Models and Ecosystem to Facilitate Breakthrough Innovations

Guides & Resources

Models and Ecosystem to Facilitate Breakthrough Innovations

Summary and Background

The Defense Advanced Research Project Agency (DARPA) has generated a series of innovations that revolutionized U.S. national security, from the Internet to GPS, speech recognition, artificial intelligence, and mRNA vaccines. The Economist called DARPA an agency that “shaped the modern world,” and it models the values, spirit, and technical structure that have the potential to spur breakthrough solutions to some of the most significant problems facing humanity.

Nonetheless, the DARPA model represents only one type of research and development (R&D) process and philosophy that has given birth to numerous innovative solutions to significant problems. Across government, philanthropy, nonprofit, and private sectors, leaders have experimented and found great success in a number of models designed specifically for generating breakthrough innovations.

In U.S. K-12 education, the need for bold, new innovations has never been greater. The 2022 National Assessment of Educational Progress (NAEP) scores for 9-year-olds revealed the largest drop in reading scores since 1990 and the first-ever decline in math scores. Recent PISA results reveal a 13-point drop in American math scores among 15-year-olds – further evidence of the learning loss inflicted by the pandemic. Worse, there has not yet been a meaningful sign of post-pandemic learning recovery. In education, we can’t afford disruptions – especially during a difficult transition – because their impacts on student outcomes deepen over time.

Meanwhile, the field of education receives a disproportionately small amount of federal investment in R&D – the vehicle for innovation – compared to fields like healthcare and defense. In 2022, the Department of Education spent $405 million on R&D, but the Department of Health and Human Services’ R&D budget was as high as $42 billion. In fact, under the President’s Fiscal Year 2023 budget request, the Department of Education was one of the two federal agencies that saw a decrease in R&D funding.

This underinvestment in education R&D means that innovation in teaching and learning has taken a backseat to basic research. Certainly, there is a need for new tools, approaches, and programs that can transform K-12 education and accelerate learning recovery so desperately needed in schools across the U.S. At a more fundamental level, there is an even greater need to identify and build the kind of infrastructure that is geared to meet the education needs of a community, has the potential for scale, and is achievable with the available time and resources.

Collaborating with colleagues and partners who care deeply about innovation, the Walton Family Foundation and The Learning Agency gathered insights into emerging processes and architectures with the potential to transform education. In our work, we have interviewed and learned from many leaders – in government, philanthropy, academia, and industry – who generously shared their expertise, supplemented by research and our experience with innovative projects in education. Additionally, we hosted a Breakthrough Innovations Workshop on Oct. 26, 2023, in Washington, DC to convene and learn from experts in and funders of “big bets” in education R&D.

This brief explores existing or emerging models and organizations – from a variety of sectors – designed to create breakthrough innovations and examines in detail the different components core to each model. To be sure, many of these models and organizations that promote breakthrough innovations are nuanced and can overlap.

In its descriptions and comparisons of different models, this brief aims to provide policy, philanthropic, nonprofit, and private sector leaders with knowledge they can use to select and leverage the model that best suits their needs and contexts.

Key Insights From Exploration

By surveying the current landscape of breakthrough innovations and analyzing their performance, The Learning Agency has gathered key insights for creating successful research and development models for breakthrough innovations. For instance, the most prominent, successful models to promote breakthrough innovations have components such as:

A focus on talent
Multidisciplinary collaboration
Clear, transformative goals
Thoughtful consideration of the end user
An appetite for risk and comfort with failure
Standards for success and failure
Fixed timeframes

Aside from their common components, these models differ in other respects. Some of these tensions across the models include:

Level of risk
Problem definition
Level of financial investment
Barriers to entry
Maturity of the desired end product
Agenda-setting
Team composition

The following sections will discuss these key qualities and decision points in detail, in the context of successful research and development models, to inform any new approach to facilitate breakthrough innovations. Then, the brief will provide an analysis of the broader innovation ecosystem to support leaders in identifying the role and possible integration of any new infrastructure into the existing system. Finally, a few recommendations are provided as promising directions for future infrastructure design and implementation. Ultimately, the brief aims to provide detailed contextual information regarding creating new R&D infrastructures that can facilitate breakthrough innovations and stimulate discussions that may contribute to fruitful ideas about how these innovation structures may seek to address some of the most pressing issues in education.

Facilitating Breakthrough Innovation

Recommended Components Across All Models

There are a variety of successful models for accelerating and achieving breakthrough developments across disciplines. This brief discusses the history, components, and applications of prominent models and approaches, including Advanced Research Project Agencies (ARPAs), Virtual Institutes (VIs), Focused Research Organizations (FROs), challenges and competitions, and talent investments. Although they may differ in goals and emphases, they share a few qualities that have been known to breed breakthrough innovation in an effective, replicable, and scalable manner. This section seeks to identify broader trends and patterns in these models and organizations that successfully promote breakthrough innovations and invite more discussions on the nuances and subtleties involved.

To address specific needs or problems, those making “big bets” in R&D tend to select models with one or more of the distinct elements below. While not a comprehensive list, these elements include:

A focus on talent. Some architectures and processes to promote breakthrough developments emphasize talent to design and execute a solution. This creates a sense of agency and independence in the problem-solving process that gives rise to creative and original solutions. By directing skilled, high-potential individuals to focus on the world’s most pressing or challenging problems, many programs are able to drive a “multiplier effect” in facilitating long-term societal progress. This focus on talent manifests in different formats – from supporting exceptional talents in executing curiosity-driven projects with full independence to helping young or mid-career talents pivot into the public sector.
Interdisciplinary collaboration. Collaboration across projects and sectors will promote new ideas and learning. To achieve breakthroughs in a specific problem, successful models have often pooled resources and experts from a variety of backgrounds and disciplines, sometimes in the form of cohort-based learning. Some of these experts may not even have previous experience in the targeted field, but they can provide a fresh outlook and expertise that catalyze the innovation process. That’s why interdisciplinary collaboration can spark novel solutions from an unprecedented combination of ideas and features, and those that would not have otherwise come to life.
Clear, transformative goals. Many effective models prioritize well-articulated, transformative goals over incremental advances. These goals, along with the focused problem area, should be kept in mind throughout the project. Goals should not be diluted by the multitude of problems that can arise throughout the lifecycle of a project. This approach allows groups and institutions to lead focused, concentrated efforts to examine some of the most significant problems facing society that are often difficult to tackle with traditional research activities, which tend to produce incremental change.
Thoughtful consideration of the end user. Even when models are able to facilitate technical success, their outcomes aren’t always adopted or commercialized. This occurs frequently when projects fail to survey and incorporate end users’ needs. To create both technical and downstream impact, innovations need to be developed with both research expertise and an understanding of targeted users. Teams can develop the latter through intentional design around user engagement through surveys, interviews, and other opportunities for input throughout the R&D process.
An appetite for risk and comfort with failure. To successfully generate breakthrough innovations, models are often built with a risk-taking culture that encourages ambitious goal-setting, experimentation of new ideas, and comfort with failures. With this risk-taking culture, “failures” are inevitable – by incentivizing teams to perceive unsatisfying results as opportunities to learn and share knowledge, many models can accelerate the research progress. From failures, teams can build on insights and make adjustments, which may lead to breakthrough transformations.
Standards for success and failure. Identifying effective metrics for success and failure is crucial to managing any project. It is important to consider the weight of a variety of factors including technical results, commercial outcomes, social value, and the short- and long-term impacts.
Fixed timeframes. Many programs for breakthrough innovations are time-bound in their existence, as well as internal processes such as team funding and program leadership terms. This approach creates a sense of urgency and mission for program leaders and talents to identify meaningful goals, efficiently plan and execute projects, document lessons learned, and share knowledge with the wider community.

Tensions To Consider When Selecting A Model

While different models for breakthrough innovations share some key traits, there are other aspects of the models that vary. These tensions help determine the type of model that may be the right fit for a specific situation.

