The NDAA military spending bill cleared the US House of Representatives on Wednesday by the thinnest of margins, sending a $1.15 trillion package toward a Senate that may not be willing to receive it. The final count was 216 to 212.

That four-vote gap tells much of the story. Republicans held their majority together just barely, and only after absorbing defections that cut in both directions.

A Vote That Ignored the Usual Lines

Party discipline mostly held, but not entirely. Six Democrats broke ranks to support the package. Seven Republicans went the other way and voted against it.

For a bill routinely described as must-pass legislation, that kind of scattered opposition is unusual. It signals that the objections this year are not simply partisan reflexes but something more specific.

The Provision Drawing the Most Heat

Much of the friction centres on a section of the bill that would deepen technological cooperation between the American and Israeli militaries.

Republican Representative Thomas Massie, who has repeatedly broken with the White House and has been openly critical of the war on Iran, condemned the outcome in a social media post. He described the legislation as merging American military technology and supply chains with Israel’s, and called Section 219 a betrayal of national sovereignty. His stated hope was that the Senate would reject the version the House had just approved.

Massie’s objection lands at an awkward political moment. Public support for continuing military assistance to Israel has been slipping, and the measure has already run into pushback from lawmakers in both parties.

The Price Tag and What Drove It

The 2027 National Defence Authorisation Act reflects an aggressive spending request from President Donald Trump. His proposal sought to lift the defence budget from roughly $900 billion the previous year to $1.5 trillion.

That would be a dramatic expansion of a military budget that already has no global equal.

Trump has justified the increase by pointing to the ongoing war on Iran, which the United States and Israel launched in late February. The conflict itself remains widely unpopular with the American public, which complicates the argument considerably.

Extra Cargo on a Must-Pass Bill

Defence authorisation bills tend to attract riders, and this year’s version carries several that have little to do with troop pay or weapons procurement.

Among the additions:

• Assistance for Ukraine, a subject that continues to divide Republicans
• The SAVE America Act, a Trump priority that would introduce new voting restrictions

The SAVE America Act is framed by its supporters as a safeguard against election fraud. That framing rests on claims of widespread fraud that Trump has advanced without evidence.

Bundling these measures into a bill that Congress almost always passes is a familiar tactic. It also gives opponents more reasons to vote no.

The Senate Problem

Passing the House was the easier half of this.

Senate Democrats have already demonstrated their resistance. In a vote last week, they blocked debate on the NDAA outright. Minority Leader Chuck Schumer characterised the bill as a permission slip for the administration’s war on Iran.

The cost is the other sticking point. Democrats have objected to authorising more than a trillion dollars in military spending while the White House simultaneously works to reduce access to social programmes. That contrast, they argue, is difficult to defend to constituents.

What Happens Now

The bill’s path forward is genuinely uncertain, which is rare for defence authorisation legislation. Congress has passed some version of the NDAA every year for decades, and that streak has survived far worse political weather than this.

But the combination of factors here is unusual. A trillion-dollar price tag, an unpopular war, a sovereignty fight over Section 219, and unrelated voting legislation attached to the package all pull in different directions.

Something will likely have to be stripped or renegotiated before the Senate moves. The question is which piece gives first.