Brandon Pugh, the Army’s principal cyber advisor, told Breaking Defense at TechNet Augusta that the service is at the stage where AI for cyber needs dedicated funding, and that it may not be Army money only. Joint solutions are common. He also said Army Cyber Command cannot fund that AI work strictly from its operational budget.
Breaking Defense later clarified the ask is for cyber capabilities across the Army, not solely ARCYBER. Read it that way. A command O&M account is the wrong pocket. A program-of-record line, or a joint pot, is the right one. Until that pocket exists, a program office cannot write the contracts small firms keep hearing about in the hallway.
“We’re at the stage now where we need to have dedicated funding to it and it might not only be Army funds, these are a lot of oftentimes joint solutions.” — Brandon Pugh, principal cyber advisor to the Secretary of the Army, TechNet Augusta, to Breaking Defense
The Context
When a DoD component says it needs dedicated funding, it is not being poetic. Operational money pays people, keeps networks up, and continues work already on contract. New capability is a different appropriation. If the line is not there, the contracting officer has nothing to cite.
Pugh advises the Secretary of the Army and certifies the cyber budget. That is why the quote matters. It is not a vendor panel. It is the official who has to sign that the cyber budget is real.
AI for cyber, especially defensive cyber, has been one of his office’s four top priorities this past year. After an April tabletop with industry, the Army stood up Rapid Defense Cyber Systems, also called Project ARDCS, and is concentrating on three first lines: Project Griffin (detect and act, not just watch), agentic deception, and agentic audit of vendor and network posture.
The System
Budget cycles and acquisition clocks are not the same clock. A need can show up in one fiscal year. The appropriation can show up in the next. Without a dedicated line or a transfer authority, the office cannot drop a solicitation that looks like a program of record.
That is the 6–18 month problem. The announcement is not the award. The justification book and the testimony are the tell.
What This Does Not Mean
It does not mean there is zero Army AI-cyber paper on the street. The same week, the Army released a commercial solutions opening for Project Griffin. DefenseScoop reported solution briefs were due August 27, with the government reserving the right to limit Phase II pitches. That is a CSO, a pilot lane, not the dedicated funding Pugh is asking Congress and OSD to create.
Do not staff a program of record to a CSO date. Do not ignore the CSO if you actually sell Griffin-shaped work. Those are different bets.
What This Means
If you sell Army AI cyber as if ARCYBER can pay for it out of this year’s operating account, you are early. The vehicle, when it comes, will tell you the shape: IDIQ, BPA, set-aside, or a joint program you are not on yet.
Watch the Army budget justification, Pugh’s next testimony, and whether Griffin stays a CSO or gets a real line. That timeline is the capture clock. The headline is not.
The Signal
Dedicated funding talk precedes contract availability by months. Track the book, not the booth. A CSO is a door. It is not the building.
Sources Checked
- Breaking Defense, Aug. 21, 2026 (updated 5:11 p.m. ET): “Army cyber defenses need ‘dedicated funding’ for AI, top official says.” Brandon Pugh interview at TechNet Augusta. Clarification: funding ask is Army-wide cyber capabilities, not solely ARCYBER. breakingdefense.com
- DefenseScoop, Aug. 21, 2026: Project Griffin CSO; solution briefs Aug. 27; ARDCS / token-cost concern. defensescoop.com
Source presentation is carried in the article body where applicable. Reported facts and analytic judgment are kept distinct.