INNOVATIVE PARTNERSHIPS: PEOPLE, CULTURE, AND ACCEPTING RISK

Kim Budil (Lawrence Livermore National Laboratory) provided keynote remarks on the value of creating and sustaining partnerships. She opened, “We are entering an era where the research ecosystem is changing in very fundamental ways. To make the progress in science and technology at the pace required demands a change in the nature of collaboration and partnership.” She explained that Lawrence Livermore, as a government-funded entity, can take a long-term, persistent approach to science and technology, using a team science model to take on complex problems. “The 17 DOE [Department of Energy] labs are natural integrators and good centers of gravity to bring entities together,” she said. Big, “audacious” ideas are rarely solved with one scientist working in a lab alone—hence the need for collaboration.

She noted that the U.S. research enterprise is the largest in the world, but China is close behind. “Our longstanding position as the preeminent provider of science and technology in the world has changed. Countries have learned from us and built research ecosystems that look very much like ours. We need to be cognizant of that, and not lose the instinct to lead and innovate, and to protect our national security,” she commented. Budil said all sectors—government, industry, academia and, more recently, international and philanthropic entities—have roles to play, based on their different time horizons and resources. She encouraged thinking about the importance of the public sector component of cross-sector partnerships as the element that enables support for research long enough to bring innovative solutions to bear—the patient investor of the ecosystem. Lawrence Livermore and other DOE labs also support the infrastructure, tools, and capabilities that other stakeholders can draw on, which can contribute to innovative partnerships.

Budil shared examples of national and international cross-sector partnerships Lawrence Livermore catalyzed or contributed to, noting that the examples have not been without challenges. The first she discussed was a public-private partnership to advance drug development with high performance computing—later called the ATOM Research Alliance, discussed in greater detail later in the workshop. Budil attributed the successes of the partnership to, “bringing together the right group of people—the experts in physical sciences, computing, biology, clinical researchers, and big pharma—to bring all the right ideas to the table at the same time.” As another example, modernization of the manufacturing enterprise associated with the DOE National Nuclear Security Administration has required working with universities and industry partners to rethink materials and manufacturing. Budil described new partnerships in this area have included cooperative research and development agreements with industrial partners to strengthen the connections between design, design optimization, engineering, and production, to address the “huge gap” between new technologies and new technology production. Internationally, the Extreme Light Infrastructure Project, funded by the European Union, built facilities in the Czech Republic, Hungary, and Romania. Budil noted the goals were not only to advance science, but also to foster social cohesion by drawing researchers and new ideas into deliberately selected locations. “This was a way for the European Union to use a novel S&T partnership to go from a brain drain to brain circulation,” she said.

In closing, Budil stressed the need to focus on building relationships, and not just on contracting mechanisms. More broadly, she commented that labeling experimental projects as ‘pilots' often provides more leeway and flexibility for trying novel partnership arrangements without raising too many institutional red flags. She also cautioned against waiting for a perfect solution, and urged, “Take the 80 percent solution on the road. Until you have road-tested it a little bit, you won't know its flaws. Try things, find the flaws, make things better.” Last, Budil noted that multiparty agreements are important where everyone is invested in success. While it is important to think through value systems, cost models, licensing and intellectual property, “People are the most important currency we have,” she concluded.

In discussion with workshop participants, Budil stressed that as taxpayer-funded facilities, the national labs are oriented toward public benefits and making their capabilities available. She explained the royalty streams they derive are used to seed new technology and buy specialized equipment. Allowing individual inventors to receive royalties is important so they “invent and stay.” In response to a question from a workshop participant on the role and importance of performance management within large institutions related to encouraging and producing partnerships, Budil noted Lawrence Livermore is shifting how it conducts performance management from an annual system of detailed documentation to an employee-centric system with frequent feedback. She also pointed to an unprecedented demographic transition at the lab: nearly 50 percent of the staff has been there less than 5 years. These ongoing transformations will likely impact the culture of the lab, and the ways its employees approach innovative partnerships and collaboration. Addressing another participant question about the labs' user facilities, Budil noted the lab can partner, support, or enable other efforts, depending on the need. She acknowledged that despite the stated vision to innovate and accept failure, risk aversion is embedded in the culture and needs to be addressed. “Leadership has to engage with employees on this,” she said. “We need to have their [employees'] backs. Visionary program managers get this.”