High Risk vs. Low Risk. Models can vary by their level of risk, based on how much evidence already exists for a particular idea or innovation. Whereas ARPAs, competitions and challenges, and FROs are more likely to invest in nascent ideas that may initially have a limited evidence base, other models like accelerators (a startup support) prefer to back ideas that are well-tested and present lower risks of failure. It is important to determine the rate of success that the model will be aiming for and to design the model accordingly.

Well-Defined vs. Loosely-Defined Problem to Solve. Most of these models offer well-defined problems for participants to solve. However, others provide more flexibility and rely more on talent to articulate the problem to solve and design a potential solution. For instance, each ARPA clearly defines a specific mission or a set of overarching goals, and program managers are given a significant amount of responsibility to select projects in various technical areas that they believe, together, have high potential to advance the overall mission. And other talent investments are less about solving specific problems, and more focused on realizing the potential of exceptionally talented individuals. While talent investments may focus on specific issue areas, like the Pahara Institute for education leaders, some support a vast array of thinkers and doers, as in the MacArthur Fellows program.

Low vs. High Financial Investment. Not all models require the same level of financial backing. Competitions and challenges can be relatively inexpensive endeavors, known for spurring significant research and development relative to their investment. Nonetheless, the amount of financial investment often includes funding for not only administering the competitions – prize competitions also necessitate more spending on marketing and development, while challenges often absorb the cost of supporting ideas that don’t succeed. In contrast, an ARPA, VI, or FRO requires a significant financial investment to explore the possibilities of innovation and engage in sufficient iteration and adaptation. With ample investment, the latter are able to gather exceptional talents and seasoned experts to concentrate on achieving the kind of long-term and enduring outcomes that can accelerate other research in the field through a “multiplier effect,” rather than being concerned with the fear of big risks and short-term failures. In considering this decision point, it can be helpful to understand the model’s endpoint – whether that’s identifying high-potential ideas, taking them to scale, or both – which often demand different levels of investment.

Low vs. High Barriers to Entry. Some models are designed to attract a wide and diverse array of innovators and therefore have low barriers to entry. Others are intended to be more exclusive or eligible to enter only by invitation or in-network referrals, therefore posing higher barriers. Prize challenges and competitions can be established with relatively low barriers to entry, with the intention to secure interest and participation from individuals or teams with wide-ranging levels of expertise or at different stages of product development. Startup supports, too, may be offered to innovators with varying levels of experience and expertise. By contrast, models like ARPAs, VIs, FROs, and talent investments tend to set a higher bar in terms of experience and expertise for their participants. These entry requirements vary, depending on the model’s intended participants – some models may bring in participants from all stages of their lives and careers; some seek to engage underrepresented communities; and others only intend to gather established experts.

Early vs. Mature Product. Not all models aim to generate end-products or innovations with the same level of maturity. Competitions and challenges often source solutions at various stages of development. Educational technology competitions, for example, may fund anything from an algorithm to an existing app or software. Although these solutions may be at an early stage of development, they often have the potential to scale widely or benefit the research community in the long run. It can be meaningful to consider whether the model targets early or mature innovations, and if it will be equipped with resources designed to support innovations at those specific stages of development. Many models designed for early products also provide the needed support from product testing to networking and scaling. ARPA-E, for instance, embeds a tech-to-market team in every research team.

Bottom-up vs. Top-down Agenda. Models like ARPAs, VIs, and FROs often aim to reach overarching, predetermined technical goals, which necessitates the creation of a top-down agenda. Meanwhile, competitions and challenges often only pose broad directions or “tracks” to facilitate a wide variety of solutions from participants who enter with their original ideas and plans to execute them. By contrast, talent investment programs often enable the highest degree of agenda-setting independence in participants, since their purpose is to support the personal and professional growth of these individuals.

Group/team vs. Network. Most R&D models support individual teams of researchers or developers that pursue different projects independently. Competitions and challenges, for instance, often take this “group/team” approach and encourage individual teams to create innovations that surpass their competitors. Meanwhile, some R&D models take a “network” approach by supporting a collective of individuals/teams and fostering a greater degree of collaboration, communication, and coordination among them. This approach can be beneficial in building new partnerships, establishing community standards for data, creating a body of knowledge, expanding a field’s overall capacity, or advancing new directions in a field. This approach is sometimes taken by the Institute of Education Sciences and the National Science Foundation. Examples include SEERNet and the LEARN Network. SEERNet is an IES-funded network of researchers and practitioners that work together to expand the capacity of digital learning platforms while the LEARN Network brought together five teams to address educational challenges brought on by the COVID-19 pandemic. The network approach is in the Learning Engineering Virtual Institute, which supports teams in reaching their shared goal of doubling middle school math progress in five years, by encouraging the sharing of progress, results, lessons learned, and expertise among teams.

Overview Of Models. What Makes Them Unique?

A variety of organizations and programs have a track record of fostering breakthrough innovations. Seeking to study and replicate their successes in education, this section will describe six models that foster breakthrough innovations and analyze their specific components for potential applications to K-12 education. Key components and examples of each model can be found in the Appendix. This list of models is not intended to be an exhaustive list, but rather a starting point for workshop conversations.

An Advanced Research Projects Agency (ARPA) was first created by the U.S. Department of Defense in 1958 in response to the Soviet Union’s Sputnik launch. Eventually called DARPA (Defense Advanced Research Projects Agency), its mission has been to “make pivotal investments in breakthrough technologies for national security.” DARPA’s high-risk, high-reward approach has been effective in generating early-stage, field-changing breakthroughs such as the Internet, GPS, artificial intelligence, and mRNA vaccines. Its success has led to the establishment of new ARPAs designed to solve challenges in sectors other than defense, and has cropped up outside of the U.S. DARPA and more nascent ARPAs have inspired the concept of a National Center for Advanced Development in Education (NCADE) at IES, which is starting to take shape as a pilot program called Accelerate, Transform, and Scale. Research outcomes from ARPAs usually have the potential to transform a field and often fall into the classification of Pasteur’s quadrant, which means that they seek a fundamental understanding of scientific problems, while also having immediate use for society. Additionally, taking a spiral development approach is critical in ARPAs, as it helps the research projects shed ineffective elements, adjust goals, and incorporate new capabilities. ARPAs have been more frequently applied to solve technology-based problems, and are also being applied to social and behavioral science. ARPAs are known for their flexible approach to program management, which shifts and adjusts the end goals of a project as research progresses or encounters obstacles.
Virtual Institutes (VIs) bring together a portfolio or cohort of teams to meet an overarching goal with their individual research. The cohort embodies diversity within the research team as well as the development team. The research team consists of both learning and computer scientists, while the development team is a mix of technologists, educators, and academics. Under such a multidisciplinary approach, this model is guided by diversity of thought, as participants virtually collaborate, share resources, and workshop solutions beyond geographical locations and national borders. VIs are commonly sponsored by government and philanthropic organizations, and fund teams of participants based on the scope of their proposals. This can be accomplished through funding mechanisms such as grants and cooperative agreements, a type of contract between a government entity and a business, and can pose additional terms and conditions that verify the performance and implementation of the proposals.
Focused Research Organizations (FROs) are special purpose organizations, inspired by startups, created solely to solve defined science and technology challenges that require complex coordinated teamwork over a many-year timescale. Examples include developing a breakthrough measurement technique for performing a large-scale, systematic study. FROs fill a significant gap – a lack of targeted and coordinated teamwork – in the current research ecosystem that is not incentivized in academia and cannot capture value sufficiently to incentivize for-profit investments. Though many FROs function as university-adjacent nonprofits or university-affiliated research centers (UARCs), they are not held to university restrictions or expectations. By operating independently of both traditional academic and commercial structures, FROs can take a “moonshot approach” to solving pressing problems, generally within a 5- to 10-year time frame. FROs function similarly to startups with a closely knit team of 10 to 30 people that is able to move quickly on a technical roadmap. Overall, FROs use organizational structures to solve problems in research and development that aren’t traditionally addressed in academia. Therefore, understanding FROs’ makeup – and what makes them successful – can help us think more deeply about the strengths and weaknesses of our research ecosystems, which can inspire additional infrastructures designed to address specific, systemic needs in education.
Public prize competitions and challenges drive innovation by presenting a given problem or target issue and providing incentives – frequently in the form of cash awards or grants – to attract participants. They give participants a specific timeframe to submit a proposal and use a set of criteria to evaluate the proposals and select winners. They may be hosted directly by a team of organizers, or with the help of online platforms such as Kaggle. Some competitions, like IES’ XPRIZE Digital Learning Competition and NAEP data competitions, are sponsored or managed by the federal government. Because competitions and challenges are a rapid and inclusive way of addressing critical issues, they have become a popular way of generating ideas, solutions, and innovations. There are many different types of challenges and competitions ranging in size and structure. They often share some of the same features but can differ in the distribution of risks between funders (who assume more risks in challenges) and investigators (who assume more risks in prize competitions). The model has grown in popularity because it embodies the “field building” approach – thinking about how an entire field can work together – which has proven to be effective in growing the capacity to achieve a common goal. Their focus on building communities, pooling talents, and encouraging growth makes them interesting case studies for the field of education. Education R&D needs more investment and attention than ever, and this “field-building” approach yields valuable lessons for education leaders seeking to build a solid foundation for R&D to thrive.
Talent investment programs provide funding and support to individuals who are considered exceptional in their field. Talent investment programs may be conceptualized narrowly to foster talent in particular areas or broad in scope, incorporating individuals from diverse backgrounds, careers, and topics of expertise. These programs are also frequently designated as fellowships and generally aim to increase the individual’s impact on society by providing robust funding and support for a set period of time. These programs often make a bet on an individual’s potential for impact versus a bet on a specific idea or innovation – and they are often structured to bring talented individuals from the private sector into public service, or deployed to solve pressing societal issues. Talent investment programs vary drastically in what sort of reporting or results they expect to receive from awardees during their fellowship periods. Some provide funding with limited to no requirements while others expect awardees to participate in set activities, give presentations, or create a report or product by the end of their fellowship. In contrast with the models above, talent investment programs share a unique focus on cultivating individual contributions.
Startup supports are designed to help founders bring their early-stage ideas or companies closer to their goals and vision. Since the 2000s, three major types of organizations – accelerators, incubators, and venture studios – have emerged and evolved. They provide the kind of resources that founders need the most, from capital to office space and a network of founders, investors, talent, and advisors. Also known as “parallel entrepreneurship,” these startup support models fill critical gaps in the process of bringing an innovative idea to life such as the need for capital, knowledge, talent, and early customers. The impact of connecting founders with these resources early on can be profound, as it fosters long-term collaboration and support that is difficult to organize on one’s own, especially as a first-time entrepreneur or innovator. Although this model is relatively young, these startup support ecosystems have grown some of the most impactful companies today including Reddit, Dropbox, Airbnb, and Moderna. Depending on their specific needs, founders or startups may be more suited for one type of startup support model than the others, which pose different requirements, fees, and competitiveness for entry. Some models may specialize in one industry or discipline, while others support ideas across fields. Similarly, leaders may design programs with different elements from these three models, to best foster the kind of innovations they would like to see – at different stages of development, with varying experience levels and end goals.

