The service wants to buy 16 of these Shahed look-alikes, with an option to buy 20 more later, as it pursues the “next generation” of counter-drone programs, according to a request for information posted last week.

“To support weapons development and integration of these weapon systems, the [U.S. government] requires that the Class 3 unmanned aerial target system be a 1:1 copy (form, fit and function) of a reverse engineered Shahed-136 suicide drone,” the solicitation said. 

The drone must be an “exact replica” of the Iranian bird with the same profile, shape, payload capacity—about 70 to 100 pounds—and must be able to fly at least 50 miles, according to the solicitation. That’s far less than the Shahed’s range of over a thousand miles, but sufficient for testing purposes.

Designed in Iran, modified and mass-produced under license by Russia, the Shahed-136 has emerged as a favorite weapon of the invading forces inside Ukraine. Their price tag is estimated at $30,000 to $40,000 apiece, a fraction of the cost of the U.S. and European missiles used to take them down. That imbalance, plus depleted stockpiles of interceptors, has Ukraine and its supporters hunting for cheaper defenses.

Several U.S. firms have already started designing Shahed-esque offerings for the Pentagon. During a drone demo event at the Pentagon last month, SpektreWorks, an Arizona-based drone manufacturer, showed off its new Low-Cost Uncrewed Combat Attack System, dubbed LUCAS, that can emulate a Shahed. Another firm, Alabama-based Griffon Aerospace, recently unveiled the MQM-172 Arrowhead, marketed as both an attack and target drone.

While the Air Force provided some guidelines in the solicitation, the government said it won’t provide a technical data package for this requirement, so companies must be able to design and develop their own copy. 

The solicitation notes that the drone will be sent to the service’s armament directorate, which develops weapons at Eglin Air Force Base in Florida.

“You tell this organization to do ‘A.’ Then you cut contracted staff by 80 percent and you have a turnover of close to what, 60 percent? Things are going to break. Things are going to get delayed. We’re in both places,” said one defense official who asked for anonymity to speak freely.

In 2021, the Defense Department comptroller launched the “Advancing Analytics” platform—soon shortened to Advana—as an accounting tool that could help keep track of what the Pentagon was buying and where it was sending it. DOD leaders pushed commanders and offices “hard,” in the words of one official, to adopt it. Within months, Advana had some 20,000 users across 42 organizations, and Booz Allen Hamilton had a five-year, $674 million contract to help expand and maintain it.

In less than a year, Advana was helping U.S. troops coordinate with European partners to track and deliver munitions and other supplies to wartorn Ukraine. The system proved so useful that in July 2022, DOD’s inspector general chided elements of the military for not adopting it fast enough.

As more DOD offices and agencies connected their systems to Advana, users across the Pentagon could coordinate their work—logistics, contracting, finance, and more—far more easily and accurately. By 2023, Advana had more than 72,000 users and far more data than initially expected.

This unexpected success strained the Advana system, so in June 2024, the Pentagon’s Chief Digital and Artificial Intelligence Office—which had taken over the Advana program—put development on hold so that its user interface could be improved and its underlying architecture could be strengthened. 

CDAO leaders also decided to split Advana into two parts: one to handle the urgent, often classified needs of warfighters, and another to handle slower, less dynamic business records, according to current and former officials. Among other things, this was intended to smooth the introduction of AI tools onto the platform. Yet the two parts would remain connected so that, for example, acquisition officials could make decisions based on battlefield data, while commanders could bring supply-and-contract data into their own war planning.  

One month after pausing Advana development, CDAO unveiled plans for a 10-year, $15 billion Advancing AI Multiple Award Contract, or AAMAC. This vehicle was intended to bring in other software makers and cloud providers to help Booz Allen Hamilton split up Advana, give it a better front end for users, and update its back end for performance and growth. CDAO officials formally announced the plan in September and opened a process for contractors to submit proposals. 

Cuts and cancellations

The election of Donald Trump brought a new team to the Pentagon. In February, a senior defense official described a desire to convert the Advana project into a program of record, ensuring consistent budgeting. In April, the Defense Department axed what it called “duplicative” consulting contracts, including at least one to Booz Allen. Two current officials we spoke to said the action reduced the amount of contractor support the company was able to provide to Advana. But a former senior official said that the point of AAMAC was to bring in additional contractors to work alongside the company. Booz Allen remains under contract to support Advana through early 2027. A company spokesperson said, “Decisions regarding current Advana contracts were made before the memo, and any staff reductions made earlier this year were due to normal contract transitions rather than the memo or DOGE reviews.” 

Contracts are reduced or changed all the time. Defense officials assured the public that the program was still on track and the upgrade was in the works.