TECHNOLOGY, INNOVATION, AND PARTNERSHIPS AT THE NATIONAL SCIENCE FOUNDATION

In March 2022, the National Science Foundation (NSF) launched its first new directorate in more than 30 years: the Directorate for Technology, Innovation, and Partnerships (TIP).⁴Graciela Narcho (NSF) explained how TIP will accelerate the NSF mission to promote the progress of science; advance national health, prosperity, and welfare; and secure the national defense. NSF's existing directorates fund fundamental and use-inspired research, and they all do have collaborations, but TIP is designed to work across the agency, and across sectors and communities of practice. The creation of the directorate comes at a pivotal moment in terms of the challenges to address—such as climate change, equitable access to education and health care, and critical and resilient infrastructure—and within a changing science and engineering enterprise. The pace of discovery is accelerating, the demand for social impact is increasing, and opportunities to leverage partnerships across sectors are expanding. TIP is designed as a “new horizontal” to expand and scale use-inspired collaborations. It embodies a paradigm shift to engage users in shaping and conducting research, involve multisectoral teams, and drive research to solve societal and economic problems.

The President's FY 2023 budget request of about $10.5 billion for NSF represents almost a 20 percent increase from the FY 2022 enacted appropriation to the agency. Of the $1.7 billion in new funding, about $880 million is planned for TIP (about $500 million added to existing programs), and the rest for new research in other directorates. TIP focuses on three thematic areas: (1) partnerships as a foundation, (2) innovation and technology ecosystems, and (3) technology translation. As an example of the first area, she described NSF's partnership with Intel to support U.S. chip production through manufacturing, R&D, and workforce development.⁵ The new directorate aims to ensure that it will becomes a resource for best practices, toolkits, and platforms to help facilitate direct partnerships between NSF and the private sector, non-profits, and other partners.

Innovation and technology ecosystems will be supported through the already existing Convergence Accelerator, as well as new programs for emerging technologies, regional innovation, and experiential and entrepreneurial learning.⁶ A new initiative, NSF Regional Innovation Engines, will balance scientific and technical goals with broader societal benefits.⁷ “What we see is an opportunity for innovations in the technology space to shed light on how we approach societal challenges, and for the societal challenges to drive these innovations,” Narcho explained. NSF Engines are an intentionally different approach than previous efforts because they are at a larger scale, are iteratively co-designed and co-created, involve cohort-based training, set up milestones for continued funding, and have focused success expectations.

Expanding geographic diversity and involving a range of different organization are essential. “We want to get to a place where the metrics of success are not just about publications and conference proceedings, but who we are training and the technological capabilities that are emerging,” Narcho said. She noted the Engines are intended to move research to solutions more quickly. Existing examples include ecosystems around robotics in Pittsburgh, medical devices in Nashville, defense in Huntsville, and electric grid innovations in Chattanooga. NSF intends for the Engines program to expand the geography of innovation, so every community has STEM opportunities like these examples.

To grow technology translation, TIP will bring together several existing NSF programs, including the National Innovation Network (I-Corps), Partnerships for Innovation (PFI), small business grants (SBIR/STIR), as well as new innovative pathways. One area under exploration is how NSF can put its funded research outputs on a pathway so nonprofits, cities, and communities can better access them.

Narcho said TIP is building diversity, equity, inclusion, and accessibility principles into all its work at the outset. For example, a goal of the NSF Engines program is the successful inclusion of minority-serving institutions, two-year institutions, and others underrepresented in the NSF portfolio. A recent series of listening sessions with some of these stakeholders pointed out challenges often not brought to the table early enough related to equitable partnerships, capacity-building, NSF policies, geographic isolation, and mentoring.

Since launching in March 2022, TIP is repositioning existing programs, planning new investments, and initiating the recruitment of support and program staff. In closing, she said TIP hopes to serve as a ramp of opportunity to cross the “valley of death” (Figure 1) from lab to market.

FIGURE 1

Potential of NSF's new TIP Directorate. SOURCE: Graciela Narcho, NSF. Presented at a workshop of the Government-University-Industry Research Roundtable on June 29, 2022.