The Broader Innovation Ecosystem

While specific structures promote breakthrough technologies, these models are just one part of a larger ecosystem that cultivates transformative ideas and accelerates their impact. Industry’s ecosystems vary considerably due to regulatory requirements, funding structures, and market demands, which are not always generative of breakthrough innovation. These challenges are often compounded by the relatively low profitability of the education sector which limits funding for R&D. For this reason, additional support and a robust innovation ecosystem are necessary components for the generation of transformative breakthroughs.

This section discusses the broader ecosystem necessary to foster breakthrough ideas and then turn those ideas into actionable, impactful innovations in the education sector and beyond. In thinking through this innovation ecosystem, we imagine a pipeline that combines talent, support, and funding, which is then followed by an incubation and development phase supported by big bet models. After this incubation and development period, breakthrough innovations then need to implement, scale, evaluate, and integrate with pre-existing systems in order to achieve successful reach and adoption. Layered, multi-phased support is fundamental in creating meaningful innovations that positively change the ways we live, work, and learn.

Building a Pipeline For Talent and Innovative Ideas

It takes time, talent, and resources to generate ideas for breakthrough innovations. However, connecting the potential creators of breakthrough ideas with the necessary support to develop and actualize those ideas requires a multipronged approach that both recruits and cultivates diverse thinkers. In thinking through the structures that support the early-stage development of breakthrough idea pipelines, innovators benefit from:

Support for emerging talent. Creating new solutions and ideas requires unique and different approaches. By bringing in out-of-the-box perspectives and fostering opportunities for collaboration across sectors and disciplines, programs conducive to breakthrough thinking often welcome new innovators and provide them with support ranging from networking opportunities to funding. Within the fields of education and beyond, there are many programs that take on the challenge of recruiting new, diverse voices and providing them with fundamental support, with the ultimate goal of generating the next major breakthrough.
Identifying problems worth solving. In the early stages of a program, it is important for leadership or teams to identify the problems that they aim to solve – from the high-level headline to a defined structure and timeframe for meeting smaller goals and objectives that will lead to eventual breakthroughs. Program managers can help their team break down large problems, develop working theories around potential solutions, and create short-term or long-term benchmarks within the project scope and sequence. Many ARPA models approach this aspect with a flexible program management framework, which allows them to adjust these broader research goals and their internal structures based on real-time progress or obstacles.
Advocacy and storytelling. To support an ecosystem that fosters breakthrough innovations, it is important to have advocates calling for sustained investment – from government, philanthropies, and the private sector – in the models and approaches known to spark and scale innovation. As an example, the Alliance for Learning Innovation advocates for better education R&D infrastructure, including the establishment of an ARPA for education. Effective advocacy would not be possible without compelling success stories. Consider, for instance, that ARPAs would not have proliferated in non-defense sectors and outside of the U.S. if not for the lore around DARPA’s origin story and the agency’s groundbreaking achievements.
Seed funding. Seed funding allows teams to get started without being beholden to investors who would expect a return on their investment. Receiving this sort of funding also indicates to breakthrough thinkers that some party or institution sees potential in the idea, which is encouraging in and of itself. Additionally, potential developers of breakthrough ideas benefit from accessible seed funding with approachable applications that notify applicants of awards in a timely manner. A quick turnaround with an approachable application also allows innovators to address time-sensitive issues and needs. The Small Business Innovation Research (SBIR) programs at NSF and IES and NSF offer examples of seed funding.

How Can Smaller Seed Funding Kickstart Developments?

While most models designed to spur breakthrough innovations operate on a relatively robust budget, some models are intended to provide smaller amounts of seed funding – less than or around $10,000 – to accelerate each research project.

With smaller seed funding, these projects target highly specific problems – in diverse fields and disciplines – on a small scale. For instance, they may aim to solve challenges in a local community, improve existing societal systems and processes, or contribute to a new and growing field. The modest amount of funding often serves to finance expenses of innovative projects at their earliest phases – in equipment, publications, survey administration, and outcome assessment.

While this approach can take shape as challenges, competitions, or crowdsourced campaigns, it can also be tailored to spark innovations from groups that are not typically funded by the existing system. For instance, Experiment Foundation supports scientists at various stages of their careers with a “science angel model” – specifically through granting highly experienced and pioneering scientists more discretion in their research, or encouraging student and early-career scientists to start working on their best ideas. Under the smaller seed funding approach, many models like these aim to establish new funding dynamics that yield innovations that may not otherwise occur within the current funding norms.

This model allows funders to kickstart a greater number and wider variety of promising research projects with a fixed amount of available funding. While these research projects start small, many have been able to leverage the seed funding to grow into highly impactful and transformative innovations over time. The wide distribution of funding also nurtures socially driven or local community-oriented projects that may be traditionally underfunded. Furthermore, platforms such as Experiment allow these small-scale research projects to more directly benefit from funders – and not in exchange for a significant share of the funding or equity – compared to other models.

