2

Increasing U.S. Competitiveness by Improving Knowledge Creation and Technology Diffusion

Over the course of the workshop, speakers explored U.S. competitiveness, gaps in the U.S. innovation system, and ways that the Manufacturing USA institutes can avoid possible market failures associated with knowledge creation, such as through collaborative public-private partnerships. Speakers discussed questions related to long lead times for commercialization for new technologies, such as: Do the collaborations enabled by the institutes help accelerate lead times? Does the participation of small firms in the institutes provide them with better access to technologies that they would not otherwise be able to access? Do the advances in workforce development across the institutes help firms, regardless of size, adopt new technologies?

In his remarks, Brett Lambert of Northrop Grumman discussed the importance of manufacturing in the economy, particularly its impact on wages. Where there is a concentration of manufacturing, the impact on wages spills over to the surrounding economy. Total earnings, he said, are much higher for those in manufacturing jobs and those in jobs that are located near centers of manufacturing. He added that these figures have been increasing over the last two decades, “both through the [economic] decline and through the upturn.” Overall, according to Mr. Lambert, after controlling for education and experience, “employees that are involved in any way in manufacturing [receive] about 15 percent more in total compensation” than other workers.

Willy Shih of Harvard Business School echoed Patrick Gallagher’s (University of Pittsburgh) thoughts on the importance of co-locating R&D and production for U.S. industrial competitiveness. In addition, he argued that the United States may have underestimated the significance of economies of scale in production for industrial competitiveness. The need for manufacturing at scale points to a worrying deficiency in the United States’ ability to match China’s manufacturing competitiveness. For example, while more than one billion smartphones are produced each year across East Asia, none are produced in the

United States. China’s automobile market—more than twice as large as the United States’—leads to more opportunity for Chinese companies to practice and develop manufacturing techniques. China’s ability to leverage the economies of scale in production, combined with its Made in China 2025 strategic plan–which is designed to enhance its high-tech capability–may also put the United States at a competitive disadvantage. While the manufacturing institutes are working to be a hub for industrial R&D, he articulated the need for policies to drive those R&D and innovation gains into consumption and production.

Dr. Shih also mentioned the need for much more workforce development in order for the United States to be competitive. He highlighted the lack of parity between China and the United States in terms of scale in industrial engineering and manufacturing talent. To visualize his point, Dr. Shih showed a video of a massive Gen 10.5 LCD factory in Hefei, China that is able to produce large sheets of glass only 0.7 mm thick in a building that is 1.3 km in length with hundreds of thousands of square meters of clean rooms. Though most of the production is automated, substantial human engineering support is needed to run the facility. “There are no engineers in the United States who know how to do this level of industrial engineering or work on this type of process,” Dr. Shih said, referencing the scale of the project. Industrial R&D is important, he argued, but so is the know-how that comes from being able to scale production.

Erica Fuchs of Carnegie Mellon University said that the ability of U.S. companies to innovate in advanced materials and processes requires U.S.-based manufacturing and that the recognition of this need, and of related national security considerations, was behind the creation of the manufacturing institutes. She highlighted the large capital investment and time needed for manufacturing to make it to market, especially in advanced materials and processes. The outputs of the domestic innovation economy in software and artificial intelligence, for example, are dependent upon underlying hardware. Dr. Fuchs said that only about 10 years may remain on existing semiconductor technology until advances are needed to accommodate progress in machine learning and artificial intelligence.

Another reason to keep manufacturing in America is that moving manufacturing overseas undermines the profitability of firms pursuing the most advanced technologies, such as in the areas of optoelectronic semiconductors and solar and battery technologies, Dr. Fuchs said, citing empirical evidence of a “global race to the bottom” in such a scenario. Firms in the optoelectronic semiconductor industry, for example, are attracted by bigger markets and lower costs overseas, but in moving offshore for shorter-term gains they abandon new and emerging technologies in the United States. The question then becomes how can that gap between production and technology be bridged in the United States, and how can products be made in the U.S. that are wanted by everyone globally?

The manufacturing institutes are thus a grand policy experiment, Dr. Fuchs said. They seek to scale the commercialization of advanced manufactured products by bringing together different parties from industry, academia, and government. Whereas past public-private partnerships like SEMATECH and the Semiconductor Research Corporation relied on models that focused on

semiconductor equipment upgrading and bringing together government funding for academic research on silicon semiconductor research, the institutes focus on general purpose technologies serving diffuse interests and broad needs across multiple areas of industry. Manufacturing USA’s usefulness comes in when market, government, and institutional failures result in underinvestment in advanced technologies (like beyond-CMOS¹ devices) critical for advanced economic growth. Because of the Manufacturing USA institutes’ varying technological missions, focus areas, and funding approaches, technological expertise from government agencies such as the National Institute of Standards and Technology (NIST), the Defense Advanced Research Projects Agency (DARPA), and regulatory bodies such as the Federal Aviation Administration (FAA) impacts the industries that the institutes serve in different ways. More can be done to enhance the institutes’ ability to orchestrate these actors and academic partners toward the needs of the diverse industries that the they serve, from photonics to advanced fabrics.