But the Trump administration took aim not just at excess consultants, but what they deemed excess staff as well. On Feb. 20, Defense Secretary Pete Hegseth announced plans to cut 5 to 8 percent of his department’s roughly 760,000-person civilian workforce, an effort aided by White House executive orders and its Department of Government Efficiency, run by Pentagon contractor Elon Musk. 

The cuts hit the effort to modernize Advana hard. The CDAO lost two top architects as well as supporting staff, nearly 60 percent of its workers. Current and former officials said the cuts hollowed out the technical workforce needed to implement the upgrades. 

The schedule began to slip. Promised upgrades were repeatedly delayed. Users across the department continued to wait. Last month, the entire AAMAC was put on hold.

Why did the loss of a few staffers, following a plan to expand the program, have such a big effect? Standing up a highly automated data environment works differently for the Defense Department than it would for a Fortune 500 company, which could call a vendor and easily purchase an out-of-the-box solution. The department has to verify the security and reliability of software, hardware, and even data according to specific standards. Even if some of the bureaucratic obstacles are unnecessary, many still exist in law and can’t just be ignored. Simply bringing in a data source to help train an AI platform can take months of review and security verification, according to one source. Then there is the issue of actually understanding Defense Department data, including where to find it across departments, offices, commands, and agencies. In short, no matter what the Defense Department decides to purchase as part of a new Advana effort, expert staff, now absent, will be necessary to make it work. 

“Literally, there’s no one else to do it,” said one defense official.

Far-reaching effects 

An indefinite delay to a data platform update may not sound like the sort of thing that could have far-reaching effects for the largest military in the world. But aside from its role in accounting—core to the stated goal of enabling the Defense Department to pass an audit— Advana has become the key data platform for many of its activities, including operational logistics and Joint Staff planning for large-scale conflict.

In future wars, defense planners anticipate “contested logistics,” making resupplying troops and other administrative functions more important, complex, and difficult. For more than a year, units have been incorporating Advana into key wargames and exercises. These experiments have demonstrated the value of easy access to as much information as possible about supply, logistics, and other “business” matters. One former senior defense official described it as essential to the goal of “getting ahead” of adversary efforts to thwart resupply and material support. 

But unless the Advana platform gets the upgrades needed to keep it from buckling under the load, it will suffer a type of collapse, current and former officials said. Already, the current architecture does not meet demand and it’s difficult to run advanced analytic tools on it. Data, they said, is returning to “silos,” meaning officers and commanders are holding it rather than incorporating it into a central repository. They worry that simple requests will go back to an analog system of phone calls and emails, exchanges that could take days or weeks as opposed to instantaneous. Some officials said this is already happening in isolated instances. 

A larger concern is that individual services, commands, and offices may build their own data analysis tools but won’t have the benefit of all available Defense Department data. And exchanging data between entities will return to phone calls and email exchanges. That will sever the connections built up over the past few years, plunging the Pentagon back into a world of separate systems, figuring out orders and drafting plans through the exchange of PDFs, and Excel spreadsheets.

It’s “a little back in the future,” said one official.

The picture is of a Defense Department moving away from faster, better, joint, all-domain command and control, getting slower as adversaries get faster.

Coming changes

Changes are already underway—for CDAO, at least. 

The office is being moved under the office of the defense undersecretary for research and engineering, according to an Aug. 14 memo. The memo gives Undersecretary Emil Michael and his fellow defense undersecretary for acquisition and sustainment 120 days to deliver a plan for Advana and the Maven Smart Systems program, another CDAO AI effort. Defense One reviewed a copy of the memo, which was first reported by DefenseScoop.

Current and former officials were cautiously optimistic that the change could provide high-level attention to the program to get it fixed— most importantly, by restoring cut staff. 

“Putting CDAO under R&E should help focus resources to do the upgrades necessary to make Advana an available platform for Fourth Estate use,” said one current official.

However, another current defense official expressed concern that the move might shift data-centralization efforts away from AI-development programs, just when the two should be more closely integrated.

A second former senior defense official said putting CDAO under the R&E undersecretary might turn AI efforts that are close to deployability into a “science project.”

“Everyone is focused on moving projects out of R&E faster and fielding faster, which is a good thing. But this puts even more on their plate in a way that seems unnecessary. It could work out just fine, but there’s a lot of risk,” they said.

Asked for comment on Advana and the merger memo, a defense official sent a statement. 

“The OUSD(A&S) has completed an assessment and delivered a recommendation on Advana, which is under review by Senior Department leadership. The Department of Defense is continuing to take decisive action to implement AI across enterprise functions and workflows, including instituting directives from President Trump’s AI Executive Order and the White House AI Action Plan. 

“Claims of dysfunction are unfounded; our efforts are focused on accelerating AI adoption to ensure the Department remains at the forefront of this new technological frontier.”

One month ago, a banner was added to the CDAO website: “More information on the Advana way ahead coming soon.”