In discussion with participants, Narcho clarified that NSF continues to support the work of investigators in the academic community but is reaching out to new communities, such as industry, community colleges, and nonprofits that have rarely or never engaged with NSF. To incorporate lessons learned from previous Engine-like efforts, Narcho noted the focus on metrics. One participant from a philanthropic organization praised the geographic innovation that TIP is encouraging but cautioned NSF might face pushback “from the coasts and larger entities.” Narcho responded that NSF would continue to work with established partners but wants to grow the geography of innovation. In addressing speed to overcome the valley of death during the time of global competition, Narcho said co-design and co-creation of research, milestones, and sharing early-stage results as they emerge can help.

Jeff Welser (IBM Research), Susan Martinis (University of Illinois at Urbana-Champaign), and John Beieler (Office of the Director of National Intelligence) provided industry, university, and government perspectives on the opportunities raised by the NSF Technology, Innovation, and Partnerships Directorate.

Welser expressed excitement about TIP and shared that IBM Research tries to maintain the balance that NSF is aiming for between exploratory basic research and development of innovations needed by IBM. “It is not always an easy balance,” he acknowledged. One of the few industrial stand-alone labs operating, IBM Research is organized by pillar (AI, quantum, cyber, and hybrid cloud) but also horizontally to feed those pillars and new areas not yet fully envisioned. In 2021, Accelerated Discovery was created as a “new horizontal” to understand how computing can accelerate science and impact society. “This new approach merges with what TIP is doing,” he suggested.

The goal of the Accelerated Discovery horizontal is to create an ecosystem that looks beyond what IBM might do on its own. IBM can develop the technology that can be applied in collaborations, for example, with medical or climate change experts, in an open environment. Economic development, training, and other incentives are incorporated to continue the work beyond the research stage. The key behind accelerated discovery partnerships are advances in computation that can he made available on the cloud and distributed equitably; the use of AI and the computational backbone behind it; and quantum computing. “Those three together will advance how to do science and supercharge the scientific method itself at every step,” Welser said. Global Discovery Accelerators will use technology to solve societal challenges, like TIP, with an Accelerator at the Cleveland Clinic the furthest along. Welser expressed hope that more synergy will accelerate the process for everyone.⁸

Martinis said she was involved in organizing an NSF workshop in 2021 to lay the foundation for a National Network of Research Institutes, with 50 percent of the approximately 60 participants from the private sector.⁹ She drew from a breakout session on regional and site-based innovation to share ideas about how to unite ecosystems. She praised NSF's efforts to bring “different people to the table,” including community colleges, community partners, industry, and others. “Workforce development is happening, but it needs to be done at scale and based on trusted, effective partnerships,” she noted. The 2021 workshop challenged NSF to consider how to incentivize partnerships beyond large academic institutions and raised the issue of how to share intellectual property among partners. She stressed this is an important step for NSF to consider.

John Beieler provided a perspective from the intelligence community about partnerships. Speaking to opportunities, he noted technology and collaboration as underpinning a lot of timely, actionable intelligence for policymaking. Beieler explained, “Intelligence-related problems are time-sensitive, hard, and unique—such as face recognition from acres away. Like the space program, solutions for hard problems in intelligence can become a net benefit to society.” He noted that three trends—the convergence of fields, acceleration of discovery, and democratization of technology—make the environment challenging during a time when economic and national security are converging. The hard problems within the intelligence community can drive the use-inspired challenges and paradigm shifts discussed by Narcho and the other respondents. From an intelligence community perspective, Beieler harkened back to Vannevar Bush's and the post-Sputnik era.¹⁰ “We as a nation have done this before,” he concluded. “We have the parts. It is venues like this that can bring together government, industry, and academia to drive these hard problems towards solutions.”

In discussion, a participant from RTI International noted a common theme between the respondents was how institutions can mentor other institutions in capacity building. Martinis stressed that mentoring occurs from all angles, not just from better-resourced institutions to less-resourced institutions. Martinis noted that many Historically Black Colleges and Universities (HBCUs) do not have offices of sponsored research and would benefit from creative learning and capacity building arrangements with R1 institutions. Beieler added that different government programs have unique proposal requirements, and workshops and other outreach could improve communications on what a good proposal looks like, especially for institutions that do not have a history of working with the federal government. Welser said IBM Research has been thinking about how to extend technology more equitably.

