Without a Treaty, How Does the U.S. Track Russia’s Nuclear Arsenal?

Last updated 1 month ago. Our resources are updated regularly but please keep in mind that links, programs, policies, and contact information do change.

On February 5, 2026, New START expired without replacement. For fifteen years before that, U.S. inspectors had walked through Russian missile silos with measuring tapes and clipboards. They counted warheads. They recorded serial numbers on individual missiles. Every six months, Russia handed over declarations stating exactly how many nuclear weapons sat on weapons ready to launch—not estimates, actual numbers. Inspectors could cross-check those declarations against what they saw on the ground. Now the United States faces a problem it hasn’t confronted since the early 1970s: monitoring a nuclear adversary’s arsenal without any legal right to verify what’s there.

Intelligence agencies can still see what matters. Satellites photograph Russian missile bases from orbit. Electronic eavesdropping networks listen to communications and technical data transmitted by missiles during tests. Human sources inside the Russian military report on deployments and readiness levels. Commercial satellite companies sell imagery that researchers use to track construction at nuclear storage sites.

But every assessment now carries more uncertainty. The question is: at what cost, with what blind spots, and with what risk that a miscalculation based on incomplete information triggers something catastrophic?

What Treaty Verification Provided

New START was an intelligence collection system disguised as an arms control treaty. You can photograph a missile sitting in a launcher from orbit. You cannot determine whether it carries one warhead or ten. You cannot see inside the part of the missile that holds the nuclear weapon to verify a warhead is present rather than a training dummy. A deployed intercontinental ballistic missile looks the same from orbit whether it’s fully armed or empty.

Russia provided technical data transmitted by missiles during tests, letting U.S. technical experts analyze how new systems performed. The treaty imposed counting rules that prevented cheating. Once a missile was tested with multiple warheads, all missiles of that type had to be counted as carrying multiple warheads, even if deployed with lighter loads.

Russia suspended on-site inspections in March 2020 citing COVID-19 concerns, then suspended full treaty participation in February 2023 and discontinued notifications in March 2023. Now those declarations are gone permanently. The U.S. no longer receives official Russian statements about warhead inventories, deployment locations, or modernization plans.

Intelligence must infer Russian intentions from indirect indicators: satellite imagery showing weapons being moved, communications intercepts revealing operational matters, human sources reporting from inside the system. This is orders of magnitude harder than having Russia provide the information directly, even in an adversarial context where deception remains possible.

On-site inspections are gone, meaning officials cannot physically verify what they see through other means. Satellite imagery showing unusual activity at a missile base might indicate a weapons deployment, a training exercise, or routine maintenance. An inspector on the ground could determine which with certainty. Analysts now make educated guesses based on contextual clues, and those guesses carry margins of error.

Some Russian weapons are mobile—missiles on trucks or trains that move from location to location. On-site inspections let teams locate mobile systems and track their movements over time. Satellites must catch mobile systems in transit, a hit-or-miss proposition. Miss the window when the launcher is moving, and it disappears.

The treaty also provided technical information that helped assess Russian capabilities. When Russia tested its new Sarmat ICBM, the treaty required Russia to share technical data from missile tests. Signals intelligence satellites intercepted these transmissions, providing direct insight into missile performance. Post-expiration, Russia has no obligation to conduct tests the same way and can take measures to prevent interception. The U.S. may now have less detailed knowledge of how Russia’s newest weapons perform.

Teams who spent years at Russian missile bases developed expertise in how Russians operated these systems. They understood the routines, the security procedures, the tricks used to hide things. They built professional relationships with Russian counterparts that, while competitive, included implicit understandings about transparency. Those relationships are severed. If arms control discussions resume, that institutional knowledge will be difficult to rebuild.

Satellite Reconnaissance: Capabilities and Limits

Today’s satellites are incomparably more capable than early reconnaissance systems. The National Reconnaissance Office maintains a constellation optimized for different intelligence tasks. High-tech cameras capture imagery from orbital altitudes with classified resolution. Radar systems create images using microwave signals that penetrate clouds and darkness, allowing observation during bad weather and at night. Signals intelligence satellites intercept electronic transmissions: radar signals, communications, technical data from missile tests.

Recent imagery has revealed major construction at five Russian nuclear weapons storage sites—triple-layered security fencing, new buildings, covered loading ramps, large antennas.

