Following the money

The federal workforce shrank by roughly 270,000 people in 2025. Federal IT contract spending hit $130 billion that same year, up from $126 billion the year before, up from $120 billion the year before that. The trend line goes through the purge as if the purge never happened.

Q4 of FY2025 was on pace to obligate nearly $50 billion in IT contracts, a record for any quarter in federal contracting history. The acquisition workforce processing those contracts is operating at vacancy rates approaching 40 percent at some agencies, according to the Professional Services Council, the trade association representing the firms doing most of the contract work. Fewer contracting officers are pushing more money out the door, faster.

This is the story of the last twelve months and almost nobody is telling it.

What was actually cut

DOGE cancelled or restructured an estimated $85 billion in contracts, heavily concentrated in USAID, climate work, DEI programs, and a handful of consulting agreements that made for good press releases. Aggregate contract spending kept rising because the work the federal workforce was doing did not disappear when the workers did. Every function that lost a federal employee retained a need. Every retained need became a contract. The shadow workforce — already two contractors for every one federal employee before any of this started, per Apollo Global Management's chief economist — absorbed the load and grew.

The Cato Institute summary of the year was that DOGE had no noticeable effect on the trajectory of spending while engineering the largest peacetime workforce reduction on record. The fiscal claim collapsed and the institutional damage held.

The acquisition crisis is the fiscal crisis

The 40 percent vacancy rate among contracting officers is the buried lede of the entire post-DOGE landscape. The people who write contracts, scope contracts, oversee contracts, and catch contracts going sideways are exactly the people you cut hardest in a generic headcount-reduction campaign. They are mid-grade career employees in agency procurement shops, sitting in unsexy roles, exactly the kind of mid-career operational employees that buyout programs target by design.

Cutting them produces no spending reduction. The money is appropriated, the work has to happen, and the contracts get written by fewer people, faster, with less oversight, against vendors who have noticed that the negotiating partner across the table is a tired person carrying twice the load.

The cost of bad contracting in FY2026 and FY2027 will be enormous, and most of it will be invisible because the function that catches it is the function that just got hollowed out. The IRS preview of this dynamic is already published: a Yale Budget Lab estimate found that losing 22,000 IRS employees would forfeit $8.5 billion in revenue in 2026 and approach $200 billion over the decade. The IRS lost more than 17,000 employees. The math runs from there.

The cohort window is closing

Eight months into the purge, agencies started bringing people back. Labor reversed deferred resignations. Employees whose roles were eliminated received notices reassigning them to different positions. OPM director Scott Kupor acknowledged in public comments that the workforce reductions were overdone. The reversal pattern was case-by-case embarrassment management run by managers who needed specific people back to do specific work that had visibly broken.

The cohort that left in the first wave is dispersing in real time. People with active clearances, agency-specific knowledge, working relationships across program offices, and demonstrated delivery records are taking private sector jobs every week. Clearance currency lapses, non-competes get signed, and institutional memory that took a decade to build evaporates over months.

A rapid rehire authority targeted at this specific cohort would be the highest-return personnel intervention available to the federal government right now. The legal authorities mostly exist through Schedule A appointments, direct hire, and reemployment after reduction-in-force. What is missing is the political will to do it systematically and the operational infrastructure to identify, contact, and onboard the cohort before they are gone.

By month 18, the cohort is gone. The cost of rebuilding equivalent capacity from new hires runs five to ten times higher and produces lower-quality results because the relationship layer cannot be hired for. The window is closing in measurable increments and nobody with the authority to act on it is treating the question as urgent.

The micrograms play nobody is running

10x is the model that already worked and nobody is running.

For readers outside federal tech: 10x was a technology investment program inside GSA's Technology Transformation Services, founded in 2015 out of a one-day pitch event called The Great Pitch. The premise was that federal employees working inside programs see operational problems first, have ideas about how to fix them, and the federal funding apparatus is built to ignore exactly those people. 10x inverted that. Any federal employee could submit a problem and a sketch of a solution in three sentences through a Google Form. The 10x team triaged the submissions and funded the promising ones through four phases of escalating investment, starting at small money to investigate whether the problem was real, then discovery to understand whether technology was the right answer, then prototyping with an agency partner, then scaling a working product across government.

Roughly one in three projects advanced from one phase to the next. Most ideas died in Phase 1 or Phase 2, which was the point — the program treated killing a project as a success because it preserved the budget for ideas that would actually deliver. Login.gov, the U.S. Web Design System, and Notify.gov all came out of this pipeline. The total annual budget was a rounding error against any single agency IT line item.

The model worked because federal teams did the work, funding was gated and incremental, ideas came from operators rather than acquisition strategists, and the overhead ran an order of magnitude smaller than any equivalent contractor engagement. Institutional knowledge stayed inside the agency. Bad bets got cut early. Real friction got targeted instead of hypothetical demand. The program ran inside the existing GSA operational stack.

10x was effectively shut down in the 2025 TTS dismantlement that eliminated 18F. The 10x AI Sandbox, one of its 2024 projects, was rebranded as GSAi by DOGE allies and pointed at a different mission. The program apparatus, the four-phase model, the intake pipeline, the evaluation function — sitting on the shelf, reconstitutable with a single appropriation and a director.

A rebuilt 10x at $50 million a year would capture more genuine capacity than $5 billion in equivalent contractor spending. The reasons this is not happening are revealing. Micrograms do not generate consulting fees, do not produce the press releases that the foundation-funded civic tech ecosystem runs on, and embarrass the contractors who would lose the work, the agencies that prefer contractors as accountability sinks, and the political appointees whose theory of change requires big announcements rather than working systems.

The civic tech funding world has spent the last year defending the civil service as an institution rather than as a function. You can lose the function while keeping the institution and end up with buildings full of people who cannot deliver because the people who could deliver are on contract or out of government entirely.

Where the money should go

Pay GS-14s and GS-15s competitively for technical roles. The labor market for the skills agencies actually need is not the labor market the federal pay scale was built for. Direct hire authority for technical positions, expanded and made permanent. Term-limited appointments for mid-career hires who shouldn't have to navigate a multi-month onboarding process. A rapid rehire pathway for the FY2025 cohort, run as a program with a deadline rather than as a series of agency-by-agency reversals. A 10x-scale internal grants program with a real evaluation function. Procurement reform that breaks the contractor stranglehold on agency IT through systems engineering and technical assistance contracts.

None of this needs philanthropic money. All of it needs political will to take on a contracting industry that funds both parties. The contractors know the rebuild is coming. The fed-spend trade press is already telling them how to position for it: agencies that lost capacity through contractor cuts will rebuild that capacity through new contracts with different requirements, different structures, and different set-aside designations. The dependency deepens with every cycle.

The diagnostic

The purge produced no spending reduction, a still-growing IT contract market, an acquisition workforce too thin to oversee what it is buying, a contractor-to-employee ratio that just got more lopsided, and zero structural reform of the hiring and procurement systems that created the dependency in the first place. The cohort that could rebuild capacity is leaving the building. The funding ecosystem that should be financing the recovery is financing conferences about the recovery.

The recovery work is sitting on the table. Somebody could pick it up.