Opportunities to build knowledge and credibility. Beyond funding and programs that welcome new perspectives, thinkers benefit from opportunities for mutual development with fellow innovators in order to foster credibility and build knowledge. In many cases, this is done via a collaborative, cross-disciplinary setting where the stakes for sharing ideas and creating connections are lowered so that innovation can flow and develop. This allows individuals and teams to work together without success or failure resting on one entity. Cross-sector knowledge-building, which can take place via rotator programs and fellowships, is an integral part of fostering “diffusion” within an innovation ecosystem. Diffusion occurs when advances by one innovator spark further exploration and imitation by the wider innovator community. With greater diffusion comes a more connected and productive innovation ecosystem.
Time and support to cultivate ideas. In order to fully develop breakthroughs, innovators need time and support to properly cultivate their ideas, including building evidence that supports the fundamental need for their particular breakthrough. By giving innovators support and time, big thinkers can separate themselves from market demands and the status quo while gathering the evidence they need to eventually launch their innovation. This allows innovators to go beyond the current constraints of their fields and think creatively, which further fosters the development of their innovations.

Example Programs Supporting the Innovation Pipeline

The following programs and models are just some examples of how innovators can receive necessary knowledge-building opportunities, talent development, and/or seed funding. Some of these programs include various types of support while others provide support more narrowly. What they have in common is the ultimate goal of creating innovations and developments that foster actionable change.

National Science Foundation Research Coordination Networks: A variety of National Science Foundation (NSF) directorates create funding opportunities via Research Coordination Networks (RCNs). Rather than create primary research, RCNs seek to bring together experts across diverse fields, geographic locations, and sectors to generate innovative ideas or drive fields in new directions. Within education research and development, the Mid-Scale Research Incubators and Conferences program is one such RCN opportunity. The program allows K-12 education researchers to collaborate with experts across disciplines as they explore ideas. The Incubator and Conference program has the eventual goal of preparing investigators to apply for funding in the Mid-Scale Research Infrastructure program.
Data Science Competitions: Data Science Competitions allow rising talent to gain access and exposure to some of the world’s toughest issues while providing space and support for collaboration and growth. For example, competitions on Kaggle, a crowdsourcing data science competition platform, seek to solve issues in health, education, housing, technology, and more. These competitions have real-world outcomes but also provide a place to learn about data science and machine learning. They develop talent while cultivating their potential to solve specific problems and provide tangible societal benefits. Competitors win by creating the highest-performing algorithms based on competition-specific metrics. Winners receive a monetary prize of various sizes, as well as significant recognition within the data science community.
Learning Engineering Tools Competition: The Tools Competition is a global edtech competition that provides seed funding and support for teams developing tools to address urgent learning needs. The program has various funding levels for applicants, including Catalyst awards that specifically target new competitors. These awards give $50,000 of seed funding to new innovators, including students and teachers, who need an initial influx of monetary support to develop their ideas from different stages, whether that be leading pilot studies or taking a prototype to scale. Beyond cash prizes, competitors also receive additional support from the Tools Competition community.
Fellowships: Fellowship programs identify, connect, and support top talent in various fields to enhance their leadership and increase their exposure to opportunities for growth. A fellowship program can be issue-specific to develop leadership in a particular field. For example, the Pahara Fellowship seeks out and supports diverse, exceptional talent in public education. Meanwhile, some fellowships, like Echoing Green, are agnostic about issue areas and instead focus on cultivating talent to support a broader mission. Echoing Green’s fellowship aims to solve the world’s greatest challenges through transformational leadership. Other fellowships seek to engage talents with specific fields – the FAS Impact Fellowship, for instance, provides short-term placements for science and technology experts in the federal government.

Structures to Realize the Potential of New Innovations

Once a breakthrough idea or technology has been conceptualized via a supportive pipeline, that potential breakthrough is at a delicate point. Without structures to then advance the development of the breakthrough, it is at risk of losing momentum. Breakthrough technologies, by definition, often work outside the conventions for their respective industries and sectors. As a result, adoption, scaling, and measuring impact require additional time and support. Additionally, those with the expertise that created a breakthrough technology aren’t always equipped with the necessary expertise and capacity to scale and implement it, meaning innovators require additional collaborators from a variety of disciplines.

In considering the successful implementation, scaling, and evaluation of a breakthrough innovation, the following components are crucial for long-term, quantitatively-proven success:

Systems that are co-designed with end users in mind. A recurring characteristic of breakthrough innovations is that they create a paradigm shift, which can be concerning for people working in the sectors that will be changed. The increased use of artificial intelligence is a salient example. Advances like AI tutors provoke concerns in some educators who believe the new technology will impact their jobs and in some parents who are worried about their child not getting instruction from their teacher. However, implementing these breakthroughs alongside the impacted users can ensure a successful implementation that centers the target populations’ needs.
Strategies that consider integration into current systems. Innovators and their collaborators must consider how the innovation will fit into pre-existing systems or how these current systems can shift to accommodate the innovation. This requires innovators to find collaborators and expert counterparts who are familiar with the systems to better strategize for a smooth and thoughtful integration of breakthrough technologies. In addition, it is a continuous and important process to find space, attention, and support for innovations within an organization, which usually takes preemptive planning of an organizational structure that allows for and even encourages productive disruptions.
Strategic, sustainable scaling. Marketing breakthroughs, recruiting users, and implementing new technologies at scale also require expertise beyond that of the original innovators. The invention of the Internet, which was spearheaded by DARPA, is an excellent example of what it takes to successfully scale a breakthrough. Apart from the original developers, the user-friendly interface of the World Wide Web and its potential to enhance commercialization enabled industry partners to support and adopt the idea quickly. Scaling supports vary significantly for different programs. Some are embedded in a government entity with procurement authority, while others are supported by internal systems for launching and commercializing innovations. DARPA, for instance, is embedded in the U.S. Department of Defense, while ARPA-E equips every research team with a tech-to-market team with commercialization experience.
Effective mechanisms for evaluation. Beyond the number of users, new metrics of success may be required to effectively measure the impact of new breakthroughs. Returning to the example of the Internet, the new technology needed to evaluate its success in terms of reliability, accuracy, and ability to successfully integrate with pre-existing technologies.

The field of education can benefit tremendously from such robust implementation of a breakthrough technology with the same reach and influence of GPS whose invention and successful scaling changed every fiber of our navigation system. It fulfilled a universal need with great nuances, but not without the solid groundwork of previous navigation inventions, continuous experimentation and improvement of its quality and accuracy, and thoughtful implementation for both military and civilian users. Similarly, education researchers are paving the road for such breakthrough innovations to occur, and a successful outcome depends on similar steps.

In considering how an education-related breakthrough could achieve the same reach and transformative power as something like GPS, there are several nuances to keep in mind. First, it is crucial to continue investing in education R&D that supports concepts with the potential for impactful evolution. The U.S. education system is complex and fragmented with structures differing from state to state and district to district. A commitment to R&D will encourage targeted and effective implementation with the target populations in mind. Meanwhile, partnerships with stakeholders in and beyond education can lead to more deliberate strategies in scaling and integrating these education breakthroughs into school districts and classrooms – with varying capabilities and access to resources. Guided by informed strategies to integrate the innovations, their design, refinement, and implementation should center their users – teachers, students, and their families. In implementing breakthroughs in education, the components above will serve to function in a way that supplements and enhances one another, thus capturing the efficacy and scaling to the full potential of breakthrough innovations, thereby leading to innovations that change how we teach, learn, and assess.

Conclusion

With the right environment and conditions, innovations can be developed, refined, and scaled to impact a field, discipline, or community to an unprecedented extent, and this vision is only within reach if leaders take a highly targeted and purposeful approach.

To sustainably harvest and nurture breakthrough innovations, leaders must approach key decisions in creating an R&D infrastructure with great clarity, especially when it comes to questions in risk-taking, goal-setting, investment, success criteria, paths to transition to scale, and barriers to entry. That is what has led to the success of the models discussed above – a clear and thorough understanding of the breakthrough innovations themselves, the context of their use and applications, and the roles they assume in the broader innovation ecosystem at large.