But satellites cannot count warheads. They can see a missile being transported or mounted in a launcher. They can photograph a submarine entering a base and infer operational status based on how long it stays in port. They cannot look inside the part of the missile that holds the nuclear weapon and determine whether it’s present or a training dummy. A deployed ICBM might carry one warhead or ten. A mobile launcher might have a missile inside or be empty. These distinctions cannot be resolved through imagery alone.

A single satellite observes a particular location once per day in standard low Earth orbit. Mobile ICBMs can move quickly. A launcher photographed at location A in the morning might be at location B by evening. Miss that window, and it disappears. Over extended periods, satellites track patterns and trends. For moment-to-moment situational awareness, they’re comparatively slow and inflexible.

Russia has spent decades developing techniques to hide military activities from observation. Mobile ICBM units hide under camouflage netting. New weapons get concealed in specially designed buildings. Russia understands where satellites are in their orbits and can time activities for periods when satellites won’t be overhead.

Even when satellites capture clear imagery of military activity, interpretation requires context and assumptions that can be wrong. Is a missile base conducting a major deployment or a training exercise? Is the activity consistent with normal operations or something unusual? Ground truth is gone—no teams verifying what’s happening—so these judgments remain probabilistic rather than certain.

Intelligence officials assess that satellite reconnaissance provides sufficiently detailed information about major Russian nuclear activities to maintain understanding of force composition and deployment patterns. The NRO has evolved its satellite constellation significantly since New START was signed in 2010, with improved resolution, expanded radar imaging, and enhanced signals intelligence collection. These improvements have partially compensated for lost treaty verification data. However, intelligence officials acknowledge that unilateral monitoring cannot achieve the same confidence level that treaty verification provided.

Signals Intelligence: Telemetry and Communications

If satellites are the eyes of intelligence, signals intelligence is its ears. The National Security Agency operates a global network of listening posts designed to intercept electronic communications and emissions. These systems sit at the heart of intelligence collection on Russian nuclear forces, though much of what they accomplish remains classified.

Telemetry—technical data transmitted by missiles during flight tests—is the most relevant signals intelligence for nuclear monitoring. When Russia tests a new ICBM or submarine-launched ballistic missile, the missile’s onboard computers transmit information about acceleration, engine functioning, stage separation, warhead deployment, and numerous other technical details. This information is transmitted at radio frequencies that U.S. satellites positioned to receive it can intercept. NSA has maintained this collection capability continuously since the late 1950s, when ground stations in Turkey first captured such data from Soviet short-range missiles.

Analysis allows technical experts to reverse-engineer the performance characteristics of missiles: range, accuracy, payload capacity, guidance system performance. Before the treaty was signed in 2010, the U.S. had already acquired decades of such data from missile tests. The treaty formalized and regularized this collection by requiring Russia to provide certain information at speeds the U.S. could receive it, and by prohibiting Russia from deliberately denying access through encryption.

Post-treaty, Russia faces no legal restrictions on encrypting its transmissions. NSA may still intercept the signals, but encryption makes the raw data unrecoverable. Russia also has no obligation to conduct tests in the same manner as before, meaning it may test missiles in ways designed to provide less information to intelligence collection. A former NSA official noted that “the absence of the treaty’s constraints on encryption means that the amount of unencrypted data available for intelligence analysis may diminish over time.”

NSA conducts broader signals intelligence collection on military communications about nuclear forces. When commanders at a missile base communicate with higher headquarters, when submarines report their status to fleet command, when logistical units arrange warhead transport—these communications can be intercepted and analyzed. NSA operates listening posts around the world, including the Ayios Nikolaos station in Cyprus, which provides coverage of the Mediterranean and Eastern Europe. Other stations in allied nations, plus signals intelligence satellites, intercept communications and pass them to intelligence analysis centers where linguists and military experts interpret what they mean.

Communications intelligence provides context for other intelligence collection and reveals Russian intentions and capabilities in ways that purely technical collection cannot. A satellite photograph showing unusual activity at a nuclear storage facility is more meaningful when combined with intercepted communications indicating that a high-level inspection or readiness assessment is occurring. Communications about weapon movements from central storage to forward deployment areas reveal operational patterns. Communications about training and exercises provide clues about Russian confidence in their systems and readiness levels.

Russia’s military has become increasingly sophisticated in ways to keep military activities secret, using encrypted communications and requiring face-to-face meetings for sensitive discussions that cannot be intercepted remotely. The FSB hunts for listening devices and communications intercept operations. Particularly sensitive discussions about strategic nuclear decisions may be conducted in ways designed to prevent interception altogether. Intercepted communications often require weeks or months of analysis to understand fully, meaning they may not provide the real-time awareness that immediate threats require.