The state of Michigan runs a commercialization shared-service mentoring program for all public universities to capacity build, a participant reported. She welcomed the ideas presented about how to build capacity. “We need a national conversation about how to raise all ships to engage in research,” she said, and thanked NSF and others who are aiming to equalize the distribution of science across the United States. In addition to accelerating the translation of research, it is important to de-risk technology, so investors come in, perhaps as start-up companies. Martinis said this comes back to who holds the IP to attract investors and suggested it may be time to revisit the Bayh-Dole Act (which covers patent rights for inventions made with federal assistance). Beieler added government should find better ways to partner with start-ups. Another participant noted the importance of having partnerships in mind from the beginning, which might mean balancing different priorities. “Better communications about mutual needs; discovering a common goal and complementary skills to achieve those needs; buy-in at the grassroots level; and devotion of time to form relationships—not just to check off the “partnerships” box—are needed,” Martinis concluded.

PARTNERSHIPS TO ACCELERATE AI HEALTH RESEARCH

As presented by its CEO, John Baldoni, the Accelerating Therapeutics for Opportunities in Medicine (ATOM) Research Alliance is a public-private partnership based on the convergence of high-performance computing, subject matter expertise, and data to accelerate discovery and development of more effective therapies for patients. As Baldoni described it, drug discovery is traditionally a linear process in which four “races” must be won: to the target, molecule, clinic, and market. It usually spans 10 to 12 years and costs $1–$2.5 billion to develop a product. One analysis showed an average 4–5,000 molecules must be tested to discover a candidate and it takes several years from discovery to human. “The industry is making people feel better and live longer,” he said, “but it is predicated on profound failure. That is the opportunity for improvement.” This “hard problem” inspired Baldoni to begin the formulation of a unique collaboration network to enable rapid drug discovery.

In 2012, realizing the vast number of data points generated and the fragmented data sources just within GlaxoSmithKline (GSK), where he worked at the time, Baldoni challenged his team to test the possibility of discovering a molecule to modulate human biology within a year. Only one person took up the challenge, but it made Baldoni realize the possibility. He began to form ATOM, which brought together GSK, Lawrence Livermore, the National Cancer Institute, and the University of California San Francisco. Support from the highest level of each organization led to a four-way Cooperative Research and Development Agreement (CRADA) in 2017. GSK provided 3 million dead (i.e., unused) compounds with which to start to experiment. Questions to resolve included how to democratize drug discovery, how to use open-source tools and capabilities, how to create a place for disciplines to come together, and how to train the next generation. The initial goal was to discover a drug molecule in less than a year, along with the development of new tools that can deal with a larger range of molecules than can currently occur. Baldoni explained the foundational ATOM molecular design workflow and target-to-clinical-trial roadmap. Sharing some examples, he noted a key takeaway of each is the value of biologists and computer scientists working together. Much of the work is in the public domain.

The ATOM Network now consists of National Labs, universities, nonprofits, and companies. He welcomed new participants as strategic and associate members into what is now a 501(c)3 organization. The United States has strategic stockpiles for all kinds of materials, he said; similarly, there is a huge strategic stockpile of data hidden in publications and in failed experiments.

Baldoni ended with personal recommendations to enable rapid drug discovery. For government: to create a Drug Discovery Data Strategic Reserve, legislation to fill it, and tools and capabilities to use it, as well as keeping National Labs focused on molecular recognition. For universities, he recommended large collaboration networks, teaching biology as a math and physics discipline, and encouraging curious and imaginative business development offices. For industry, he urged pharma to release failed compound dark data and development of technology based on what users say they need.

PUBLIC-PRIVATE PARTNERSHIPS AT NASA

Phil McAlister (National Aeronautics and Space Administration [NASA]) reported on NASA partnerships with industry for capability development, which he described as representative of a huge innovation and culture change for the agency. The idea of these partnerships began with the phase-out of the Space Shuttle program and the need to re-supply the International Space Station. The prevailing view within NASA at the time was that using private industry was too risky—that industry would cut corners in order to make money or would not be capable of fulfilling the mission. The idea did not get traction until May 2012, when SpaceX successfully delivered cargo to the Space Station. According to McAlister, this changed the game—NASA astronauts are now being launched on privately owned and operated spacecraft.

