America’s Innovation Pipeline Is Leaking — and We’re Losing Game-Changing Startups
Foreword by Kumar Garg
In recent years, there has been a growing movement to bring the scientific method to scientific grant-making - to rigorously test different strategies for how to support scientific ideas and talent, and scale what works.
Often, these debates focus on discovery research - how to identify and bet on breakthrough ideas. What many of these debates leave out is that we need this same method of testing and scaling for the other major component of scientific enterprise: commercialization, or “lab to market.”
One of the greatest challenges in science and technology today is the consistent translation of those ideas into organizations and products that can deliver impact at scale. For all the federal investment we make in research, too many discoveries never cross the gap between the lab bench and the market. They are lost not because the science fails, but because the commercialization pathways that could carry them forward remain underdeveloped or unevenly applied.
Unfortunately, commercialization is an area where everyone has an opinion - or a complaint. Very few people have a systematic view of how we build commercialization as a discipline. That’s where Orin Herskowitz comes in.
I’ve had the privilege of working with Orin Herskowitz for many years, and one of key things you learn as you talk to him is that commercialization is a system and to improve it, you have to understand the system, test strategies for improving its components, and scale what works.
At Renaissance Philanthropy, this perspective deeply resonates with how we see the world: science, technology, and innovation are among the most powerful drivers of change, and philanthropy has a catalytic role to play in ensuring that more of these ideas become real organizations, startups, and platforms that benefit the public. For philanthropists, this means thinking beyond funding discovery alone, and instead also fixing the “pipes” of commercialization that allow ideas to flow into the world.
The recommendations that follow provide a roadmap: practical, proven steps that institutions can take to reduce bottlenecks, accelerate the launch of deeptech ventures, and ultimately strengthen the nation’s innovation ecosystem. For academia and industry, they offer guidance on how to make the journey from discovery to impact easier for researchers and entrepreneurs.For funders - both private and public - they highlight where strategic support can deliver outsized returns for society. In fact, Orin is now leveraging this playbook in his new role, as President of the Mussallem CHD Alliance, a foundation dedicated to saving and improving the lives of people born with congenital heart defects, no matter where in the world they live.
What follows is not a call for billion-dollar labs that can do it all, but a field guide for plugging the leaks in the innovation pipeline – ensuring that world-class research translates into world-changing impact. We hope you use it - and add to it.
Orin Herskowitz, Dmytro Pokhylko, and Kumar Garg
Every year, the U.S. government pours over $200 billion into research. That investment has helped fuel breakthroughs like advanced semiconductors, mRNA vaccines, and life-saving medical devices. But too often, brilliant ideas never make it from lab to market - not because the science fails, but because the process does.
The U.S. already has the world’s strongest basic research infrastructure and funding mechanisms - and, via the Bayh-Dole Act, a globally-envied way for those innovations to get launched out of research labs - which collectively led to the creation of thousands of new inventions, licenses, and startups per year. Collectively, over 19,000 startups and 200+ new drugs have been brought from academic labs to market in the past 30 years (source).
While all of these deeptech innovations will attempt to follow roughly the same path from the lab to the market, few will make it all the way to the end goal.
The “Pipe” Through Which Many Federally-Funded Research Innovations Flow: Representative Lifecycle for Science-Based Deeptech Innovations
Of course, many promising research innovations don’t make it to market for perfectly understandable and hard-to-fix reasons. For instance: the science fails, or works suboptimally compared to other approaches; the idea is ten years too early or too late; the science works, but without a rational business model; or the founding team collapses; the product fails in clinical trials.
But many other innovations won’t reach their potential for predictable and addressable reasons. To try to understand why, over the past two years we conducted over a hundred interviews with scientist entrepreneurs, Federal agencies, venture capitalists, tech transfer executives, and other relevant stakeholders. These interviews and analyses have helped identify over 60 avoidable failure points, which could be resolved at scale. These 60+ “leaks” generally fall into five broad categories:
Lack of access to appropriate training and incentives
Overly bureaucratic & slow processes
Lack of awareness of existing best practices
Lack of access to existing talent
Lack of access to existing funding sources
Our country’s main approach so far has been to push more and more water into the pipe - i.e. funding more research, launching more initiatives, adding more programs - without doing enough to fix its leaks. And we all know how that story ends: while some water will reach the end, even more water still leaks out and gets wasted in the process, leaving the bucket at the end half empty.
