With New START gone, The U.S. and Russia both claim defensive intent, but history suggests neither can see clearly enough to be believed.
In the Marine Corps, I flew hundreds of missions that were briefed about deterrence, protection, and stability — responses to threats, not contributions to them. That wasn’t dishonesty. It was something more structurally embedded, and considerably more dangerous.
New START expired February 5. For the first time in over half a century, there are no legal limits on U.S. and Russian nuclear arsenals. Both governments have offered their explanations: Washington wants a better agreement; Moscow cites NATO encirclement. Both describe their nuclear posture as defensive, and neither appears to register how structurally identical those justifications look from outside the system.
The perceptual gap
This is not a new dynamic. Russian strategic culture, whether under tsars, Soviets, or Putin, has consistently framed expansion as existential necessity. A history of catastrophic invasion from the west created a logic in which buffer zones equal survival. That logic makes sense from the inside. From the outside, those same moves are viewed as an encirclement of their neighbors. The gap between those two interpretations has historically produced wars, not prevented them.
The American version of this operates with a different vocabulary. The preferred terms are partnership, leadership, and responsibility, not empire. The strategic framework I operated under as a Marine aviator was internally consistent: forward presence prevents conflict; alliances stop wars; positioning deters aggression. Whether other countries experience American presence the same way is a question the framework has never answered very well.
The recent renaming of the Department of Defense to the Department of War, a rhetorical shift that would have been politically unthinkable a decade ago, illustrates how that gap is widening. Other states register these signals even when we don’t. Language that signals dominance rather than defense changes how every deployment, every alliance expansion, and every modernization program gets read in Moscow, Beijing, and Iran.
Why this matters now
The expiration of the New Start treaty creates a specific danger that can’t be reduced to simple budget calculations. A January 2026 poll found 91 percent of Americans, including 85 percent of Trump voters, support negotiating a replacement agreement. Nearly three-quarters believe removing all nuclear limits makes the United States less secure. The public understands the stakes more clearly than the current policy posture reflects.
What polling can’t capture is the dynamic that makes miscalculation most probable. Without verification programs and shared data exchanges, both sides now rely on intelligence assessments and inference. Modernization programs, which both governments describe as qualitative improvements to existing forces, look from across the table like preparation for first use. Systems designed to preserve deterrence get interpreted as threats to it. In the absence of institutional mechanisms that constrain that interpretive spiral, the spiral accelerates on its own.
The Congressional Budget Office estimated in 2025 that U.S. nuclear modernization will cost $946 billion over the next decade. The Sentinel ICBM program alone is already 81 percent over budget, and the National Nuclear Security Administration is running seven modernization programs simultaneously. A numerical arms race layered on top of this would compound the cost while delivering no security return. Russia’s comparative advantage is industrial capacity; a warhead-count competition plays to its strengths and drains American resources away from conventional priorities (ships, drones, cyber) that actually shape the current threat environment.
The structural problem arms control has avoided
Arms control negotiations are typically designed to address bad-faith behavior: verification catches cheating, transparency builds confidence, and inspections deter violations. That framework is necessary, but also insufficient. The more enduring problem is that each party’s position warps perception even when the other party is operating in good faith. States with overwhelming power tend to read their own deployments as necessary responses, and others’ matching moves as unprovoked threats. This isn’t self-serving spin, it’s how one’s position shapes analysis from within.
I spent years flying missions whose briefings were completely logical as delivered. They were also, as I came to understand later, incapable of asking certain questions. The missions made sense as responses to conditions we’d identified as threatening. Whether our presence had helped generate those conditions was not a question the framework answered. That cognitive limitation wasn’t unique to the Marine Corps or even to the United States. It’s what structural position does to analysis.
History offers limited encouragement here. Dominant powers rarely adjust because they conclude they’ve been wrong. They adjust when the cost of sustaining a self-justifying posture exceeds the cost of revising it. Currently, both nuclear powers remain in the phase where the narrative is still cheaper than the correction.
What a realistic path forward requires
The immediate practical steps are not complicated. Both sides could publicly commit to honor New START’s core limits while a successor agreement is negotiated. Notification protocols before major nuclear exercises and missile tests could be restored without a treaty. Crisis communication channels between military commands, degraded or inactive across multiple recent crises, could be reopened. None of this requires trust. It requires recognition that the alternative is more dangerous and more expensive for both sides.
The harder prerequisite is whether either government can develop the institutional capacity to see how its posture registers externally. That is not a question traditional arms control negotiations ask. It probably should be a foundational one though, because agreements reached without that shared understanding tend to collapse the moment either side perceives a relative disadvantage.
The last time a U.S. president was willing to ask something close to that question seriously, Reagan and Gorbachev’s joint declaration that a nuclear war “cannot be won and must never be fought,” it produced the agreement architecture we just allowed to expire. That architecture was imperfect and incomplete. It was also the product of two leaders who, briefly, managed to see their own positions from the other side’s vantage point. We have no comparable starting point now, and the window for building one is narrowing.
In a nuclear system without limits, the cost of misjudgment is not theoretical. That much, at least, both sides agree on, even if agreement on that point alone is not enough to prevent it.