Human Intelligence: Intentions and Capabilities

Satellites and signals intelligence provide technical data about what Russia is doing with its nuclear forces. Human intelligence—information acquired through espionage and human sources—provides insight into why Russia is doing certain things and what Russian leaders intend.

The CIA and Defense Intelligence Agency maintain networks of human sources in Russia who report on developments within the military and government. These sources range from military officers who provide information about weapons systems and operational plans, to civilian employees in defense ministry facilities who report on personnel movements and administrative changes, to officials in government ministries who provide insight into policy discussions.

A source within a strategic forces command post might reveal the readiness status of a weapons system—whether it’s fully operational or degraded. A source in a weapons design bureau might reveal the true timeline for development of a new missile system or information about design flaws that other intelligence sources might interpret differently. A source with access to leadership discussions might provide insight into strategic doctrine and intentions that no amount of satellite imagery can convey.

Security services conduct spy-catching operations designed to identify and eliminate sources inside Russian institutions. In recent years, several human sources within the government have been identified and prosecuted, indicating that espionage networks have suffered significant compromises. The difficulty of recruiting new sources, the risk to sources once recruited, and the time required for sources to move into positions of access where they can provide valuable information all create constraints on how quickly the intelligence community can replace losses.

With the treaty in effect, some cooperation on arms control verification meant that teams and Russian military officials had regular contacts and opportunities to cultivate sources. Those relationships are now severed. With the treaty in effect, official Russian declarations and inspections provided reference points against which to evaluate human source reporting. If a source claims Russia has deployed a certain number of warheads, there’s no longer treaty data available to confirm or contradict that claim. Human source reporting must be evaluated against other intelligence sources with greater uncertainty about whether sources are reliable or have been deliberately fed false information by spy-catchers.

Commercial Satellite Imagery

In the decades before commercial satellite imagery became available, intelligence on Russian nuclear forces was exclusively a government function. The NRO operated the only reconnaissance satellites capable of high-resolution imagery, and that imagery was classified and available only to cleared government officials. This changed dramatically in the 1990s and 2000s, as private companies began launching commercial satellites capable of capturing high-resolution imagery and making it available for purchase.

Companies like Maxar Technologies and Planet Labs now operate satellite constellations that capture imagery of virtually every location on Earth. Maxar’s WorldView satellites can resolve objects on the ground smaller than a foot across—degraded in resolution from classified NRO systems but still sufficient for many intelligence purposes. Planet Labs operates a constellation of over 100 small satellites that provide daily coverage of most of Earth’s land surface at medium resolution.

The Federation of American Scientists has used commercial satellite imagery to track nuclear weapons storage sites, identifying major renovations at the Kulikovo facility in Kaliningrad that suggest Russia is preparing to store nuclear weapons in that region. The Middlebury Institute’s James Martin Center for Nonproliferation Studies has used commercial satellite imagery to monitor uranium mining sites and nuclear testing facilities.

Commercial satellite imagery faces inherent limitations for intelligence collection. The resolution is less precise than classified NRO systems. Availability of imagery is sometimes constrained by cloud cover or by the orbital patterns of commercial satellites. Most critically, everyone has access to the same imagery, meaning allies have it, but so do Russia and other adversaries. Adversaries can view what commercial satellites are observing them with and take countermeasures.

Commercial imagery has become an important supplementary source for monitoring Russian nuclear forces, filling some gaps created by treaty expiration. The Intelligence Community now regularly integrates commercial imagery into its assessments of nuclear activities, alongside classified NRO imagery. Commercial imagery also provides a source of information that can be used for unclassified intelligence assessment and sharing with allies—no need to compromise classified collection methods.

Allied Intelligence Sharing

The Five Eyes alliance—comprising the United States, United Kingdom, Canada, Australia, and New Zealand—shares signals intelligence and other classified information on a basis of mutual trust. Britain and France, as European nuclear powers themselves, conduct independent monitoring and share selected intelligence with U.S. agencies and NATO. Germany and other NATO members contribute intelligence from their own sources.

The treaty had provided a common baseline for understanding Russian nuclear forces—official declarations that both U.S. and allied intelligence services could use as a reference point for validating their independent assessments. Allied intelligence services now have no common standard against which to compare their findings. Britain, France, and other European nations have less transparency into assessments of Russian nuclear force composition and may make different judgments based on their own intelligence collection.