The frustrating part? The U.S. already has the hardest part covered: world-class science, a strong federal funding base, and a network of research institutions producing thousands of inventions each year. The path from lab to market for deeptech innovations is robust at many research institutions with good tech transfer offices (e.g., Stanford, MIT, Columbia, Berkeley) in established startup hub cities (e.g., New York, San Francisco, Boston, Los Angeles).
However, the vast majority of research institutions and cities elsewhere in the country, when compared on a research-funding-per-startup-created basis, are significantly behind at navigating this process.
Some of these advantages are too expensive or complex to easily replicate. For instance, not every institution can launch The Engine, as MIT did, with $150 million in dedicated funding and a whole building filled with expensive prototyping equipment. And not every city teems with serial entrepreneurs and venture capitalists the way San Francisco does.
However, many of these challenges can be addressed not by making massive investments or focusing on underperforming regions, but rather by simple process improvements. We are confident that many of these “leaks” could be resolved with minimal funding and effort.
Fixing this isn’t about building billion-dollar labs or hoping for unicorn founders to appear. It’s about sharing what already works - standardized best practices for tech transfer, accessible training for scientist-entrepreneurs, coordinated talent-matching programs, national toolkits for mentorship and patenting - and addressing financing gaps across the entire commercialization journey, from concept to global scale-up. We share some examples of possible ways to fix these leaks in the appendix below.
The parts of the graph above that are particularly prone to “leaks” are the university tech transfer and startup stage. At high-performing institutions, many resources are available to support campus scientist-entrepreneurs: gap funds to validate and accelerate innovations to the point of startup creation; talent matching programs to identify fundable entrepreneurs; executives-in-residence programs to provide high value-add coaching and mentorship; startup postdoc programs to give scientist-entrepreneurs the time and training to harden their innovations; entrepreneurial leave policies; streamlined access to shared core facilities; standardized and startup-friendly terms for IP license agreements. However, at most of our country’s universities and Federal Labs, some or all of these resources are lacking, making the path from the lab to the market much more perilous than it needs to be.
The result of the 100+ interviews mentioned above is a series of strategic and tactical recommendations that institutions ranging from academia and venture capital to industry and government can utilize to help fix the nation’s applied innovation plumbing. They can be found here.
If we plug these leaks, we won’t just get more startups. We’ll see:
Faster delivery of life-saving tech to the public
Stronger regional economies, including areas outside traditional hubs
The creation of thousands of high-quality jobs
A more competitive America in the industries that will define the 21st century
And the opportunity cost of these leaks is staggering; fixing them could have a massive impact on America’s economy and strategic competitiveness for generations to come. For instance, if we could simply bring the “lowest performing” large research institutions (as defined by rate of startup creation per million dollars of Federal funding for institutions with over $200M per year in total research) up to the mean (Washington State University’s level) based on 2022’s data, America’s research institutions could launch over 75 more critical and emerging technology (CET) startups every year. If we could increase the rate of startup creation by 20% over the mean (Arizona State’s level), America’s research institutions could launch 120 additional new startups compared to today’s output. If we improved everyone up to the middle of the top quartile (University of Arizona or NYU), we could create 335 more new CET startups per year compared to today’s rate. Other process improvements could help ensure that these ventures have a streamlined path towards the market, receiving the support they need just when they need it. And many of these startups could be the next Google, Moderna, or Nvidia.
The ideas and talent are already here. The question is whether we’ll fix the pipes so they can flow more robustly. Whether your organization is a university, a research lab, a philanthropy, or a government agency, we encourage you to use these recommendations to help fix your applied innovation plumbing.
Reach out to us at leakypipes@renphil.org with feedback on how these recommendations worked (or didn’t!) for you, if you have ideas you think should be added to the list, or if you want to explore ways to pilot these ideas.
