How Arc Institute is bringing science into the century of biology

If physics ruled the 20th century, the 21st is shaping up to be the century of biology. The past two decades have brought us the completion of the Human Genome Project, the advent of CRISPR technology, the development of mRNA drugs, advances in personalized medicine and stem-cell therapies, and, more recently, AI advances that promise to accelerate our understanding of diseases and drug discovery. 

We’re seeing a “Cambrian explosion in experimental and computational technologies,” bioengineer Patrick Hsu said in 2021.

But how science is organized and funded today may be dragging us down as we try to unlock this wealth of new ideas. That’s one of the motivating ideas behind the Arc Institute, a nonprofit biomedical research organization cofounded by Hsu in 2021 to accelerate medical breakthroughs.

The institute, based in Palo Alto, exists to pursue “high-risk, high-reward science,” backed by more than $650 million from funders like Vitalik Buterin of Ethereum and Patrick and John Collison of Stripe. Researchers come to Arc Institute with varied backgrounds and focus, but all work toward the same goal: to make concrete progress in understanding and treating complex diseases.

“Complex” here is not quite the same thing as “hard” — it’s a technical term, meaning the diseases are not caused by a single factor, such as one specific gene or virus. Without a lone culprit, scientists don’t fully understand all the factors that spark and drive diseases like cancer, Alzheimer’s, or immune disorders. This makes curing or slowing these killers down excruciatingly difficult.

“They’re so fundamentally different from the other types of diseases that we’ve made much more progress on — they do kind of require a different scientific approach,” says Silvana Konermann, Arc Institute’s cofounder and executive director.

What kind of different approach? While there’s no single solution, there is one obstacle that reliably blocks innovative ideas from making it from the drawing board to the lab: funding.

Chasing the money

What sets Arc Institute apart from other research groups isn’t its goal but rather its philosophy of freeing researchers from the red tape that clogs the pipeline of conventional research: lengthy grant applications, long lags between submission and receiving funding, high competition, and award systems that stifle creative ideas.

Like politicians on the fundraising trail, researchers spend tons of time seeking money for their projects, with more than 90% agreeing the total time spent on this is too much. Some researchers report having to spend half or more of their time on writing grant proposals.

Some funding comes directly from universities and philanthropy, in the form of university endowments or donations to private research centers. But the bulk comes from the “800-pound gorilla” of medical science funding: the National Institutes of Health (NIH), the world’s largest public funder of biomedical research.

“Today, if you have an idea, you have to go write a grant,” says Hsu, who’s also a core investigator at Arc Institute. “And the sort of standard grant that you write is called an [R01, awarded by the NIH]. It’s $250,000 a year, and you pay from that direct costs, budget, everything — salaries, benefits, equipment, reagents, sequencing, stem cell culture, animal costs — literally whatever comes out of that budget.”

Those budgets feel increasingly tight. In recent decades, the buying power of NIH grants has decreased while competition over them has increased. These conditions stem partly from the federal government’s move to start increasing the NIH budget in 1998 — doubling it from $14 billion to $27 billion over the brief span of five years.

Great news for biomedical research — but there was a downside. The rapid budget boost led to a significant expansion of research infrastructure and an influx of new scientists, resulting in larger research teams and larger projects. When budget growth froze in 2003, inflation began to erode the real value of the grants, creating a highly competitive environment where a recently expanded pool of researchers struggled to secure adequate funding to sustain their work. It was a “perfect storm” of increased demand, inflation, and flat budgets.

Today, Hsu describes grant-seeking as a “crazy” process where researchers are asked to “do more with increasingly less” — with ever more years of experience required on their CVs in order to even compete for funds.

“The average age to get these grants is getting older and older,” Hsu says. In May, Nexus noted that the median age at which researchers received their first R01 is 42. “So, you’re halfway between graduating from college and getting Social Security. And so that I think is a crazy observation.”

Even if you’ve got enough experience, write enough grants, and craft a winning proposal, you’re still looking at a long delay in actually starting the research.

“You submit this grant [and] it takes a year to get the money — and that’s if your grant is approved the first time,” Hsu says. “And if it gets rejected, which is 90% of the time, […] you have to do it again. And then now it’s like a two-year cycle.”

Applicants also must strategically consider the incentives of the NIH, whose funding priorities are determined by various byzantine councils — for instance, Hsu says, the “hilariously named” and “very Lord of the Rings” Council of Councils.

NIH funding also generally flows to safer bets, Konermann says. After all, the NIH doles out taxpayer money, so it prefers betting on research that is defensible and low-risk. Betting on high-risk, high-reward projects that ultimately prove fruitless would put NIH in a bad spot. 

“You might get a lot of questions from your Congress people and from anyone else in the public about like, ‘Why did you fund this crazy idea that turned out to be totally wrong, by this, you know, 25-year-old that doesn’t even have a PhD?’” Konermann says.

As for the safe bets that do get funded? “They’re probably going to do something reasonable with the money, but it might not be the most innovative or the most important,” Konermann says. “The system is […] acting rationally, you know, within those incentives. But you probably want to have more than one incentive system in a scientific ecosystem overall.”

Arc Institute aims to provide an alternative in that broader ecosystem, building on the small but highly successful tradition of mostly privately funded biomedical research centers, like the Howard Hughes Medical Institute and the Broad Institute. 

Instead of researchers scraping together their own grants, Arc Institute gives researchers no-strings-attached funding for eight years (with renewable terms), fosters collaboration with scientists in other fields, and provides access to cutting-edge technology in their labs.