Andrew Bowd, chief strategist of the American Institute for Manufacturing Integrated Photonics (AIM Photonics, or AIM), highlighted the institute’s role in providing infrastructure, design, and knowledge transfer through private sector and academic collaboration in the integrated photonics space. It is important to teach engineers and scientists to design, build, and manufacture light devices, said Mr. Bowd, in order to sustain and advance technology proliferation. Like AIM, all of the Manufacturing USA institutes support similar initiatives to create design ecosystems in which knowledge is transferred among workers, firms, and academia. AIM itself has connected firms ranging from defense contractors to medium-sized businesses with domestic foundries, making available the use of its own semiconductor fabrication plant (“fab”), packaging facility, and intellectual property (IP).

Speaking in his capacity as director of emerging and external technologies for Ashland, Inc., a specialty chemicals company, Joseph Fox described his role as connecting Ashland with other companies, universities, and federal laboratories that have already developed technology and products it can use—and the value that has been brought to the company as a result of its membership in the Institute for Advanced Composites Manufacturing Innovation (IACMI). Mr. Fox explained that IACMI is focused on reducing the cost of composite fabrication as well as job creation and improving the recyclability of composite materials. It has been effective in bringing together industry, universities, and federal laboratories to collaborate on industrially relevant technology, particularly on driving the adoption of advanced composites by manufacturers like Boeing and BMW.² Key to its mission, he said, is the co-location of IACMI’s commercial focus areas—lightweight vehicles, wind

___________________

¹ Complementary metal-oxide-semiconductor.

² IACMI is a public-private partnership that consists of seven centers, two federal labs (Oak Ridge National Laboratory and National Renewable Energy Laboratory), and five universities (Purdue University, Michigan State University, University of Tennessee, University of Kentucky, and University of Dayton Research Institute).

turbines, compressed natural gas storage, and process technology—in sector-dominated regions like Michigan (automotive production), Ohio (compressed natural gas storage), and western Pennsylvania (shale gas). Ashland’s collaboration strategy has seen the firm involved with university research centers and cooperative research programs, and more recently as a member of IACMI. There is considerable overlap between IACMI’s interests in advanced composites and Ashland’s needs for advanced composites technology, said Mr. Fox, and a key benefit of membership in IACMI has been access to the technology and expertise of federal laboratories in addition to universities.

IBM’s Kathleen Kingscott offered remarks regarding important questions for IBM including how, in a competitive global environment, the institutes create knowledge, enable technology transfer, and bridge the “valley of death” between innovation and commercialization. As a global business that participates in a variety of industry consortia around the world, IBM is compelled to look at how the institutes operate with respect to the “global state of play,” she said, citing the need for U.S. leadership on certain technologies like artificial intelligence.

Ms. Kingscott discussed the importance of having the right composition of organizations in the manufacturing institutes, including both large companies like IBM, with their broad set of capabilities and deep organizational structure, and small companies that typically have niche experience in a technology area that is of value to the larger organization. Putting together these kinds of partnership groups makes sense, but only if there are leaders in these groups and not just followers. There can be 100 players in an institute but only two or three leaders, something that “can be a troubling characteristic,” Ms. Kingscott observed. This is true not just of the manufacturing institutes but of other consortia as well.

Unlike the metrics used by NIST to evaluate the institutes, IBM evaluates Manufacturing USA with its own metrics based on the cost and return of participation in the institutes in terms of time, technology, and partnerships with other members. The closer that a consortium like the Manufacturing USA institutes meets the goals that were jointly set at the beginning—and the clearer the goals and conditions so that all actors are on the same playing field—the more useful that consortium is to a firm like IBM. Strong leadership from each of the partner sectors is a critical criterion for a decision by IBM to join an industry consortium. Other important considerations are whether a technology fits into IBM’s strategic needs and provides value, who the other members of the consortium are, and whether the other members are willing to bring their assets to the table and to exhibit leadership.

One of the biggest challenges for any consortium like the institutes is how to deal with IP and the ownership, licensing, and assigning of royalties to partners. A model IP contract is one in which, in IBM’s view, members own what they bring into a consortium, but what is created in the consortium is jointly owned among members. While affirming that the institutes have been successful in the area of technology transfer, Ms. Kingscott described remaining questions

about whether some institutes are really tied to the needs of industry and are not just technology showcases. Another consideration is whether the flexibility of the institutes is sufficient enough to accommodate industry differences in time frames and capital requirements. Such flexibility is an important criterion for IBM to think about when it considers membership in the institutes. There are also frictions between industry and other participants around the appropriate time frame for technology development, and there is a need to be mindful of federal and state funding and to ensure that the industry share of institute funding ramps up so that institute efforts do not collapse at the end of the five-year federal period.

