Project Maven and Colonel Drew Cukor tell a blunt story about how Pentagon bureaucracy can chew up a smart, effective leader who tried to bring commercial-style software practices into national defense. This piece follows how he shifted acquisition thinking, clashed over intellectual property, endured anonymous attacks, survived invasive investigations, and ultimately saw his career damaged even after an unclassified report found no fraud. It’s a portrait of how institutions protect themselves and how innovation gets punished when it up-ends the status quo.
In 2017 Cukor treated software as software, not as fragile hardware wrapped in procurement rituals. He argued Maven should be bought as a continuously evolving capability, with steady costs across its life rather than one massive up-front payment. That ran against a Pentagon habit of treating software like finished hardware, which slows updates and stifles real improvement.
To make this real, Cukor used Broad Agency Announcements to categorize work as RDT&E and keep the program agile. Commercial vendors arrived with mature products and decades of investment, not blank slates needing basic research. The result was faster delivery, closer iteration, and systems that actually improved in the field rather than rotting on a shelf.
The sticking point became intellectual property, because the government reflexively wants to own everything it pays for. Cukor pushed back hard, insisting that if vendors cannot monetize the tech after partnering with DoD, they won’t keep collaborating. As he put it, “If you [the company] can’t monetize this after working with us, then what’s the use of doing this? Why would you hand over your IP ever?”
That stance protected vendor incentives while still keeping government data and mission-specific logic secure. ITAR and other controls prevented vendors from misusing government material, while vendors kept proprietary model weights and platforms. In practice Maven preserved vendor IP while giving the government rights to mission configurations, enabling a healthy public-private ecosystem.
For that approach Cukor drew fire. He became an easy target inside the acquisition community for upsetting entrenched flows of money and influence. Opponents filed a steady stream of anonymous complaints that treated innovative contracting as if it were criminal behavior, not reform, and the bureaucracy went to work piling on investigations.
The allegations ranged from absurd to malicious: illegal contracting, luxury cars stored in his home, lavish entertaining to win contracts, and even harboring foreigners in his basement. Some charges were plainly nonsense tied to his sponsorship of immigrant mathematicians, but the process proceeded anyway. Cukor defended his choices and pushed for transparency, noting “I had some very strong captains that would happily tell off a colonel or general if they were wrong. We had a climate of moving fast and getting things done.”
Investigations conducted by rival services and NCIS found no evidence of criminal conduct and no mystery bundles of cash. Still, the campaign left scars: threats of demotion, career lists blocked, and relentless IG reviews that dragged on for years. As Cukor observed, “You just have to understand this: When one group of people in the Pentagon get ahead of everybody else, the natural reaction is to kill that thing and get everyone back in line. That’s the Pentagon.”
The Office of Inspector General later released a redacted evaluation that cleared Maven of fraud and found the program ran “in accordance with FAR, DFARS, Defense Grant and Regulatory System, and contract requirements.” The IG’s main gripe was documentation of monitoring approaches, even while admitting acquisition rules don’t capture AI methods well. The report essentially said Maven worked, but the institution resented the way it worked.
That institutional reflex—protecting the machine at the expense of outcomes—gets promoted. “Those that ascend are a rare breed: they’ve figured out how to survive in an environment where people can log any complaint against them and start investigations that jam up everything.” By contrast, Cukor moved at speed, experimenting with vendors outside the Beltway and shielding teams so they could perform, earning a nickname: the “iron dome of Pentagon bullshit.”
Cukor’s story is not a sermon, it’s a warning. Rotating officers too fast, rewarding caution over results, and weaponizing oversight all discourage leaders who can actually deliver new tech quickly. If we want a military that wins and out-innovates rivals, we must protect people who get things done rather than the processes that keep everyone the same.
