If there is a recipe for military innovation, it is acute crisis and infinite money. While that’s a repeatable phenomena, it’s harder to muster the same effort, energy and resources when long-running conflicts remain quiet enough to lack that sense of emergency. How, exactly, can a more status quo force find a way to still innovate and iterate to meet the changing demands of modern warfare?

“Army Futures Command has really embraced academia and industry,” said Patrick O’Neill, director, C5ISR Center, Combat Capabilities Development Command, U.S. Army, speaking at the 18th annual C4ISRNET Conference June 6. One part of that embrace is tracking who is involved and who is funded, another part is embedding in academia itself, like the AI task force situated at Carnegie Mellon, and another part of that is making sure the funding in the department goes outside the usual suspects.

“For every dollar in our budget, 60 cents goes to industry and 40 cents stays in-house,” said O’Neill.

Funding alone is a start, but one of the harder challenges for military acquisitions is abandoning dead-end projects when it is cheap to do so.

“I don’t want to say we’re doing failure wrong, I’m just not ready to commit myself that we’ve changed and are ready to fail fast,” said Andrew Hunter, director, Defense-Industrial Initiatives Group and Senior Fellow, International Security Program, CSIS. “Learning to fail fast is hard, and in a culture that says failure is not an option, I think sometimes it might be.”

Ultimately, this becomes a task of prioritization, of finding what can be acquired easily from the commercial world, what can be adapted to military purpose, and what takes dedicated funding to develop — the kind of funding that can only come from the Pentagon.

“Industry is not driven by the commercial market to develop electronic warfare,” said O’Neill.

And even when commercial tech has a direct military application, it often needs modification and modification that is aware of the demands of the operating environment.

“Small innovative companies don’t understand what you’re deploying to,” said Ken Rice, deputy director for research, National Geospatial-Intelligence Agency. “It needs to be useful to people with high school education working in the desert; you can’t just ship the existing box.”

While legacy defense companies might not have the same innovation as smaller start-ups on the market, with VC funding and an existing relationship with the Pentagon, they can serve as important connectors and translators matching startups to military projects. And it means that the Pentagon may have to change how it buys to adapt to the the predominance of software-first design in the commercial world.

“Traditionally, it hasn’t been profitable to write software for DoD,” said Hunter. “You make profit when you deliver hardware.”

Better software can lead to better ways to use existing sensors and data collected. That’s important, because processing data is a software problem first. It’s a problem already tackled in a myriad ways in the commercial world, and one just waiting for easier modification to the military customer.

“What we have more of than anybody that we haven’t tapped yet is data,” said Rice. “We are tapping a very small potential of the data.”