A New Model for Philanthropy?

Two respected former leaders of the Defense Advanced Research Projects Agency, Regina Dugan and Kaigham J. Gabriel, have teamed up to lead a new philanthropy-funded, DARPA-like entity—Wellcome Leap. It is supported by the United Kingdom’s Wellcome Trust, an independent charity focused on health science. In “Changing the Business of Breakthroughs” (Issues, Summer 2022), Dugan and Gabriel propose this as a model for other DARPA-like entities to be funded by philanthropy.

Science policy theorists have long studied two innovation factors: research and development levels and directions and the talent base behind that R&D. The first focus stems from work by the economist and Nobel laureate Robert Solow, who argued that the dominant factor in economic growth was “technological and related innovation.” The second stems from work by the economist and Nobel laureate Paul Romer, who argued that “human capital engaged in research” was the vital factor behind the R&D for technological advance. These can be considered two direct factors behind innovation (as opposed to a multitude of indirect factors). However, a third direct factor, innovation organization, is less understood and has received less scrutiny. Dugan and Gabriel are, in effect, arguing for its centrality, pressing a new approach upon us.

They suggest that an era of complex technologies makes innovation organization a necessity. The lone innovator in the garage never happened; complex innovation (as opposed to examples of discovery or invention) requires a collaborative process involving a mix of skills and disciplines, putting innovation organization at a premium. This is not a new reality; it has been true since Thomas Edison’s group at Menlo Park developed the incandescent lightbulb and proposed the electrical system behind it. But getting to the right innovation organization is a minefield, littered with many inadequate models.

Dugan and Gabriel focus on DARPA, a famously successful model they know well. It has focused on taking high risks for high rewards, on breakthroughs not incremental advances, and it relies on empowered program managers to find the best research groups to implement new technology visions. They cite former DARPA program manager Dan Wattendorf’s vision that led to a critical DARPA effort a decade ago to advance mRNA vaccine technology. The model has been successful enough to have spawned successful DARPA clones, ARPA-E (for energy technologies) and IARPA (for intelligence technologies). A new clone, ARPA-Health, is now in the offing.

However, governments have faced challenges in creating ARPAs. Within the Department of Homeland Security, HSARPA was well-staffed at the outset by former DARPA program managers, but was never allowed to operate independently by its departmental overseers who limited its freedom of action. Other countries that have attempted an ARPA model have faced problems of locating it within an established agency, which can limit the needed entrepreneurial culture; of controlling the level of risk it can undertake; and of finding ways to link the ARPA to the scale-up efforts that must follow initial prototyping. Governments always have trouble with failure—of spending taxpayer dollars on high-risk ventures, whatever the potential rewards.

Could philanthropy be an alternative? Dugan and Gabriel suggest that it could face fewer of these restraints, citing their own Wellcome Leap effort. They argue that while governments must innovate within national borders, many technology answers, particularly in health, will be found by creating networks across borders, and philanthropy can operate internationally.

A potential problem for philanthropy is mustering the scale of funding needed. DARPA is a $3.8 billion a year agency. But how much funding do you need to make a difference? ARPA-E has shown that you can have a tenth of that funding level and spur important progress.

Also, philanthropy has been teaming up lately. Cooperation across leading foundations working on climate technologies is now widespread. Fast Grants has brought together some of Silicon Valley’s most successful—the Chan Zuckerberg Initiative, the Collinson brothers, Elon Musk, Schmidt Futures, Reid Hoffman, and others—collaborating to pool funding for projects such as a universal coronavirus vaccine.

Could there be too many DARPAs? In the 1940s IBM chairman and CEO Thomas Watson allegedly said there was a world market for about five computers. We’re now at about 2 billion and counting. Science has truly turned out to be an endless frontier that keeps building on itself; the more innovation the more opportunities there are for more. The DARPA innovation model has proven an unusually viable one; there seems no good reason not to bring on the clones.

Maybe we should encourage this?

William B. Bonvillian

As a former CEO and senior tech executive at companies such as Xerox PARC, Sun Microsystems, and Google, I have been a direct beneficiary of the DARPA model that Regina Dugan and Kaigham J. Gabriel describe. The Defense Advanced Research Projects Agency’s critical role in creating the internet is widely appreciated, but it also helped to enable many other technological revolutions, including design software for computer chips, handheld GPS receivers, speech recognition, and intelligent assistants. Google itself grew out of a DARPA-funded project at Stanford University on digital libraries. So I am a big believer in the DARPA approach of recruiting world-class technical program managers with a compelling vision, setting ambitious but measurable goals, and backing multidisciplinary teams to achieve those goals.

I am also delighted to see the growing support for the DARPA model, including the United Kingdom’s planned launch of the Advanced Research and Invention Agency (ARIA), funding from the US Congress for an ARPA for Health, and Wellcome Leap, led by Dugan and Gabriel. 

I’d like to pose three questions that, if addressed, could increase the impact of these and other future ARPAs.

What takes the place of Defense Department procurement for other ARPAs?

The original DARPA has benefited from the fact that Defense Department procurement will often create markets for the technology developed by DARPA’s R&D investments. What’s the equivalent of that for other ARPAs? Will market forces be sufficient to commercialize the results of ARPA-funded research programs, or will they get stuck in the “valley of death” between the lab and the market? One possibility to explore is what economists call “demand pull” (as opposed to “technology push”) approaches. For example, DARPA’s investment in the development of mRNA vaccines was complemented by Operation Warp Speed’s advance market commitment to purchase hundreds of millions of doses of a COVID-19 vaccine from Pfizer and Moderna.

What other pressing problems would benefit from a public or private ARPA?

For example, President Obama proposed creating an ARPA for Education, and the US House of Representatives recently provided funding for a National Center for Advanced Development in Education. What other economic, societal, and scientific challenges would benefit from an ARPA? What goals might these ARPAs set, and what are examples of projects they might support to achieve those goals?

What can we learn from the original DARPA, and what experiments should new ARPAs consider?

The original DARPA has been operating for over 65 years, and I think there is more we can learn by studying the different strategies used by DARPA program managers. For example, one highly successful DARPA program was called MOSIS, which provided shared access to semiconductor fabrication services to academic researchers and start-up companies. This accelerated the pace of innovation in microelectronics by providing access to an expensive resource, allowing more people to get involved in semiconductor design. There are dozens of these DARPA strategies that new ARPA program managers should learn from. New ARPAs should also take advantage of their ability to experiment with new models, such as Wellcome Leap’s Health Breakthrough Network.

Eric Schmidt