Quoting Darwin, McAlister observed it is not the strongest or most intelligent species that survives, “but the one most responsive to change.” NASA needed to change, he asserted. The NASA budget will likely not return to its earlier levels; major space programs have been cancelled and buying power has declined. Facing these challenges, NASA looked at new approaches, including public-private partnerships in which “resources, technical capabilities, and risks are shared.” For these arrangements to work, alignment between all participants is necessary. Everyone must have skin in the game—a big change from NASA's traditional approach in which NASA owns, pays for, and defines all aspects of its business. In the new, partnership approach, industry owns the hardware, NASA invests based on milestones, and the industry partner defines how to do the project. As an example, the Space Shuttle had between 10,000–12,000 requirements. In contrast, privately developed cargo missions now must fill about 250 requirements and missions with crews about 650 requirements. It was not easy—in the old model, NASA was in control and thought that control was key to success. “To let go of that control and share accountability is hard,” he acknowledged.

McAlister stressed both the traditional and nontraditional approaches are appropriate depending on the need. One-of-a-kind programs where NASA is the only customer might call for the traditional, NASA-owned approach. Programs with other customers—with no technological breakthrough required—and a strong industrial base may lend themselves well to the partnership approach. Applied in the right way, public-private partnerships combine the strengths of the government and the private sector. Government has experience, expertise, access to financial resources, and ability to forge consensus. The private sector has a laser focus on cost effectiveness and speed. “The combination creates a healthy tension and better results than either entity working individually,” he concluded.

In discussion, McAlister invited participants to see how they might collaborate with NASA through the International Space Station (ISS) National Laboratory.¹¹ He reiterated that relinquishing government control represents a huge change in mindset, such as recognizing private industry “partners,” not “contractors,” with a seat at the table. Applying the NASA case more generally, McAlister concluded that early successes are important to help people recognize what is possible. A portfolio helps mitigate risk; having competition and a Plan B also help.

PARTNERSHIPS TO DRIVE DIGITAL TRANSFORMATION

As brought up by presenters in this session, digital transformation can lead to solving problems in all fields of science, and partnerships are critical given the size and complexity of the technology. Perspectives on the topic were shared by university, industry, and government representatives.

To Kendra Ketchum (University of Texas at San Antonio [UTSA]), the goal of information technology is to “reduce the time to science,” so researchers can go on to the next innovation and not worry about infrastructure, standing up licenses, and information technology-related issues. She stressed, “Digital transformation is a human transformation.” IT can help meet the increasing demand for higher education institutions to provide exceptional experiences that involve all students, faculty, and staff. In addition, the digital divide issue is significant—particularly in the San Antonio area—and institutions must consider the additional challenges associated with connectivity and equity.

In partnership with Dell Technologies, UTSA developed an advanced cyberinfrastructure research platform for the information capital circulating across the university's many institutes and other programs. Ecosystems built for researchers can be torn down and repurposed. The decision of whether to use the UTSA platform or to use the cloud is determined on a case-by-case basis. She characterized the system as a “strengthened core with an agile edge.” More broadly, Ketchum offered seven essential steps to digital transformation. She advocated focusing on the mission of each researcher, mapping out the most appropriate services available, iterating, and learning each time. Ketchum urged participants to reach out to connect with their local tech divisions.

Ketchum also touched on enhanced network security architecture to prevent cyberattacks. She explained the need for segmentation for some data, and open gateways for those resources that do not have to be segmented. Flexibility is again essential, and analytics—including detection rules, heuristics, machine learning, complex systems integration, and automation and orchestration—can reveal behavior and analytical patterns so that institutions can be better prepared for a cyber event.

Deborah Stokes (Dell Technologies) shared Dell's perspective about the partnership with UTSA. She underscored the value of the Triple Helix partnership model—government to provide framework and governance, academia to provide research and skilled knowledge, and industry to provide resources and commercialization. “The magic happens when even two of these three partners join forces,” she said.