The analysis above makes clear the importance of gearing processes and architectures for the ideation, development, and implementation of breakthrough innovations. Based on research and communication with experts in innovations across fields, The Learning Agency has identified the following recommendations that can both spur and sustain breakthrough innovations in education:

Education leaders should design research and development models with clearly defined goals for outcomes and impact. The meaningful design, effective development, and fruitful implementation of R&D infrastructures can often be attributed to a clear understanding of their intended impact and outcomes. Successful infrastructures have been instrumental in generating breakthrough innovations that address pressing or significant problems in certain fields and disciplines, but it can be challenging to create them for that exact reason – their design and implementation are tremendously intentional. That’s why it is critical for education leaders to carefully evaluate what they hope to achieve with the infrastructures, how they will measure their impact, and how they will scale innovations for the level of use, adoption, and accessibility that they need.

Prior to designing infrastructures for innovation, it is important to evaluate the broader innovation ecosystem and consider the role and position of the new infrastructure. The innovation ecosystem operates with various components, each with its specific role and function that contribute to the successful generation and application of breakthrough ideas. For example, a pipeline for talent and innovative ideas is crucial to the cultivation and support of diverse, multidisciplinary talents whose collaboration and combined potential will spark early ideation of breakthrough innovations. With a reliable talent pipeline, the innovation ecosystem then benefits from structures that help to continue the early momentum of these innovative ideas. They seek to realize the full potential of innovative ideas through expertise support, user engagement, market assessment, and business development.

In implementing infrastructures for innovation, education leaders should consider the decision points listed above and select compatible components that serve the specific needs of their communities. There is no doubt that education leaders designing infrastructures for different communities will have access to vastly different types or levels of resources. They are also building these infrastructures to meet specific needs, gaps, and objectives that vary from community to community. Therefore, it is crucial for leaders to consider the types of risk, problem, investment, and engagement that they anticipate or intend to cultivate.

Acknowledgments

The authors of this brief would like to thank several individuals who contributed insights and feedback. Joanna Cannon and Bridget Cherry of the Walton Family Foundation provided valuable guidance and feedback throughout the writing process.

The following experts in breakthrough innovations reviewed an early version of the brief and made it better through their contributions: Anastasia Gamick, Convergent Research; Dan Jarrett, Schmidt Futures; Russ Shilling, Russ Shilling Consulting; Sara Schapiro, Federation of American Scientists; Steve Buchsbaum, Bermuda Associates; and Youssef Kalad, AlleyCorp. Thanks to David Lang, Executive Director of Experiment Foundation for his contributions to the inset on seed funding.

Finally, we are grateful to all Breakthrough Innovations Workshop participants who shared their expertise and insights so generously with us. Your thoughtful engagement in the workshop enhanced the content of this brief.

Appendix

This Appendix outlines each model’s critical components and examples. The “critical components” section includes attributes that make each model unique. The “examples” section outlines existing applications of each model, across sectors, industries, and countries. This section serves to provide a more in-depth look into the structure and application of each model discussed above.

Advanced Research Projects Agencies

Advanced Research Projects Agencies (ARPAs) are guided by a high-risk, high-reward approach. Modeled after the Defense Advanced Research Projects Agency (DARPA), they operate independently and execute research and development projects that spur breakthrough innovations in fields and disciplines. Their successes have led to a growing body of work that explores their unique designs and effective outcomes.

Critical Components:

Clear objective. A specific mission must drive the research an ARPA conducts, and under the umbrella of the mission, there must be a continuous effort to explore, articulate, and break the mission into subsets of problems whose timelines and benchmarks will be adjusted based on the progress or setbacks in research. This clear yet flexible structure of the program’s end goal must also be informed by the context of broader trends and opportunities in scientific investments.
Interdisciplinary collaboration. ARPAs invest in “moonshots” and create interdisciplinary programs with a diverse group of experts to reach their ambitious goals in 3 to 5 years. Most ARPAs use a version of the Heilmeier Catechism, a set of clear, straightforward questions to decide which risks to take. With their high risks come high rewards – ARPA projects are best applied to problems that are addressable but intractable, so breakthroughs can greatly accelerate the field. Nonetheless, the moonshot approach can contribute to groundbreaking innovations without a clear pathway for deployment, so it is important to direct teams to engage with end users and identify possible adopters.
Organizational independence. ARPAs operate independently from other branches within its affiliated federal agency and ARPA directors report to top leadership (e.g., the ARPA-E Director reports to the Secretary of Energy). With flexible hiring and contracting, they may hire outside of the federal government’s standard hiring requirements and bring on top talent quickly and with more competitive pay. ARPAs can use grants, contracts, and cooperative agreements to fund performers, but they also have “other transactional authority” (OTA) when needed (e.g., technology investment agreements with specific terms for intellectual property).
Focus on talent. Program managers are top talent from academia, industry, or government who come to an ARPA to serve an ambitious 3- to 5-year term. The limited term creates a sense of urgency and an outcome-driven approach to managing research projects and often serves as a launching pad for their future roles and opportunities. They have ample discretion in how they design their programs and allocate their funds. Program managers are actively involved with researchers for each project, receiving regular updates, providing feedback throughout the process, and making adjustments to milestones, the budget, and timelines as needed. The success of ARPA models has been largely attributed to this mechanism, as the position of program managers combines the control over research funds with the responsibility for research design. Their key influence and discretion in ARPAs necessitate a diverse and rigorous selection process.
Flexible program management. ARPAs take a flexible approach to program management. They lead a continuous process of setting appropriate goals for specific programs, adjusting the goals based on the progress and obstacles in research, and collaborating with research teams on meeting those shifting goals to maximize outcomes.

Examples:

Defense Advanced Research Project Agency (DARPA). The original ARPA, founded in 1958, strives for transformational change in the interest of national security. DARPA consists of about 220 government employees, including almost 100 program managers overseeing about 250 R&D programs. DARPA is behind some of the most significant innovations of our time including the Internet, GPS, speech recognition, and mRNA vaccines. It is important to note that traditional R&D models take the “pipeline approach” where research is produced and then entered into the innovation pipeline, awaiting opportunities that may arise. On the other hand, DARPA takes a unique “right-to-left” approach where it first identifies the needs from the innovation pipeline, and then initiates research efforts to meet those exact needs.
Advanced Research Project Agency-Energy (ARPA-E). Launched in 2007 to advance “high-potential, high-impact energy technologies that are too early for private sector investment.” Since its inception, ARPA-E has provided nearly $3.5 billion in R&D funding for over 1,500 energy technology projects. This has led to the formation of 135 new companies, 321 licenses for ARPA-E, and over $11.4 billion raised in private-sector follow-on funding. ARPA-E projects have resulted in over 6,000 peer-reviewed journal articles and over 1,000 patents.
Germany’s Federal Agency for Disruptive Innovation (SPRIN-D). Established in 2019, SPRIN-D’s goal is to “create new disruptive innovations from Germany” that solve the “social, ecological, and economic challenges of our time.” In the last few years, it supported the development of a number of projects including a supercomputer that simulates the human brain. Currently, the agency is also supporting projects that aspire to cure viral disease, revolutionize cancer and Alzheimer’s treatment, boost renewable energy, and more.
Japan’s Moonshot Research & Development Program. Established in 2020, this ARPA comes from Japan’s Science and Technology Agency and serves to fulfill moonshot goals. Within the last 2-3 years, the Moonshot R&D program has already supported groundbreaking research on topics including the use of oxytocin to prevent obesity, the development of AI in extracting useful knowledge and removing defects in quantum computers.
Advanced Education Research and Development Fund (AERDF). Launched in 2021, AERDF is a national nonprofit, founded by philanthropy partners, that uses key elements of the ARPA model to support “ambitious, inclusive” 3- to 5-year education R&D programs. AERDF focuses on solving education challenges that disproportionately impact PK-12 Black and Latino students, and all PK-12 students experiencing poverty.

Virtual Institutes

Virtual Institutes facilitate collaboration among a cohort of experts from a variety of fields, disciplines, and geographical locations. Usually funded by government and philanthropic organizations, VIs support multidisciplinary teams in executing their proposals and ideating solutions that could transform a field or discipline at large.

