 To fully convince the executives of global semiconductor companies and t