Ms. Kingscott then discussed new approaches to technology development, transfer, and competitiveness in the context of a new advanced technology center in Albany, New York, built in the same facility as the AIM Photonics institute. Rather than starting with a broad technology, the new center will work with core requirements from a government agency. Instead of having a broad coalition of organizations from the start, it will involve only a few partners who bring very specific technological capabilities. A common IP library will be created that enables these partners to bring technology developed at the center back to their organizations, unlike how the institutes traditionally work in this field. “There are lots of models for success. . . [W]e need to try a variety of things” to ensure the United States remains the leader in the current strategic environment, she said.

Ms. Kingscott indicated that the origins of this collaborative IP model is a program funded by DARPA, now known as the Joint Program, where the Department of Defense entered into a “50/50” relationship with industry partners, and where background IP remained the property of its creators while jointly created IP was the property of the consortium. Mr. Bowd remarked that although AIM Photonics has tried a similar model, it has found it difficult to execute a collaborative IP model due to IP challenges between members. It does help, he added, to have visionary and influential people in large corporations who stand up to, and even break, corporate norms.

Ira Moskowitz of the Massachusetts Technology Collaborative (MassTech) cited Massachusetts’ long-standing model in which companies share IP that is jointly developed while keeping their own IP. MassTech is a state agency of the Commonwealth of Massachusetts that oversees the investments of the state in innovation, technology, and advanced manufacturing. There is an element of equal gain and degree of trust that partners in the institutes expect from one another, and both parties stand to suffer if one does not work with the other. There is “a risk of diffusion,” though, for entities who bid on a project call and display to at least a subset of the institute what they are doing.

More broadly, Mr. Moskowitz spoke about the role that MassTech plays in accelerating the creation and diffusion of advanced manufacturing technology in Massachusetts. Innovation and manufacturing are inextricably intertwined, he said, as the knowledge gained by manufacturing products has led to the next generation of products and applications. In order for the state of Massachusetts to

participate in the manufacturing supply chain, it must retain its innovation economy, and MassTech oversees the state’s investments in that retention effort.

Referring to the long lead times and high costs required for developing advanced semiconductor technology, Mr. Moskowitz said that the manufacturing institutes do accelerate those time frames and reduce costs, but that the acceleration can be even greater with a more impactful state partnership: “I do believe that there are things we can do better with Manufacturing USA going forward,” he said. The Commonwealth of Massachusetts recognizes those challenges, committing to advance the Manufacturing USA infrastructure across the state with an investment of $100 million in a number of institutes in Massachusetts, including Advanced Functional Fabrics of America (AFFOA), AIM Photonics, the Advanced Robotics for Manufacturing (ARM) institute, and NextFlex. This investment, he said, is used to form partnerships to “provide connectivity between entities across the state to promote the acceleration of the advancement of these technologies in these institutes.” By doing so, it deepens the institutes’ ability to form partnerships between entities across the state, while emphasizing regionality to amplify and enhance the success of the partnerships. With the state investment, institutes, research centers, and regional industry partners now have joint access to a multi-million-dollar secured foundry at MIT Lincoln Labs, and local companies are able to develop new capabilities and products and do trusted-foundry defense work. New facilities funded by the investment such as the Discovery Center enable the institutes, universities, businesses, and national laboratories to collaborate in the intersections of their technology within one building. Also, the state has become a node of the NextFlex institute, allowing companies and universities in Massachusetts to collaboratively bid on project calls put out by NextFlex instead of competing against one another and against local universities. (See Figure 2-1.)

With regard to who leads the partnerships set up under MassTech grants and the scale of grants awarded thus far, Mr. Moskowitz said that a partnership needs, at minimum, to be between an industry and university member, either of which can be the lead principal, and that over two dozen grants totaling about $50 million have been awarded to date.

Mr. Moskowitz indicated that job creation is the most important metric that companies have to meet in Massachusetts in awarding grants. He added that grant recipients are typically not single companies but a collaboration of multiple entities and companies, all generating jobs. Eight additional criteria are used for evaluating grant applicants, including business expansion and the number of people trained in advanced manufacturing skills. In parallel, the relevant institute also has to approve the grant, ensuring that it advances the institute’s mission.

Joseph Fox highlighted additional benefits that Ashland has experienced as a result of its membership in IACMI. Its membership has given the company

access to equipment and expertise that it did not have to buy or hire but that already existed at its collaborator firms and IACMI centers. It has also resulted in access to money from federal, state, and industry sources, allowing Ashland to leverage its resources with external funding. Perhaps most notably, he highlighted the networking and partnering opportunities with member companies—many of which are small and medium-sized enterprises—spread throughout the composite value chain, observing that these connections have advanced technologies within the company and led to new business. For example, Ashland led a project that involved three IACMI centers and three industrial partners to develop resin and sizing technology for carbon fiber composites, a collaboration serving both Ashland’s technological needs and fulfilling IACMI’s mission of finding ways to recycle process waste and end-of-life waste in the manufacturing cycle. In the end, Ashland was able to develop a system with advantages in processing, molding, recycling, and re-use relative to the existing technology. New opportunities in non-automotive applications for its technology, originally designed for lightweighting vehicles, also emerged. Because of Ashland’s

involvement in this project, Mr. Fox said that the company became a resin supplier for two other IACMI projects, making a name for itself in the advanced composites community.