Critical components:

Clear objective. VIs are usually set up to support transformative advances in some of the most complex or important problems in a field. Their fundamental purpose is often to create large-scale, revolutionary impact, which means that VIs usually have clear directions in identifying research projects with promising topics and measuring the quality of their results. The latter is especially important for VIs, with many of them leading strong annual evaluations that take into account both the short-term evidence and the long-term potential of research projects. Many VIs are established with a specific theme or field of study, which allows them to concentrate technical expertise and financial resources on addressing a particular research need.
Interdisciplinary collaboration. Collaboration is at the core of VIs. Instead of competing to provide the “best” solution, interdisciplinary teams work together to tackle the problem from multiple fronts via interconnected projects. The VI model facilitates information exchange and idea generation from experts of different disciplines that are not usually facilitated in a conventional academic setting.
Driven to serve the public good. The VI model is not driven by profits or commercialization. Instead, it is designed to generate findings that can be shared globally for the greater good and benefit humanity in the long run. This is why many existing VI programs implement a data-sharing policy that allows interested parties to learn from and build on the results.
Time-bound. VIs often provide research teams with funding that spans multiple years. They offer awards of various durations through funding mechanisms such as grants, contracts, and cooperative agreements. Typically, funding is multi-year, in order to provide stability for student researchers who may be involved as well as flexibility to respond to any change in the high-level direction of the program or organization. While funding opportunities are limited to a fixed timeframe, some VIs themselves are time-bound as well and conclude at the end of a designated period.

Examples:

Learning Engineering Virtual Institute (LEVI). In 2022, LEVI was launched to tackle a widespread challenge in K-12 education: middle school math performance. Student proficiency in math has declined in the past decade and significantly worsened throughout the pandemic, yet math performance is critical to access to a living wage. To double the rate of middle school math progress, LEVI supports seven teams with compelling proposals that embody the spirit of collaboration and experimentation across disciplines. Like most other Virtual Institutes, LEVI is guided by a topic that is best approached through working across institutions and disciplines, versus centralizing research in one place. LEVI is a five-year program that will conclude in 2027.
Virtual Institute for Responsible Innovation. Other entities have launched VIs. For instance, Arizona State University launched its own virtual institute – Virtual Institute for Responsible Innovation – that existed from 2013 to 2016. It received NSF funding through the now-dormant Science Across Virtual Institutes program and brought together participants from eight countries.
EC2U, which is an alliance of seven European universities, also has several virtual institutes, though these VIs pull upon a pool of participants that is limited to the participating universities.

Focused Research Organizations

Focused Research Organizations (FROs) take a highly targeted and coordinated approach to addressing specific problems in a field or discipline. They bring together interdisciplinary teams in a fast-paced and outcomes-driven environment to spark innovative solutions that go on to serve as a “force multiplier” for the research community at large.

Critical components:

Clear objective. FROs are designed to build a tool, system, or dataset that is designed to transform a key area of R&D. While FROs may use some preexisting techniques or technologies, the goal is to develop something new that exponentially accelerates progress, reduces cost, or increases the reliability of scientific study. Therefore, they focus on new technology or developments with the potential to dramatically elevate subsequent work in the field. This “technical goal”-driven approach is a distinction from open-ended research.
Interdisciplinary collaboration. Focused Research Organizations are designed to address a need in the science ecosystem by taking on projects that require a high level of coordination and system-building. They involve interdisciplinary teams of 10-30 tackling projects that are challenging to accomplish in academia. FROs address a critical gap in the research community, given that academic structures award researchers based on individual merit, disincentivizing systematic teamwork across disciplines over time. They take a more focused approach than in academia by seeking to address narrowly defined and quantitatively tracked problems within a finite period – usually around five years.
Organizational independence. FROs greatly resemble startups. They are led by a full-time founding team – including a CEO/CTO – and staffed by a dedicated team of expert professionals from academia and industry. Similar to that of a tech startup, the culture is typically fast-paced, results-driven, and self-directed. To ensure consistent progression, FROs typically drive their work with quantifiable metrics such as objectives and key results (OKRs). However, unlike startups, the FRO is not held to shareholder expectations and does not have to create a marketable product.
Driven to serve the public good. FROs are driven to create impactful outcomes that catalyze the development of science and technology. With a focus on the greater good, FROs disseminate their products in a way that benefits the research community and the public at large – including through open-sourcing data, creating nonprofits, and furthering fruitful partnerships.

Examples:

Convergent Research. Convergent Research supports FROs from the conception and proposal phase to funding and ultimately operating the FRO. FROs in Convergent Research’s portfolio include E11 Bio, which creates tools to understand the human brain and accelerate future neuroscience innovations including treatments for brain disorders and brain-inspired computing. Another FRO supported by Convergent Research is Cultivarium, which creates open-source tools for life scientists to study microorganisms and develop innovative biotechnologies. EvE Bio is on a 5-year mission to create a first draft of a map of the pharmome – a public domain knowledge map of FDA-approved drugs’ functional protein binding partners across thousands of human gene products. PTI is scaling up proteomics to illuminate the link between protein patterns, functional interactions, and disease.
Equity Accelerator. While FROs have the highest level of representation in the hard sciences, Equity Accelerator is a first-of-its-kind FRO with a base in the social and behavioral sciences that seeks to address learning and working environments. Guided by its mission to create more equitable learning environments for all students, Equity Accelerator develops affordable, scalable tools that directly address institutional barriers to inclusive success and has reached over 110,000 students at more than 350 colleges and universities.
The USC Institute for Creative Technologies. The USC Institute for Creative Technologies (ICT) is a Department of Defense (DoD) University Affiliated Research Center (UARC), sponsored by the U.S. Army. Harnessing Hollywood-derived creativity with academic innovation and military-domain expertise, ICT conducts award-winning R&D in AI, computer graphics, geospatial sciences, human performance, learning sciences, modeling, simulation & gaming, mixed reality (MxR), medical VR, narrative, and virtual humans.

Competitions/Challenges

Public prize competitions and challenges incentivize participants to propose promising solutions to a clear problem. They are driven by a “field-building” approach to bring together teams of talents with various experience levels and strengthen an enduring network of researchers and developers to test and refine their solutions throughout and beyond the competitions/challenges.

Critical components:

Clear, compelling objective. While competitions and challenges may feature participants addressing a given problem from a variety of perspectives, participants all work toward one or more common goals or to address a particular problem within a short but reasonable timeframe. The existence of defined problems – along with a set of fixed benchmarks – can pose constraints to the kind of innovations that competitors create. The nature of competitions and challenges also necessitates the creation of simple yet effective success criteria, which can be difficult to narrow down. It is important to note that for challenges and competitions, the clear and overarching objective often guides the creation of multiple subsets of objectives. With an effective strategy, solutions to these subsets of objectives can often be assembled to answer the greater question at hand.
Inclusivity. A collaborative field of innovators can drive true change. Many competitions and challenges have low barriers to entry – in terms of cost, experience, and disciplines. This allows the organizers to pool a wide variety of talents and resources to solve problems. Participants fill in different pieces of the “puzzle” and approach the issue from diverse angles, thus growing into a technical network that will accelerate problem-solving in the field in the long run. Nonetheless, organizers may choose to design a challenge or competition in a way that forces unlikely and sometimes uncomfortable collaborations to be competitive for selection.
Driven to serve the public good. With inclusivity at its core, effective competitions and challenges offer diagnostic feedback to help participants improve their solutions and grow as innovators. This is driven by the desire to stimulate learning of the research community at large. They often implement mechanisms like data- or solution-sharing that enable participants to start on “last year’s model” and build on previous progress (however, data sharing can be complicated in education, as they often concern the data privacy of children). Many passionate participants who were not selected as winners often continue to refine their product or find another source of funding, so the impact of the diagnostic feedback extends well beyond the current competition. Ultimately, all innovations sparked by competitions or challenges, whether they are selected or not, may eventually contribute to solving an important problem.

