Replicator's Ghost: What Survived the Handover

In August 2023, Deputy Secretary Kathleen Hicks stood before the defense industry and promised to field thousands of attritable autonomous systems within 18 to 24 months — a deadline chosen to beat the acquisition system at its own game. Three years on, with the administration that birthed it gone and its successor indifferent, the honest accounting can be written. Replicator did not fail. It did something more instructive: it dissolved.

What the Deadline Was For

Replicator 1.0 was never primarily a procurement program. It was a proof-of-concept about velocity — Hicks's wager that the Pentagon could move commercial autonomy from selection to fielding on a wartime clock if the Deputy Secretary's office personally airlifted it over the acquisition bureaucracy. The budget was deliberately modest by Pentagon standards, roughly $500 million per year across FY24 and FY25. The named picks said everything about the intent: AeroVironment's Switchblade-600, Anduril's Ghost-X, PDW's C100, uncrewed surface vessels — systems that exist, fly, and could be bought by the pallet rather than the prototype.

The logic was Indo-Pacific. The 2027 Davidson window hung over every briefing: if deterrence required mass and shipbuilding could not deliver it, attritable autonomy would. Hicks called it "making the PLA think again."

What Actually Fielded

The systems themselves mostly arrived — that is the part the obituaries get wrong. Switchblade-600 went from selection to Army and Marine Corps orders, with AeroVironment's loitering-munition line running at a tempo that would have been unthinkable in 2022. Ghost-X matured from a quiet Anduril product into a fielded ISR platform with Air Force and Navy buy-in. The C100 became a standard-issue quadcopter across Army units, its $15,000–20,000 price point turning the company-level drone from an experiment into an expectation. Unmanned surface vessel experimentation fed directly into the services' own programs.

What Replicator did not deliver was the slogan: there is no formation of "thousands" operating under a Replicator flag, and there never was going to be. The initiative was a procurement accelerant, not a fleet. The hardware flowed into normal service channels — which was, depending on the brief, either the quiet success or the quiet evaporation of the whole idea.

The Pivot and the Patron

The first fracture came in September 2024, when Replicator 2.0 pivoted from delivering drones to defeating them. The counter-UAS focus was defensible — Langley's 17 nights of unexplained drones in December 2023 had made base defense a scandal in waiting — but it confessed something: the original promise of attritable mass was harder than the counter-drone mission it was replaced by. Counter-UAS had an urgent, fundable customer. Swarming mass in the Pacific had a PowerPoint.

Then came January 2025. The administration changed, Hicks departed, and Replicator lost the only institutional asset it ever truly had — a patron with the authority to convene, override, and sign. The initiative was never stood down; it was simply no longer stood up. The Defense Innovation Unit, which had run the plumbing, kept the contracts. The brand went quiet. By late 2025, senior officials had stopped saying the word.

What Outlasted It

Here is the comparison that stings. While Replicator faded, the services' own drone programs — the ones with program managers, budget lines, and service chiefs behind them — kept moving. The Army's Transformation in Contact brigades bought C100s and Ghosts at unit level and kept buying. The Marine Corps' loitering munition programs hardened into requirements. The Navy's unmanned surface efforts survived their own scandals and reorganizations because they belonged to someone.

The lesson is the oldest one in the building. Signature initiatives launched from the Secretary's suite die with the signature; programs owned by a service survive transitions because a service is a permanent interest with a permanent budget. Replicator accelerated real hardware — but every system that outlived it did so by getting adopted, quickly, by an institution that would still exist after the next inauguration. The ones that stayed orphans are powerpoints now.

The Verdict for 2027

Replicator's real legacy is conceptual, and it is everywhere. Attritable mass is now a budget category, not a heresy — the FY27 request is full of line items that would have been laughed out of a 2021 POM cycle. The phrase "affordable mass" appears in service documents without attribution, the way all victorious ideas eventually do. DIU's role as the fast lane for commercial autonomy is institutionalized in a way it was not before Hicks proved the lane could carry weight.

But the institutional lesson runs the other way, and the building has absorbed it: never launch a marquee initiative without a service owner, a program element, and a lifespan longer than a deputy secretary's tenure. Replicator 1.0 took 18 months to prove the Pentagon could move fast. Replicator's ghost is proof it cannot move fast for long — not without a home.

Replicator didn't field a fleet. It fielded an argument — and the argument won the budget while losing the war for a name.

What to Watch

- FY27/FY28 attritable-systems budget lines. Whether line items for loitering munitions and attritable ISR grow under their own names is the true measure of Replicator's absorption. - DIU's post-transition role. If DIU retains its Replicator-era convening authority under the current leadership, the fast lane survives; if it reverts to a scout-and-broker function, the lane is closed. - A 'Replicator 3.0' under new branding. Watch for a mass-autonomy initiative relaunched with a new name and a service sponsor — the concept is too useful to stay dead, and too radioactive to keep its name. - Counter-UAS follow-through. Replicator 2.0's counter-drone pivot intersects with this month's base-defense push; whether those funds flow through Replicator's old plumbing or new channels reveals who owns the mission. - Pacific fielding counts. Actual unit-level autonomous system counts in INDOPACOM exercises — not press releases — are the scoreboard the 2023 promise should be judged against.

The initiative is gone. The idea is in every budget line that matters, wearing someone else's name — which is how the Pentagon keeps its ghosts useful.