Appendix: select examples of possible ways to plug the leaks
Below is a selection of a handful of “leaks” identified through this effort, along with suggestions of potential ways to address them. The full list can be found here.
Streamlining licensing. In Spring 2020, seven universities, six venture capital firms, and three of the nation’s largest startup law firms banded together to streamline the process by which IP is licensed by universities to startups in C.E.T. areas. Beginning with life science, those parties first developed a series of process guidelines for how such deals could be negotiated; followed by a term sheet outlining the key terms and proposed clauses for these deals; and, finally, a full license agreement with all of the required clauses already pre-populated. These template documents were openly posted to both the AUTM (Association of University Technology Managers) and NVCA (National Venture Capital Association) websites, so that all universities, startups, and VCs can benefit from starting from a common ground. Beginning in Summer 2023, a similar group created the same documents for climate technologies, based on the original life science templates. The climate documents were celebrated by White House OSTP, NSF, and DOE in April 2024. The full life science and climate agreements can be found on the web portals of AUTM and NVCA.
Best-practices training and knowledge banks. At high-outcome universities and at high-performing labs, scientists such as George Church, Bob Langer, David Baker, and others have “seen the movie” repeatedly on how to quickly and effectively launch startups from their labs based on promising innovations. However, at other universities without track records of serial startup success, would-be entrepreneurial scientists (faculty, postdoctoral researchers, PhD and graduate students interested in exploring commercialization as a path beyond bench research) lack the experience or training materials to emulate that success. Consistent and easy-to-access training materials on best practices from high-impact labs and prolific scientist-entrepreneurs could be created and made available nationally.
Playbooks for patenting heuristics. Tech transfer offices may file suboptimal patents or none at all. There is a need for creation and dissemination of patent filing best practices based on learnings from high-output university tech transfer offices at MIT, Stanford, Columbia, Penn, U Michigan, and others. By doing so, we would create a toolkit that would provide a leg-up to the less experienced tech transfer offices to effectively evaluate their IP assets and devise a robust filing process that would have significant timing and commercial benefits.
National Executives-in-Residence pools. Would-be scientist-entrepreneurs lack access to experienced mentors and coaches with relevant industry experience. Some universities have established Executives-in-Residence (XIR/EIR) programs to address this gap, whereby experts with decades of industry and founder experience would provide valuable guidance to early-stage researchers who are getting on the path of launching deeptech ventures. The assistance to universities would come in the form of a shovel-ready XIR toolkit that would include templates, training, SOPs, and support for setting up XIR programs. As this effort expands with time, a natural next step could be setting up a national XIR matching program, with additional focus on critical emerging technology areas for regions that may lack local expertise in certain fields.
Streamlined and consistent review processes. University conflict-of-interest (COI) review process may require lengthy and confrontational reviews and approvals. This is oftentimes caused by the fact that COI committees conduct their evaluations on a case-by-case basis, and exacerbated by ambiguity and unsynchronized COI policies across federal agencies. As a result, the path to startup launch is slowed down, and unnecessary tensions arise. We propose convening relevant leadership from Federal science agencies to discuss providing clarity around guidance to university COI committees, explore ways to extend PIs’ ability to continue building their deeptech ventures once they outgrow academic labs, and review the makeup of COI committees to ensure diversity of experience along commercialization lifecycle. Additional benefits of these convenings could be more synchronized COI policies across the primary Federal agencies that fund academic research (NIH, NSF, DOE, ARPA-E, ARPA-H), as well as development of a more robust guidance for universities on how those policies should be interpreted for emerging startups.
User-friendliness of federal programs. Despite significant potential, many excellent Federal commercialization programs like SBIR/STTR are difficult to discover, and even more difficult to navigate for many scientist entrepreneurs, especially at universities without robust support systems. Some of the ways to increase utilization of such programs could include the creation of an easy-to-navigate repository of available resources, as well as allocation to participating tech transfer offices a certain number of hours with pre-selected SBIR/STTR consultants to provide hands-on advisory to applying PIs.