Examples:

Grand Challenges in Global Health. Initially launched by the Bill & Melinda Gates Foundation, Global Grand Challenges in Global Health works to provide innovative solutions for 14 major scientific challenges including creating single-dose vaccines for after-birth use, immunological methods to cure chronic infections, and chemical strategies to deplete disease-transmitting insect populations. It is part of a family of Grand Challenges initiatives that have awarded over 3,600 grants to investigators in over 110 countries against more than 100 follow-on challenges at times building on but usually distinct from the original 14 Challenges launched in 2003. Ultimately, they support global teams in solving global health and developmental problems, while innovating with emerging technologies across fields including AI and Large Language Models.
Learning Engineering Tools Competition. Tools Competition is a global prize competition for edtech innovators with the goal of creating and implementing technological solutions for issues in education. Participants design solutions for all types of learners – from early childhood to postsecondary education – and address specific target areas with their tools. In addition to monetary prizes ranging from $50,000 to $250,00, participants also have opportunities to network and receive feedback from expert judges. This competition is helpful to identify new entrants to the education R&D field and accelerate their growth. For instance, Rising Academies Network, one of the fastest-growing education companies in Africa, was named a winner from the first cycle of the competition to build an AI chatbot that helps deliver personalized audio content to learners, covering various K-12 topics, via mobile phones. Given the success of the tool they developed from the prize, they were able to apply to join the inaugural cohort of the Learning Engineering Virtual Institute (LEVI).
VITAL Prize Challenge. The Visionary Interdisciplinary Teams Advancing Learning (VITAL) Prize Challenge, which was created in a partnership between the National Science Foundation (NSF) and three philanthropic organizations, provides funding, mentorship, and support to interdisciplinary teams creating K-12 edtech solutions. Teams selected for the Discovery Round engage in a 7-week immersive training from the NSF Innovation Corps program for entrepreneurial education and mentorship. Those who advance to the next Semi-Final Round are paired with an “educator mentor” who provides valuable feedback on the feasibility of the edtech product. In this progressive, down-selecting challenge, teams have the opportunity to receive up to $250,000.
NASA’s Centennial Challenges. Initiated in 2005, NASA’s Centennial Challenges seeks innovative solutions from traditional and non-traditional sources that address problems of interest to NASA and the United States – from basic and applied research to technology development and prototype demonstration. Driven by the mission to directly engage the public in advanced technology development, Centennial Challenges offers prizes to independent inventors including students and small businesses. Recently, its Watts on the Moon Challenge – which seeks solutions to problems in energy storage, management, and distribution for space flights and on the lunar surface – awarded four winning teams $400,000 each to advance to the next phase of the competition and further develop and test their prototypes.

Talent Investments

Rather than facilitating collaborations to solve clearly defined problems, talent investment programs zoom in on the potential of individuals. They support exceptional individuals in formulating and realizing their own pursuits – with funding, resources, and connections to valuable networks – with the goal of benefiting society at large with their outcomes.

Critical Components:

Open-ended objective. Talent investment programs generally do not have a specific problem in mind, but rather support the individual ideas and passions of the fellows. The goal is to support extraordinary thinkers in their own pursuits, small- or large-scale, instead of having a narrow issue area. Some programs do have a relatively broad focus on a certain field such as healthcare or technology, with the goal of advancing specific fields or industries.
Focus on talent. While occasionally talent investment programs support teams, most center around fostering individual people with extraordinary talents. Instead of giving money to an organization or entity, talent investments target individual changemakers who are identified as having high potential. These individuals receive significant recognition and support, from finance to professional development opportunities and support networks.
Time-bound. Talent investment programs generally accept a group or class of fellows as a cohort that repeats with regular frequency. During the usual year-long fellowship, fellows focus on their projects as part of a cohort and a wider community that many continue to be active in after the program.
Monetary and other support. Most talent investment programs provide a mix of financial support for the duration of the fellowship and other professional development, such as mentorship, resources to create projects, and connections to other support networks. Being selected for one of the major talent investment programs also comes with significant recognition at the national and global levels.
Application or nomination process. Many fellowship programs are nomination-only and depend on networks of nominators and evaluators to select fellows. Other talent investment programs include an application process where an individual can present themselves for the fellowship, but these are still rigorously considered and the programs are highly competitive.

Examples:

MacArthur Fellows. Colloquially referred to as “Genius Grants,” the MacArthur Fellows program provides unrestricted awards of $800,000 to extraordinary talent from any field. The program is nomination-only and prioritizes creativity, a history of impressive accomplishments, and future potential.
FAS Impact Fellowship. The FAS Impact Fellowship provides a pathway for diverse scientific and technological experts to participate in an impactful, short-term “tour of service” in the federal government. Impact Fellows help ensure that science and technology are inextricably linked with policymaking as our nation confronts existential challenges and pursues ambitious opportunities. From education to clean energy, immigration, wildfire resilience, national security, and fair housing, Impact Fellows are serving across a variety of federal agencies in roles that augment existing government teams as they confront some of the greatest scientific and social challenges of our time.
Radcliffe Fellowship. Based at Harvard University, the Radcliffe Fellowship program provides housing, a $78,000 stipend, and professional support for fellows to live on campus and work on a specific project. Fellows apply directly as individuals or small teams working on the same project and can come from a variety of careers and disciplines. Fellows are expected to be active participants on campus, which includes presenting their work.
Health Innovators Fellowship. In 2015, the Aspen Institute launched the Health Innovators Fellowship in partnership with Prisma Health. It is designed to provide support and learning opportunities for senior executives and leaders in reflecting on and addressing challenges in the U.S. healthcare system. Fellows must be nominated by a third party and are expected to design and launch a venture of their choosing based on their learning experiences during the fellowship.

Startup Supports

Startup supports were created to serve the founders’ needs. Designed to assist teams at different stages of development, the three main types of startup supports discussed below have varying degrees of focus on networking, product development, or company growth. They help entrepreneurs meet critical gaps in bringing their ideas to life, whether in capital, expertise, or community support.

Models:

Venture studios are also known as “idea factories.” They source ideas with high potential, directly help build a marketable Minimum Viable Product (MVP), and then recruit founders to run and scale these products. Venture studios are unique in the way that their internal teams are responsible for the bulk of the development process, from testing the product with customers, and assessing the product-market fit, to ultimately, identifying a business model that works for scaling the idea at hand. Unlike accelerators, they don’t operate on a fixed timeframe and in the process of product testing and product-market fit assessment, kill many of the ideas that lack the potential for profit generation or high growth. A survey shows that venture studios can ask for 30% to 80% of equity from the startups. As of July 2022, there were more than 780 venture studios around the world. While some operate independently, many others work with government or industry partners to source ideas or intellectual property.
Incubators nurture innovative ideas or earlier-stage startups. Typically, startups accepted into an incubator program relocate to a physical location and co-work in a shared space with other startups throughout the startup process – from revising the ideas, to drafting a business plan, leading market analysis, and navigating tax and legal issues in launching a startup. Incubators can operate independently or be sponsored by or housed within venture capital firms, angel investors, corporations, and universities. Some may have an open application process, while others accept ideas from within their networks. Unlike accelerators, most of them don’t charge fees upfront except for the coworking space or staff’s advice.
Accelerators are more developed startups than incubators. Their focus is on rapid growth, rather than product development. To do this, accelerators provide a cohort of startups with fully immersive bootcamps over a period of 6 to 12 weeks, in exchange for a share of future equity in the companies. They typically seek founders with an existing team and company, and the expertise to grow their startups beyond the bootcamps. These accelerator bootcamps mostly serve to plug founders into a network of other driven founders and seasoned startup advisors, and conclude with a “demo day” where founders pitch their ideas in front of venture capitalists, angel investors, and members of the media. They are highly selective and have a low acceptance rate.