In addition to the infrastructure described by Ketchum, Dell has worked with UTSA and other institutions from a hiring perspective for many years. Dell designates senior staff to serve as campus executives to support engagement and looks at academic research in such areas as AI, quantum, and cloud to identify areas for collaboration and to involve students and faculty. “Interactions and relationships are the starting points for the innovation ecosystem that creates value,” she observed. Dell is working with Arizona State University, UTSA, and other institutions as part of the IBM Quantum hub (Q hub). Dell has also connected with a professor at University of California San Diego on data evaluation and blockchain security who is also working with UTSA on cyber. “Bringing pieces together to solve big puzzles is important,” she concluded.

Laurie Locascio (National Institute of Standards and Technology [NIST]) discussed how NIST enables digital transformation through partnerships. NIST is the U.S. government leader in many aspects of advanced communications technologies and has been involved in semiconductors to solve measurement challenges since 1968. Since its founding in 1901, NIST has focused on ensuring the United States captures the full potential of its innovation and competitiveness, from conducting fundamental and applied research to extramural programs in manufacturing to standards development activities. In every state, NIST works with small- and medium-sized manufacturers, industry, and other centers and labs. NIST work extends across a broad portfolio, including AI, cybersecurity and privacy, quantum computing, advanced communications, semiconductors, and more. In AI for example, NIST scientists are working on scalable research-based methods to assess risk and ensure AI is used in responsible, equitable, and beneficial ways. The agency is leading development of an AI risk management framework to better manage risks to individuals, organizations, and society associated with AI.¹²

NIST released a Cybersecurity Framework in 2014, created with extensive stakeholder input from across all sectors of the economy.¹³ In June 2022, NIST announced it will update the framework, which—along with a privacy framework developed in 2020—is a voluntary tool to identify and manage risk without holding back innovation. Locascio also spoke about NIST's programs to create the cybersecurity workforce of the future, including the National Initiative for Cybersecurity Education, and development of tools to create a privacy-literate workforce.

A key aspect of the NIST research program in quantum computing are three collaborative institutes—JILA, the Joint Quantum Institute, and the Joint Center for Quantum Information and Computer Science. “These types of partnership are critical to NIST success,” Locascio stressed. “They are more than just grants. They are collaborative research programs where we send scientists to work side by side with the partner institute.” NIST is also completing a multi-year project in post-quantum cryptography.

Locascio closed by underlining that NIST strongly supports collaboration as part of its work to develop standards that underpin the world's digital economy and more broadly to promote U.S. economic security and innovation.

Co-chair and moderator Al Grasso asked about the expectations for external partners related to cyber security to ensure the protection of the ‘weakest link.' Ketchum noted that at UTSA, the internal infrastructure and cloud services can be brokered to ensure external entities reach the internal levels of operations, so that security principles can be applied from end-to-end. At NIST, Locascio said, partners are responsible for the cybersecurity on their campuses, and added that NIST urges partners to implement its Cybersecurity Framework. In so doing, NIST learned it is hard for small and medium-sized organizations to comply with the framework, so the agency is looking at how to improve its applicability for all organizations. At Dell, Stokes said, partners must have the same level of security and training as Dell expects for itself.

A participant from UIUC asked how to work toward incorporating essential digital transformation talent who are foreign-born individuals. Locascio noted NIST has partnerships all over the world but must think about what information can be shared and how to wall off some activities. To the larger issue, she continued, the United States has about 330 million people compared with 1.3 billon in China—making the point that while it is important to capture all U.S. talent, the U.S. enterprise needs talent from other countries too. Ketchum added that the UTSA cyber program aims to prepare students who can be placed nearby to improve the security of the regional ecosystem. From an industry perspective, Stokes said Dell, as a global company, has structures to determine which teams can work on which projects from a security perspective. She urged not overlooking that a lot of top talent is outside the United States.

When asked about the role of standards in ensuring American competitiveness, Locascio said if the United States invents the global standard for a product or service, U.S. companies benefit. She explained that standards exist to open markets, noting that standards and technical regulations underpin approximately 86 percent of global trade.¹⁴ Discussions around 5G and electronic vehicles are top of mind in this area. Traditionally, Locascio noted, the best technical solution has resulted in a global standard, but world standards are becoming more political. NIST is traveling and talking with other countries to protect the emphasis of standards development on identifying the best technical argument to ensure that the process does not become politicized.

