Join War on the Rocks and gain access to content trusted by policymakers, military leaders, and strategic thinkers worldwide.
In the span of two weeks, the White House issued two of the most ambitious artificial intelligence directives in American history. On June 2, President Donald Trump signed an executive order mandating rapid AI adoption and hardened cyber defense across the government. Three days later, National Security Presidential Memorandum 11 directed every element of the national security enterprise to accelerate AI adoption, anchored by four pillars: adoption, adaptation, assurance, and accountability.
The strategy is sound. The harder challenge is execution, and a fiscal squeeze on the Fiscal Year 2026 operations and maintenance accounts threatens to stall it before it begins. On June 24, the White House sent Congress an $87.6 billion emergency budget supplemental designed to alleviate that squeeze. But it includes no dedicated funding for the AI software that the president’s directives require.
That is the problem in one line. The software that those directives call for is bought with operations and maintenance dollars, and the Iran war has drained those accounts.
As a provider of AI-enabled military planning software now used across multiple combatant commands, we at Onebrief do have a commercial interest in the AI software funding this piece argues for. But we argue here for replenishing the accounts that pay for command-level AI software, whether we fill those orders or our competitors do.
Private capital has made a generational commitment to American defense technology, yet the sector still draws less than 1 percent of total Pentagon contract dollars. Investors who backed these companies expect returns, and most of these firms depend on government contracts to earn them. They are building exactly what the president’s directives say the government must have, in a fiscal year that is running short on both time and money.
Operation Epic Fury, launched Feb. 28, has imposed costs that no budget anticipated. Then-acting Pentagon comptroller Jules Hurst told the House Appropriations Committee in May that the bill for operations in Iran had reached $29 billion, a figure the department acknowledged was incomplete. The Center for Strategic and International Studies has since put the war’s incremental cost at $34 billion to $42 billion, with the $87.6 billion request running higher because it also rebuilds munitions stocks and funds priorities beyond the war itself.
This strain is visible across the services. Chief of Naval Operations Adm. Daryl Caudle made it clear that his FY2026 budget wasn’t prepared for Epic Fury, and that training exercises, flight hours, and recruit instruction are already being cut as a result. Funding pressure that normally arrives in September — the last month of the government fiscal year — has hit months early.
This hits AI software at the command level. The Pentagon created a dedicated software appropriation in recent years, but it remains a limited pilot, and most fielded command tools are still bought the way other recurring services are, with operations and maintenance dollars. That puts AI software in direct competition with training, maintenance, and operating tempo for the same dollars, and the commands that most need decision-speed tools are often the least able to afford them.
The administration built parts of its FY2027 modernization plan on the assumption that a third reconciliation bill would include roughly $350 billion in defense funding. But top Republican appropriators have poured cold water on that plan, leaving the supplemental funding bill as the most realistic legislative vehicle for bringing in much-needed funding.
But while the White House’s supplemental request directs $67.1 billion to defense, including $17.3 billion to cover operational costs and $21 billion to replenish munitions, it does not include dedicated funding to implement the directives the president issued in June. The closest is $5.1 billion for “cybersecurity and autonomy,” with no detail on which capabilities that would fund. The bill eases the broader operations and maintenance pressure without protecting the software accounts that command-level AI adoption depends on. As Congress takes up the request, it should add an explicit line for AI software procurement.
It is fair to ask whether a dedicated line is even necessary. Supplementals rarely carry narrowly labeled software accounts, and the broad operating relief, together with the $5.1 billion for cybersecurity and autonomy, might reach AI adoption without one. But without dedicated line item funding, there is a significant risk that the supplemental funds will be diverted to the training, flight hours, and maintenance needs Caudle described, leaving command software last in line. The only way to guarantee that money survives contact with other priorities is to name its destination directly.
The executive branch also has additional tools to make this possible without Congress. The Office of Management and Budget retains apportionment control over the $152 billion in reconciliation funding that the Pentagon plans to spend within FY2026, much of which remains available. The administration could direct portions of those funds toward AI software adoption before September 30 without new legislation.
The Pentagon also has reprogramming authority under the FY2026 appropriations act. Used intentionally, that flexibility could move meaningful dollars into the technology and software accounts that fund commercial software licenses in the weeks that remain.
There is reason for optimism, and it rests with the people running the building. Secretary of Defense Pete Hegseth has made procurement reform a priority. Deputy Secretary of Defense Stephen Feinberg is implementing it with a unique mix of leaders, many drawn from a world where capital flows to what works. Strong commercial instinct with the need to get it right quickly can turn these directives into fielded capability.
The defense-tech ecosystem has answered the president’s call. The companies building this software are ready now and validated in the field. If FY2026 closes without meaningful new awards in the AI and commercial technology categories — not because the technology failed, but because operations and maintenance funding ran dry and no one acted — there will be real damage.
That outcome is preventable. The directives say what the government needs. What remains is directing the right resources to the right accounts before Sept. 30, a challenge well within the reach of an administration that has already done far harder things.
Adam Lackey is the chief operating officer at Onebrief, a Series D defense technology company and member of the 2026 NatSec100. Onebrief provides AI-enabled military planning software deployed across multiple combatant commands.
**Please note, as a matter of house style, War on the Rocks will not use a different name for the U.S. Department of Defense until and unless the name is changed by statute by the U.S. Congress.
Image: Petty Officer 1st Class Alexander Kubitza via DVIDS.
