WASHINGTON — The Pentagon spends $3.3 billion a year on its 216 laboratories, which have piled up thousands of patents, often for technologies which may never see the light of day, let alone a battlefield. But this morning, the Department’s CTO, Under Secretary for Research & Engineering Emil Michael, publicly launched a two-pronged crusade to change that.
“[It’s] a frustrating point: Why do these innovations — and we have thousands of them in the labs, billions of dollars worth of IP that’s been created by the great minds in the labs — why does it not get all the way out there to the warfighter?” Michael asked a packed conference room in downtown Washington, DC. “In part, it’s because you don’t know where to go to find them. They’re all over the place. They’re not categorized, they’re not available.”
Hence his two-part plan:
Step one, effective immediately, is to make roughly 400 carefully picked patents available online for a free two-year trial period. Specifically, any company that wants to try out one of the 400 technologies in its own research, development, and products can get what’s called a Commercial Evaluation License (CEL) without the usual fee.
Those 400 technologies — everything from a Navy-developed drone tracking system to novel Army mortar fuses — were chosen out of the thousands of possibilities by Michael’s staff, with an eye to his recently announced top six Critical Technology Areas. There were so many options from so many labs, he said, that they had to use AI to help sort through them.
“Here are the patents we think are important, are interesting, have merit, that you can develop on and potentially productize,” Michael said. “We’re going to give you a two-year patent holiday, royalty-free.”
If the project goes well and the company wants to keep using the patent beyond the two-year free trial, well, in true Trumpian fashion, Michael says he’s ready to make a deal.
“See what you could do with them, see if you can make a business out of them, and then come back to us … and let’s figure out a long term-arrangement,” he told the executives at the Pentagon-backed conference, hosted by consulting firm SMI.
It’s not as if the Pentagon is giving up a lot of revenue by sharing this intellectual property for free, he said. While it does license some patents to industry already, Michael told the executives, “the amount of money that we make from patent fees today is infinitesimal — and it’s not because they’re not good patents, [it’s] because you don’t know about them, and we haven’t created enough of a way for you to get to them and develop on them.”
More Data, More Problems
Step two, in progress, is to put all those thousands of patents from all 216 labs into a single searchable database for the first time, using a longstanding public-private partnership called TechLink and an interagency database called iEdison. (Explicitly not included: classified patents for technologies who very existence is kept secret.)
After almost two years of work behind the scenes, things are now moving fast, said Bethany Loftin, director of the Technology Partnerships Office at the National Institute of Standards & Technology (NIST), the Commerce Department agency that runs iEdison. That database currently holds ideas from some 36 federal agencies that fund research, including 10 of the Defense Department’s labs. But now an interagency Memorandum of Understanding has been thrashed out to bring in the other 206.
RELATED: Hegseth presses defense execs to move faster in speech laying out sweeping acquisition changes
“I keep checking my phone this morning because the final MOU for that relationship is on my boss’s desk for final signature,” Loftin said excitedly on a panel after Michael’s keynote speech. “So hopefully, maybe even before the end of the day we’ll be able to officially start the process of getting DoW, as a whole, onboarded onto iEDISON.”
Those thousands of patents won’t be available for free, Michael made clear — although, again, he’s willing to negotiate.
As for the first 400 royalty-free patents, they’re more like the free samples a supermarket puts on display to get customers in the door, he told reporters after his speech.
“It’s the freebie … the door-buster … the loss-leader,” Michael said. “Then hopefully you’ll get interested enough that you could look at the whole broad portfolio.”
That said, if the first 400 attract not only a lot of interest but actual investment that starts turning into usable military gear, “maybe we expand it,” he told the reporters. “That’s why it’s a pilot, right? We’re trying to see what happens when you put things out in the wild.”
In fact, the whole “Patent Holiday” idea came out of Michael’s desire to hype up the patent database and get things moving quickly, one of his subordinates told the assembled executives.
“I was like, ‘I want to build a data estate,’” said Steve Luckowski, the Pentagon’s director of Technology Transfer, Transition, and Commercial Partnerships.
Luckowski said Michael told him, “Let’s curate the patents. Let’s analyze them. Let’s make them available to industry. Let’s not wait. Let’s move fast.’”
AI was essential to that speed, Michael told reporters. “We used our best minds [on] manufacturing, biotechnology, [etc.], had them do the prompts … and try to distill it down to something that they thought was usable. So it had a kind of machine and human component to it.”
In the longer run, putting all the Pentagon patents into a single, searchable database is a classic big-government, big-data problem. There are thousands of files scattered across hundreds of organizations with no central clearinghouse or common standards. Again, it will take AI to tame the chaos.
“You heard Hon. Michael talking about how all these assets are all over the place. They’re literally scattered amongst the 216 laboratories,” said Clara Asmail, a contractor working for Michael’s office as senior program manager for technology transitions. “It’s very challenging to be able to compile, department-wide, all of those assets. So that is the crux of what our office is now engaged in doing.”
“That characterization cannot be done manually,” Asmail told the conference. “Everybody would agree here the reason that it’s never been done, but we now finally have nascent AI tools that, if we are careful and apply them in a way that we are intentional … we can start that processs.”