The Five Eyes alliance includes information sharing agreements and classification restrictions that don’t necessarily apply to all NATO members. Some NATO members have access to intelligence assessments of Russian forces, while others don’t. France, which maintains an independent nuclear deterrent and independent intelligence services, may have different assessments than the United States does. These differences can complicate NATO planning, which historically assumed a common understanding of Russian nuclear capabilities based on treaty verification data.

If the United States relies entirely on its own intelligence collection, it’s vulnerable to Russia deliberately hiding activities in ways designed to exploit collection patterns. British and French satellites may be observing the same facilities from different orbital angles and at different times than U.S. satellites. German listening posts on NATO’s eastern edge intercept communications from angles and locations where collectors might not be positioned. When allied intelligence is pooled, the combined effect provides richer, more redundant, and less easily defeated observation than any single nation’s collection.

Intelligence sharing of classified information remains constrained by legitimate security concerns. If the United States shares classified intelligence about Russian nuclear forces with a NATO ally, that intelligence must be protected by the ally’s classification and security classifications. If that ally is compromised by espionage or spy-catching operations, the intelligence becomes available to Russia itself. For this reason, the deepest and most sensitive intelligence is shared only with the closest allies, particularly the Five Eyes partners. Broader NATO sharing uses versions with secret sources and methods removed.

Historical Precedent: Cold War Monitoring

The current situation—reliance on unilateral intelligence collection to monitor Russian nuclear forces—has a historical precedent from the Cold War. The United States monitored Soviet nuclear forces from the 1950s through the early 1970s, relying entirely on reconnaissance satellites and signals intelligence collection.

Early reconnaissance programs had significant blind spots. Mobile weapons were particularly difficult to track. When the Soviet Union deployed mobile ICBMs, satellites could sometimes photograph them, but sustained monitoring of mobile systems was inconsistent. The Soviets conducted sophisticated deception operations, using mock-ups of missiles and decoy deployments to confuse analysis. Intelligence sometimes overestimated or underestimated Soviet capabilities because unilateral collection alone couldn’t provide absolute certainty.

The development of arms control treaties in the 1970s was driven partly by the recognition that unilateral monitoring, while valuable, had inherent limitations that created strategic instability. Each side tended to assume the worst about what the other side might be hiding. Mutual transparency was gone, so strategic planning became more conservative and potentially more destabilizing.

The SALT I treaty (1972) and subsequent agreements incorporated provisions for mutual verification, including on-site inspections and data exchanges, specifically designed to reduce the uncertainty that unilateral monitoring alone created.

Unilateral monitoring provides sufficient information for advance notice of major threats—intelligence could detect if the Soviet Union was preparing a major attack or conducting a major new weapons deployment. But it doesn’t provide the kind of detailed, routine, everyday confidence in understanding the other side’s forces that treaty-based verification does. The burden of conducting unilateral monitoring falls entirely on the monitoring state. The United States had to dedicate enormous intelligence resources to observing Soviet forces—no efficiency gains from cooperation in providing some information directly.

Resource Costs of Unilateral Monitoring

The expiration of New START has had tangible resource implications for intelligence. Maintaining the treaty verification regime required dedicated resources—teams trained in nuclear weapons verification, analysts who reviewed declarations and inspection data, and NRO resources devoted to monitoring specific facilities in ways that would support verification. The treaty provided efficiency in intelligence collection: because the U.S. knew that Russia was required by treaty to declare certain information, collectors could focus on verifying that information rather than searching for information Russia might be hiding.

Intelligence agencies must now dedicate more resources to compensating for lost verification data. More satellites must be tasked to observe nuclear facilities with the redundancy necessary to ensure persistent monitoring. More analysts must be assigned to synthesizing information from several sources to arrive at assessments that once came directly from declarations. More human intelligence resources must be devoted to acquiring information that was once available through official channels.

The intelligence budget isn’t unlimited. Resources devoted to monitoring Russian nuclear forces must come from somewhere. Either Congress must increase intelligence budgets to provide the funding required, or intelligence agencies must reduce collection efforts in other areas. Resources dedicated to monitoring Russia’s nuclear arsenal cannot simultaneously be devoted to monitoring China’s rapidly modernizing nuclear forces, tracking terrorist threats, or conducting spy-catching operations to protect national security.