PARTNERSHIP FOR INCREASING SEMICONDUCTOR MANUFACTURING COMPETITIVENESS

In introducing the final panel, GUIRR co-chair Darryll Pines noted the critical role of semiconductor manufacturing in the past decade and looking ahead.

Grace Wang (Ohio State University) described what is happening around semiconductor manufacturing in central Ohio. With Intel planning to open new facilities, she said, it is an exciting time to build an entire semiconductor ecosystem. As projected by McKinsey & Company and others, semiconductors will continue to be a vital enabler of the growth of critical areas of the digital economy.¹⁵ While U.S. semiconductor companies account for 47 percent of global sales, only 12 percent of semiconductors are manufactured in the United States (down from 37 percent in 1990).¹⁶ The top priority for adding manufacturing capacity is the development of skilled talent in the workforce. Wang noted that the United States invests in semiconductor R&D as a percentage of sales at a higher level than other countries, including China, but most of the value-added of U.S. investments is in R&D-intensive aspects—very little is for materials, fabrication, and assembly.

“Research investments must translate into impact for the economy,” she said. She urged building on best practices—pointing to resources developed by the University-Industry Demonstration Partnership—to grow the talent pipeline at scale, increase the density of co-located partners, and enable a highly connected network (Figure 2).¹⁷

FIGURE 2

Evolving cross-sector partnership models. SOURCE: Grace Wang, Ohio State University. Presented at a workshop of the Government-University-Industry Research Roundtable on June 29, 2022.

With Intel coming to the region, Ohio State has focused on partnerships to build the semiconductor and STEM workforce involving Columbus City Schools, career technical centers, community and technical colleges, and 17 institutions that offer bachelors and graduate degrees. Columbus City Public Schools has about 50,000 students, most from lower socioeconomic communities. A new initiative, STEAMM Rising Columbus, will create a summer institute to train and develop 500 STEAMM teachers in the next five years.¹⁸ Keeping scale in mind, Wang noted the program will “train the trainers,” so that participating institutions enable new pathways for learning, teaching, and partnering across STEAMM fields. In addition, Ohio State is building an Innovation District to co-locate students, faculty, industry, entrepreneurs, and others to work together, including an area for semiconductor research and collaboration.

Gabriela Cruz Thompson (Intel) shared Intel's thinking about academic collaboration and partnerships with the purpose of creating world-changing technology that enriches the lives of all. She shared two priorities—first, to innovate with boldness, which gives permission to fail; and second, to fail fast, or learn fast. “This cannot just be done with corporate resources,” Thompson said, and consequently noted Intel considers relationships with academia as critical for innovation. She also shared that Intel is also looking for ways to increase representation and support a vibrant culture for its 121,000 employees in 53 countries. She related a quote from former chairman Andy Bryant: “The ingredient we start with is sand. Everything else is value added by people.”

Intel wants to grow manufacturing operations in the United States, with increased capacity in Arizona and new facilities in Ohio, resulting in tens of thousands of direct and indirect hires. To accomplish this, Cruz Thompson shared that Intel partners with academia, government, and other companies, using outreach mechanisms to several ends: to sense new ideas and breakthroughs through large-scale collaborations; to transfer knowledge through midsized, more Intel-focused centers and individual grants; and to grow talent, including Intel's Academic MindShare and campus recruiting. The company is now also concentrating on curriculum needs and reaching out to community colleges in addition to universities. Intel projects the future STEM workforce will be comprised of Ph.D., master, and bachelor degree holders (each at about 20 percent of the whole) and associate degrees or certificates (the remaining 40 percent). Through Intel RISE (Responsible, Inclusive, Sustainable, Enabling), Intel is seeking to make a positive impact on society, business, and the planet.

Eric Evans (MIT Lincoln Lab) spoke about defense-focused R&D collaborations for advanced microelectronics (ME). He noted that ME fabrication objectives are aimed at both economic security, as discussed throughout most of the workshop, in addition to Department of Defense (DoD) needs. DoD's unique needs for ME technology include creating an asymmetric advantage for the United States; securing access to commercial off-the-shelf parts in acquisition; fabrication of very specialized parts; maintenance of a high level of trust for the products; and mitigation of the issue of obsolescent parts for some systems. There are fabrication capabilities in labs across the nation, including at Lincoln Lab, DOE national labs, and elsewhere. “There are opportunities for collaboration with academia and industry to work with these large facilities,” he said.

