If President Joe Biden’s last big science project was a moonshot, his new one has goals that are light years further.

The proposed Advanced Research Projects Agency would deliver breakthrough treatments for cancer, Alzheimer’s, diabetes and other diseases and reshape the government’s medical research efforts, by adding a nimble new agency modeled on the Pentagon’s Defense Advanced Research Projects Agency, or DARPA, which laid the groundwork for the internet.

But the way Biden would make “ARPA-H” and its $ 6.5 billion budget part of the sprawling National Institutes of Health is raising concern within the research community and in Congress about whether it will bring a new approach to old problems or become a duplicative bureaucracy with a lofty mandate.

“Most of us did not support putting this in NIH, for the simple reason that if NIH were capable of doing this, it would have done it,” said one person outside the government familiar with the planning who’s worried NIH’s staid culture and leadership will bog down the effort.

A half dozen individuals both inside and outside the administration who were involved in discussions about the plan told POLITICO there are alternative approaches being discussed, like putting ARPA-H well outside of Washington, to escape some of the Beltway’s inertia and turf battles. More autonomy could, in theory, speed up the way scientific discoveries are turned into drugs and diagnostic tests.

But the prevailing view is that making the new agency part of NIH’s infrastructure will give it a foundation to spring off — and foster communication to head off unnecessary duplication. As Congress prepares for hearings on the first budget proposal, administration officials are expressing confidence ARPA-H can carve out a distinct identity, wherever it is.

“[The established NIH culture is] a valid concern and we have to do everything to prevent that from being the default,” NIH Director Francis Collins told POLITICO. Referring to his agency’s many constituent parts, he added, “This is not going to be the 28th institute.”

Biden has long aspired to build a broad successor to the Cancer Moonshot, the $ 1 billion initiative he launched as vice president during the Obama administration in the hope of fostering a decade of cancer research in half the time, three people familiar with his vision said. He announced the Moonshot in the same emotional 2016 Rose Garden speech where he said he would not run for the presidency, citing the pain of losing his son Beau to brain cancer. Within a year, President Donald Trump was in office and White House focus on the massive research initiative dried up.

Making ARPA-H a federal agency would go a long way toward ensuring it could survive such political winds and power transitions. But the bar is high, especially with budget hawks in Congress and other skeptics who question what a new multibillion-dollar agency can do that others can’t.

“If it’s a vision of merely doing advanced translational research … we have the ability within NIH to do that already,” Rep. Andy Harris (R-Md.) told Health and Human Services Secretary Xavier Becerra during a hearing on his department’s fiscal 2022 spending plan.

Discussions about a new federal agency to cut through research barriers originated with a plan dubbed ‘HARPA,’ first pitched by the pancreatic cancer group Suzanne Wright Foundation to President Donald Trump in 2017. But critics said the original vision was too narrow, and a proposal for the agency to track mentally ill consumers in a bid to head off mass shootings proved too controversial. Collins himself questioned the need for HARPA, two people familiar said. The NIH director told POLITICO that the Covid-19 pandemic and Biden’s support for ARPA-H helped change his thinking about what is possible.

The NIH already has a department, The National Center for Advancing Translational Sciences, that focuses on finding practical applications for scientific discoveries. Harris and other Republicans argue that type of work is better done by the private sector, while NIH’s mandate is basic science — like mapping genes — that can be a foundation for other studies.

Collins said ARPA-H could bridge an important gap between academic research and industry, and spark collaborations across multiple federal agencies. “There’s often this gap, this valley of death,” between basic science and practical use where “we could play a really important role,” he added.

Part of DARPA’s model — which critics argue is hard for NIH to duplicate — is an organizationally flat structure that gives project managers vast autonomy over their work and funding decisions. It is a model that lets scientists quickly succeed but also quicky fail and then move on to new work, said Ellen Sigal, chairperson of Friends of Cancer Research who sits on NIH’s council of public representatives. Sigal, who has been involved on talks about the new agency, said, “If we’re going to do something audacious like DARPA, we have to have the mentality of knowing that we’re going to fail.”

Collins acknowledged the benefits and limitations of the system he oversees. “Our NIH process for how we fund research is both the best in the world, because of the rigorous peer review system, but it’s also a little slow, maybe a little conservative, and it isn’t necessarily going to embrace the really big transformative projects that would result in somebody sending you a grant application.”

But many of those big decisions, from which disease areas to tackle to a focus on basic or practical research will land on the shoulders of ARPA’s first leader, setting up what could be a high-stakes search for its founding director.

“The first director of this organization is going to be an incredibly important decision because they are going to set the stage for the future,” said Tara Schwetz, assistant director for biomedical incentives in the White House’s science office, who is shaping a plan for the new agency along with the White House’s top science official Eric Lander. “Culture is really difficult to change; it’s really important to get all this stuff right from the beginning because it’s hard to fix down the road.”

If Congress goes along with Biden’s funding requests, ARPA-H will have an annual budget of $ 6.5 billion available to spend over three years. That timeframe gives the agency some runway to fill out its ranks and fund its first key projects. But Schwetz and others admit it will be important to rack up quick wins to gain momentum and build sustained congressional support. They agency will also have to set priorities as advocates for cancer, Alzheimer’s disease and a range of other conditions clamor to shape its work.

“You get one shot at doing this and it really needs to have the necessary ingredients to succeed,” said a person familiar with talks between government groups and outside advocates. “ARPA-H needs to have an ambitious enough agenda that it is dealing with the most important problems we have in a way that allows people to take those and run with them.”