The Confidence Gap in Intelligence Assessment

Prior to treaty suspension, intelligence assessments of warhead deployments were based on a combination of declared data, inspection findings, and unilateral collection. Each source provided different pieces of the puzzle. Declarations stated the exact number of deployed warheads at any given time. On-site inspections verified specific deployed missiles and warheads at specific locations. Satellite imagery tracked the movement of weapons systems. Signals intelligence monitored communications about deployment activities. Together, these sources allowed the intelligence community to construct a reasonably confident picture of how many weapons are deployed and where.

Now, the intelligence community must assess warhead deployments based primarily on satellite imagery showing which missile bases have launchers with weapons and inference about the likely deployment status of those weapons. If a satellite observes a deployed ICBM at a missile base, analysts infer that it’s deployed. But an ICBM sitting in a launcher could be deployed with a warhead, deployed with several warheads, or deployed with a dummy warhead used for testing. Distinguishing among these possibilities becomes an exercise in probabilistic reasoning.

The intelligence community maintains its assessment that Russia has remained generally within treaty limits even after the treaty’s expiration, but the confidence in that assessment has declined. Intelligence assessments are typically categorized by confidence level—high confidence, moderate confidence, low confidence. Assessments that once would have been “high confidence” based on confirming sources are now “moderate confidence” based on fewer sources.

A policy decision made on the basis of high-confidence intelligence that Russian forces are at a certain level will be different from a policy decision made on moderate-confidence intelligence. The uncertainty compounds when it affects military planning decisions—the more uncertain the intelligence, the more conservative the strategic assumptions must be to maintain stability.

Anti-Satellite Weapons: A New Vulnerability

The intelligence community’s capacity to monitor Russian nuclear forces through satellites faces a novel threat: anti-satellite weapons. In 2021, Russia conducted an anti-satellite missile test that shot straight up and successfully destroyed one of its own decommissioned satellites in low Earth orbit.

In March 2022, Russia launched what U.S. officials believe is an experimental satellite that could carry a nuclear warhead, representing a potential threat to reconnaissance satellites. These developments suggest that Russia may be developing the capability to disrupt satellite reconnaissance in future conflicts.

If Russia could successfully attack or disable reconnaissance satellites, the intelligence monitoring capability that compensates for treaty verification loss would be severely degraded. Intelligence agencies have developed multiple satellites watching the same areas for this reason—different wavelengths and at different times to ensure that the loss of one satellite doesn’t completely eliminate observation capability. However, the redundancy isn’t infinite. If Russia deployed several anti-satellite weapons simultaneously, space-based reconnaissance could be substantially degraded, particularly for monitoring nuclear force movements during a conflict.

With the treaty expired, satellites are more vulnerable to attack, precisely when they’re most critical to compensating for the loss of treaty-based verification.

Conclusion: Monitoring Without Verification

The loss of New START hasn’t created a complete intelligence failure. Intelligence agencies retain the technical capability to monitor major Russian nuclear force developments. Satellites continue to observe military installations. Signals intelligence collection continues to provide information about military activities. Human intelligence networks continue to report on developments within the military and government. Allies continue to share intelligence. The intelligence assessment that forces remain at roughly treaty-limited levels, though with reduced confidence, continues to guide strategic policy.

The loss of treaty-based verification has created a new normal characterized by greater uncertainty, higher resource requirements, and increased risk of miscalculation. Official Russian statements against which to measure unilateral collection findings are gone. Strategic planners must plan for worst-case scenarios about Russian nuclear capabilities more routinely than when treaty data provided a common reference point. The vulnerability of space-based reconnaissance to anti-satellite weapons has become more operationally significant now that satellites are the primary source of information about forces, rather than one source among several.

The expiration of New START has interrupted a 50-year tradition of mutual verification designed to reduce the risk of miscalculation and build confidence between nuclear adversaries. Negotiators who crafted arms control agreements after the Cuban Missile Crisis understood that nuclear war could result not from deliberate aggression but from miscalculation based on incomplete information about the other side’s capabilities and intentions. Treaty-based verification was designed to minimize that miscalculation risk. The return to reliance on unilateral intelligence collection, whatever its technical sophistication, represents a step backward in that direction.

A restored arms control framework that includes agreements where both sides can check on each other would provide the intelligence community with valuable information while reducing resource demands and allowing those resources to be devoted to other national security priorities. In the absence of such a framework, the intelligence community continues to develop new technical capabilities and to work with allies to maintain understanding of nuclear force developments. It’s a functional but less stable alternative to the transparency that a negotiated arms control agreement provides.

Our articles make government information more accessible. Please consult a qualified professional for financial, legal, or health advice specific to your circumstances.