The strategy is to accelerate research that would have a hard time getting funded through conventional means — and do so quickly. Rather than innovative ideas waiting a year to get funding, Hsu says Arc Institute’s model is: “Tomorrow.” 

Innovation, faster

Tomorrow isn’t much of an exaggeration. Before founding Arc Institute, Konermann and Hsu collaborated on Fast Grants, a program created by Hsu, economist Tyler Cowen, and Stripe cofounder Patrick Collison (who is married to Konermann). Launched in the thick of the pandemic, the program awarded grants ranging from $10,000 to $500,000 to COVID-related research projects. Submitting an application took 30 minutes. Funding was delivered within two weeks, rather than a year or two — sometimes as little as 48 hours.

A total of about $50 million in funding helped advance some notable research, including that of Yale’s SalivaDirect, which created a saliva-based COVID test and proved it was as effective as nasopharyngeal swabs.

Fast Grants showed, for Hsu, “that giving grants quickly doesn’t make the quality of granting worse.” Konermann says Fast Grants was an informative “dry run [for Arc Institute] of what it looks like to distribute capital at scale in a way that can hopefully be an effective model.”

Besides speed, the Fast Grants program also highlighted other problems with conventional funding. 78% of Fast Grants recipients said they would change their existing research programs “a lot” if they could spend their money however they wanted. They couldn’t, though, because their funding was awarded for a specific project, and that comes with specific constraints.

Arc Institute gives researchers a similar amount of funding as conventional grants but without tying them down and creating “weird incentives,” Konermann says. “So that means, a month in, you’re like, ‘This project isn’t working, like, why in the hell would I keep working on this … I should be working on this other project over here that’s going really well and that’s going to have a major impact,’” then the researcher is in charge — they can pivot with no questions asked.

It’s a model that Hsu and Konermann saw working firsthand when they met in the 2010s at the Broad Institute of MIT and Harvard, where both worked in a lab run by CRISPR pioneer Feng Zhang.

At the time, the Broad Institute was the newest entry to a relatively short list of biomedical research centers that operated outside the usual university and NIH funding pipeline, relying heavily on private philanthropic funding.

“In the United States, there’s a very long, 125-year history of focused biomedical research institutes that have this kind of middle space [that] sort of mix all of the other missions that happen at universities, but [are] purely focused on doing research breakthroughs,” Hsu says.

The tradition dates back to Rockefeller University, the nation’s first biomedical research institute, which was founded in 1901 by John D. Rockefeller Sr. after his grandson died from scarlet fever.

Other notable examples include the Salk Institute, the Whitehead Institute, and the Howard Hughes Medical Institute (HHMI), where Konermannn was accepted as a fellow in 2017. Like Arc, HHMI funds people instead of projects, providing researchers with flexible funding over eight-year terms to pursue high-risk, high-reward biomedical research. Since 1978, HHMI alone has supported more than 30 researchers who went on to win Nobel Prizes.

Federal institutions like the NIH are even trying alternative funding strategies, too. For example, the Common Fund of the NIH provides flexible funding for high-risk, high-reward biomedical research. But the Common Fund accounted for about 1.5% of the NIH’s total budget of $47.1 billion in 2024.

Teams on the cutting edge

Funding is only part of the equation. Hsu says it’s difficult to create research centers that sustainably and consistently focus on the most important problems. One limiting factor is the fast-changing nature of the field. “Biomedical research is fundamentally different as an entire landscape and industry than it was 20 years ago, say, when the Broad Institute was founded in 2004.”

The past decade or so has unlocked several generational breakthroughs and “tremendous opportunities for trying to understand and tackle complex human diseases, which is our core mission,” Hsu says.

Biotechnologies now exist that can read, write, and model nucleic acids (DNA and RNA) at scale. Large-scale computation and machine-learning techniques are now able to model and interpret complex biological systems. And researchers now have “the ability to start to do these experiments not in simple cancer cell lines,” Hsu says, “but in complex cellular models like organoids, or directly in vivo in an intact physiological environment to study disease.”

Leveraging all of these emerging technologies requires being interdisciplinary: blending neuroscience, immunology, computation, technology development, and chemical biology — and bringing them together to tackle complex diseases “under a single physical roof, which is what we’re doing,” Hsu says.

“You need interdisciplinary people at the interface that can kind of bridge each of these fields or dimensions.”

One product of Arc Institute’s interdisciplinary approach is Evo. As a project that combines machine learning, computational biology, and experimental biology, Evo is an AI model that can predict and design DNA sequences. The “DNA foundation model” is similar to a large language model like ChatGPT — except it’s not trained on English-language text but rather the language of evolution.

Unlike previous models that could only handle short sequences and specific tasks, Evo can analyze and generate whole genomes, offering a window into how complex systems emerge from subtle interactions between DNA, RNA, and protein components.

“We’re interested in using this to understand the language of evolution, to be able to predict drug response, […] design new types of medicines, and then, you know, model the complexity of an entire cell. […] This is sort of the size of project that is not $250k a year, R01 modular grant. It’s not the kind of thing you can go and raise money for from venture capital.”

That sweet spot — big and bold ideas unlikely to be funded in the university or VC space — is where Arc Institute is placing its bets.

“It’s very early and we’re just getting started,” Hsu says. “We’re tackling this sort of middle space of institute-scale problems like this for our virtual cell or our Alzheimer’s disease initiatives, that can’t really be done in other places.”

This article was originally published by our sister site, Freethink.