The ME Innovation Ecosystem Concept has its hub at MIT. They look at disruptive concepts, often projecting out decades. For example, early investments to develop digital focal plane array technology began in 2004 and was considered mature enough to be applied to DoD systems in 2018. According to Evans, from a national security standpoint, disruptive areas for extended investments lie in advanced computing and software, edge-based AI and autonomy, directed energy, integrated sensing and cyber, and space technology. “The best way to drive collaboration and development of new technologies is to present big problems to solve,” Evans said. He shared several examples of big problems, including the need to replace aging radar infrastructure with better-performing, less costly, phased arrays, or to reduce the enormous amounts of energy currently needed to run data centers. Lincoln Lab recently broke ground on the Compound Semiconductor Laboratory Microsystems Integration Facility (CSL-MIF), which Evans said will be another opportunity for collaboration with the academic and industry communities.

In discussion, Pines asked the panelists about the need behind the CHIPS Act for U.S. competitiveness.¹⁹ Thompson predicted the importance of the legislation for development and innovation over the next 10 to 20 years, noting the inclusion of the directive to the Department of Commerce to launch a National Semiconductor Technology Center, which would bring industry, government, academia, and others together for experimentation. “It will be the way to train the next generation of semiconductor workers,” she added. Wang suggested that CHIPS funding will lead to long-term impacts and industry contributions of $1 trillion by the end of the decade. Evans pointed out several parts of the legislation that would strengthen national security.

A participant from UIUC noted that COVID-19 vaccine development crashed the prevailing view that vaccine development required 10 years or more and asked about ways to crack the time barrier in semiconductor technology. Thompson address current barriers, including the necessary coordination of large-scale collaborative experimentation and a trained and ready workforce. She saw the first barrier as something CHIPS might address, but the second as both a short-term and a long-term problem. She noted that deans of engineering are finding students moving from away electrical engineering to computer science.

Another participant from NIH suggested tapping into young, tech-savvy students, even if they do not currently have the necessary education. Wang said this applies to many needs beyond semiconductors. “The country needs to rethink formal and informal engineering education,” she suggested, and to reach out to elementary schools. Thompson observed that affordability of education is an issue, but in reflecting on the efficiency of a three-week boot camp experiment with high school graduates she noted that ten years ago the same content might have taken a whole year to convey. Evans agreed that hands-on training is the best approach, some of which can be done through kits and interactions with experts online. A participant suggested workers whose jobs are being automated could be retrained and credentialed.

Grasso asked how to sustain progress and momentum on semiconductor manufacturing policy this time. Wang concurred with the need for a cohesive, long-term effort. “It took three decades to lose some capacity; it will not come back in 1 or 2 years, no matter the level of funding. We are at an inflection point, not just for semiconductors, but in quantum, EV, gene therapeutics, and more. There is no magic—just do it consistently,” she said. Thompson posited that creating a new ecosystem with new economic possibilities is important. She expressed hope that the NSF Engines program and other efforts can encourage more entrepreneurship and entry by more semiconductor design and manufacturing companies. Evans said federal support plays a role because federal funding attracts great people and creates new self-sustainable markets.

In closing, GUIRR co-chair Pines reflected on the workshop. He noted it began with a call for innovative partnerships and ended with optimism about breakthrough technologies and disruption. He suggested that GUIRR continue to serve as a convener for conversations that tease out these new models and new ways of coming together to launch new endeavors. “Things do not happen by chance,” he underscored. “The passion of the individual brings initiatives to fruition.” Even though the U.S. system is sometimes inefficient and untidy, Pines said, through cross-sector partnership, the U.S. enterprise continues to enable individuals and institutions to innovate.

Suggested citation:

National Academies of Sciences, Engineering, and Medicine. 2022. Enhancing U.S. Science and Innovation with Novel Cross-Sector Partnerships: Proceedings of a Workshop—in Brief. Washington, DC: The National Academies Press. https://doi.org/10.17226/26830.