Examples:

Venture Studios

America’s Frontier Fund. Founded in 2021, America’s Frontier Fund is an investment platform driven by the mission of “reinvigorating our nation’s innovation and manufacturing prowess in critical frontier technology sectors.” Its team consists of scientists, founders, investors, policy and national security experts who are committed to the economic competitiveness and national security of the U.S. It partners with companies and governments to research and invest capital in frontier technologies, building companies and shared platforms that support their development, and bolster regional tech ecosystems. In 2022, New Mexico pledged $100 million to back its first fund to companies in industries that are “strategically important” to the U.S.
Flagship Pioneering. Founded in 1999, Flagship Pioneering is an independent venture studio and venture capital company based in Cambridge, Massachusetts that focuses on building and investing in “life sciences companies that invent breakthrough technologies to transform healthcare and sustainability.” It has a portfolio of successfully exited and current companies in human health, technologies, and sustainability. In late 2010, Flagship Pioneering formed LS18, the company that became Moderna.
Rocket Internet. Founded in 2007, the Berlin-based Rocket Internet incubates, invests in, and provides operational support to internet and technology companies across the globe. Its active portfolio includes more than 200 companies with over 42,000 employees and a total valuation of over €30 billion. In 2008, Rocket Internet founded Zalando, which emulates the business model of American online retailer Zappos and has become one of Europe’s largest online retail stores. It is an example of a venture studio that is “industry-agnostic,” meaning that it doesn’t specialize in a specific industry and applies its expertise across sectors.

Incubators

Idealab. Founded in 1996 by businessman Bill T. Gross, the California-based Idealab is one of the first technology incubators that has created over 150 companies with more than 45 Initial Public Offerings (IPOs) and acquisitions. In addition to capital, Idealab also provides startups with resources including office space and services including product and graphic design, marketing, financial advice, human resources, legal, accounting, and business development.
StartX. Founded in 2011, StartX is an independent non-profit business incubator affiliated with Stanford University. It requires no fees and doesn’t take equity in companies, but to be accepted through its application process, at least one founder of the company must be affiliated with the university. It helps entrepreneurs launch companies across industries and has a specialized track for medical startups. StartX helps participating companies hire talent, secure funding, and connect with a network of 1800+ serial entrepreneurs, industry experts, and 700+ well-funded growth-stage startups launched through the incubator including Patreon, LimeBike, Kodiak Sciences, and more. The element of a network is key in this incubator, as participating companies have access to the talent pool affiliated with Stanford, as well as startup founders who can serve as collaborators and users who will provide early feedback on their products.

Accelerators

Y Combinator. Y Combinator hosts two 3-month cohort-based programs each year, and it aims to help a cohort of startups – at all stages of development – get into a dramatically better shape at the end of the 3-month period. It makes a $500,000 investment in every company on 2 separate Simple Agreements for Future Equity (SAFEs), a common early-stage financing contract between an investor and a startup that gives the investor a future equity stake in the company if certain triggering events occur. During the 3-month period, startups have access to office hours, an in-person retreat, speaking events, and opportunities to network with other startups in the YC community and receive their feedback and engagement as first customers. They remain connected as alumni through online groups, in-person events, and YC’s private network for founders called Bookface.
Techstars. Founded in 2006, Techstars is one of the largest pre-seed investment companies in the world, having invested in more than 3,500 early-stage ventures. It provides 3-month accelerator programs located across the world that accept 10 companies each. During the 3-month period, companies work with 3-5 mentors who will advise them in any areas of business that need support, identify and achieve Key Performance Indicators (KPIs), and finally, create and refine their pitch decks in preparation for a “demo day.” Startups will also have access to a network of corporate partners, investors, and program alumni. Participating startups receive $20,000 and a $100,000 convertible note (a type of bond that can be converted into common stock), in exchange for 6% to 9% of common stock given to Techstars.

References

Andre, David (Chief Science Officer at X, the Moonshot Factory) in discussion with The Learning Agency, June 2023.

Anon. 2021b. Research collaborations bring big rewards: the world needs more. Nature 594, Nr. 7863 (16. June): 301–302. doi:10.1038/d41586-021-01581-z, https://doi.org/10.1038/d41586-021-01581-z.

Anon. 2023. Accelerator vs. Incubator (Similarities and Differences). https://ca.indeed.com/. 10. March. https://ca.indeed.com/career-advice/career-development/accelerator-vs-incubator.

Anon. About FROs — Convergent research. Convergent Research. https://www.convergentresearch.org/about-fros.

Anon. Inside a Techstars Accelerator: What To Expect From the Three Months. Techstars. https://www.techstars.com/the-line/advice/inside-a-techstars-accelerator-what-to-expect-from-the-three-months.

Anon. NAEP Long-Term Trend Assessment Results: Reading and Mathematics. https://www.nationsreportcard.gov/highlights/ltt/2022/.

Appendix B GPS history, Chronology, and budgets – CMU school of … Accessed September 21, 2023. https://www.cs.cmu.edu/~sensing-sensors/readings/GPS_History-MR614.appb.pdf.

Blank, Steve. 2022. Entrepreneurs, is a venture studio right for you? Harvard Business Review. 19. December. https://hbr.org/2022/12/entrepreneurs-is-a-venture-studio-right-for-you.

Buchsbaum, Steve (former director of the Office of Chemical, Biological, Radiological, Nuclear and Explosive Defense at the Homeland Security Advanced Research Projects Agency and former deputy director of Grand Challenges at the Bill & Melinda Gates Foundation) in discussions with The Learning Agency, June and September 2023.

Christina Hildebidle (Vice President, SRI International) in discussion with The Learning Agency, June 2023.

Cohen, Paul R. 2005. If not Turing’s test, then what? Ai Magazine 26, Nr. 4 (15. December): 61–67. doi:10.1609/aimag.v26i4.1849, https://doi.org/10.1609/aimag.v26i4.1849.

“‘The DARPA Model for Transformative Technologies’: An Interview with the Authors.” YouTube, October 22, 2020. https://www.youtube.com/watch?v=Of4UbOIxs4o.

The Economist. 2021. A growing number of governments hope to clone America’s DARPA. The Economist, 5. June. https://www.economist.com/science-and-technology/2021/06/03/a-growing-number-of-governments-hope-to-clone-americas-darpa.

Federation of American Scientists. 2023. Focused research organizations to accelerate science, technology, and medicine – Federation of American Scientists. Federation of American Scientists. 13. June. https://fas.org/publication/focused-research-organizations-to-accelerate-science-technology-and-medicine/.

Federation of American Scientists. 2023. Fund Organizations, Not Projects: Diversifying America’s Innovation Ecosystem with a Portfolio of Independent Research Organizations – Federation of American Scientists. Federation of American Scientists. 30. April. https://fas.org/publication/fund-organizations-not-projects-diversifying-americas-innovation-ecosystem-with-a-portfolio-of-independent-research-organizations/.

Floare, Flavius. What any investor needs to know about venture building studios – Grai Ventures. Grai Ventures. https://www.graiventures.com/articles/what-any-investor-needs-to-know-about-venture-building-studios.

Flood, Mark (Program Manager at DARPA) in discussion with The Learning Agency, May 2023.

Gates, Bill. “Why Our Foundation Takes on Grand Challenges.” gatesnotes.com, October 9, 2014. https://www.gatesnotes.com/Why-Our-Foundation-Takes-On-Grand-Challenges.

Jarrett, Daniel (Schmidt Futures Fellow) in discussion with The Learning Agency, September 2023.

Harris, Laurie A. “U.S. Congressional Research Service. Federal Research and Development (R&D) Funding: FY2023 (R47161; June 10, 2022).

Lang, David (Executive Director of the Experiment Foundation) in discussion with The Learning Agency, October 2023.

Paschkewitz, J., & Patt, D. (2023, November 1). No, We Don’t Need Another ARPA. Issues in Science and Technology. https://issues.org/arpa-catalyze-diffusion-paschkewitz-patt/.

Salman, Javeria and Javeria Salman. 2023. Is a ‘DARPA for education’ finally happening? The Hechinger Report (1. February). https://hechingerreport.org/is-a-darpa-for-education-finally-happening/.

Shilling, Russell. “Embracing the DARPA Model for EdTech Innovation: Charting the Course with GPT-4 and Beyond.” Getting Smart, April 10, 2023. https://www.gettingsmart.com/2023/04/11/embracing-the-darpa-model-for-edtech-innovation-charting-the-course-with-gpt-4-and-beyond/.

Zasowski, Nick. 2023. Disrupting the venture studio landscape – Morrow. Morrow. 20. April. https://morrow.co/disrupting-the-venture-studio-landscape/.

This brief was written by Tasha Hensley, Meg Benner, Lizzie Jones, and Yueying Yu.