I keep coming back to a blunt line that landed this week from Moscow: the talks that are not happening should have started yesterday. Not next quarter. Not after the next headline cycle. Yesterday. If you have spent any time watching how great-power bargains actually get made, that wording is not theater. It is a clock. The last treaty that put a hard ceiling on deployed American and Russian strategic nuclear forces expired in early February 2026, and nothing with legal teeth has taken its place. Markets shrug at nuclear diplomacy until they do not. I have found that the shrug is usually the expensive part.
There is a strange double track running through this war season. One track is creative, almost transactional: ideas about easing pressure in exchange for prisoner releases, quiet channels, trial balloons about what a thinner sanctions regime might buy. The other track is louder, sharper, and aimed at Europe. Those two tracks do not cancel each other. They sit on top of a third problem that almost nobody prices properly. The nuclear file, in the Kremlin’s own phrasing, remains completely unaddressed. That is not a slogan. It is a description of the legal landscape.
Why The Missing Treaty Is Not A Technical Footnote
For a decade and a half, the 2010 strategic arms reduction pact was the last standing rulebook between the two largest nuclear arsenals on earth. It was signed by an American president and a Russian president who are both long out of those offices. It capped deployed strategic warheads at 1,550 a side. It capped deployed delivery systems, the missiles, bombers, and submarines that actually carry those warheads, at 700. It also shaped how each side thought about the other’s political and military centers. That last piece is easy to skip in a briefing note. It is the piece that keeps a crisis from becoming a guessing game about what the other capital considers worth a launch.
When a ceiling like that disappears, nothing explodes on day one. Arsenals do not double overnight. Warheads do not walk out of storage by themselves. What changes is the permission structure. Planners who used to argue inside a box now argue about the box. Verification visits that used to be routine become political favors. Data exchanges that used to be dull become bargaining chips. I have sat through enough risk memos to know the boring machinery is the part that fails first, and it fails quietly.
A Kremlin spokesman put the practical problem in plain language. This is a complex matter. It needs lengthy, meticulous, expert-level negotiations. It will take a long time in any case. Read that twice. If the work takes years, and the work has not started, then the gap is not a pause. It is a deficit that compounds. You cannot sprint a treaty that was designed to be slow.
The nuclear issue remains completely unaddressed, and the talks that would address it should already be underway. Delay is not neutral. Delay is a choice about how much uncertainty both capitals are willing to live with.
Diplomatic reading of the current Moscow line
What Actually Expired, In Plain Numbers
People toss around the phrase strategic stability as if it were a mood. It used to be a spreadsheet. The old pact did three jobs at once. It limited warheads that were mated to delivery systems and counted as deployed. It limited the launchers and bombers themselves. And it forced a rhythm of notifications, inspections, and exhibitions so that each side had a recent picture of the other, not a rumor from a satellite pass and a leaked slide.
Those numbers still matter as a reference point, even if they are no longer law. Fifteen hundred and fifty deployed strategic warheads is not a small pile. It is enough to wreck the industrial core of a continent several times over. Seven hundred deployed delivery systems is not a museum count either. It is the fleet of intercontinental missiles, submarine-launched missiles, and heavy bombers that make the warhead number real. Without a shared definition of what counts, each side can tell a story that flatters its own restraint and exaggerates the other’s buildup. Stories are cheaper than inspections. They are also worse.
| Constraint | Old Treaty Ceiling | Status After Early 2026 |
| Deployed strategic warheads | 1,550 per side | No binding cap |
| Deployed delivery systems | 700 per side | No binding cap |
| Inspections and data swaps | Routine, treaty-mandated | Political, optional |
| Rules on political and military centers | Embedded in the framework | Unwritten, contested |
| China in the same deal | Not a party | Washington still wants inclusion |
Look at that last row before you romanticize a simple restart. Washington has said for years that a landmark deal which ignores Beijing is a deal that ages badly. Moscow has treated that demand as a way to stall, or as a genuine structural problem, depending on the week. Both readings can be true. China’s arsenal is smaller than the American or Russian stock, but it is no longer a rounding error, and it is growing. A two-player treaty in a three-player world was always going to creak. The creak became a break when renewal failed against the backdrop of the Ukraine war and a thick stack of sanctions.
The Alaska Conversation That Did Not Become A Treaty
In mid-August 2025, the American and Russian presidents met in Alaska with a crowded agenda. One of the quieter aims, according to the way the meeting was framed at the time, was to sketch next stages that could include agreements on control of strategic offensive weapons. That phrase is bureaucratic on purpose. It means the big missiles, the boats, the bombers. It does not mean a photo and a handshake.
Nothing binding came out of that room on the nuclear file. The old treaty still died on schedule the following February. I do not find that surprising, and I do not find it comforting. Summits are good at naming a problem. They are bad at counting warheads. Counting warheads is a job for people who argue about telemetry, bomber attribution, and what a launcher looks like when it is empty on purpose. Those people were not given a mandate that survived contact with the war.
Perhaps the most interesting aspect of the Alaska framing is how modest it was. Not a new grand bargain. Next stages. A door left ajar. Doors left ajar do not inspect missile bases. If the political layer cannot even authorize the expert layer, the expert layer does not magically appear because a spokesman says the work should have begun yesterday. The sentence is a complaint and a warning at the same time.
Two Tracks, One Unfinished File
The creative track is easy to underestimate. Ideas have circulated about sanctions relief tied to the release of political prisoners, about narrower deals that lower the temperature without pretending the war is over. That kind of bargaining is familiar. It treats pressure as a dial. Turn it down here, ask for something concrete there. It can work on hostages and bank access. It does not, by itself, rebuild a verification regime.
The louder track runs through European capitals. Rhetoric has hardened. Force posture talk has hardened with it. A document said to have been passed toward the alliance warned that Russia would be ready to use its entire arsenal, including nuclear weapons, to defend territory if there were an attempt to cut the Kaliningrad region off from the rest of the country. Treat that paper as alleged until it is authenticated. Even as an allegation, it tells you how the conversation is being framed: isolation of a geographic pocket is being described, by someone in that channel, as a nuclear problem, not only a conventional one.
Kaliningrad is a strange piece of map. It sits apart from the main body of Russian territory, bracketed by alliance members, home to forces that matter in any Baltic scenario. You do not need to accept every threat as a plan to see the negotiating consequence. If one side is willing to describe a regional isolation attempt as grounds for the full arsenal, then arms control is not a side conversation. It is the guardrail people reach for when maps get tight. Right now the guardrail is language, not law.
- A sanctions-for-prisoners idea can move fast and still leave warhead ceilings untouched.
- A summit can name strategic offensive weapons and still produce no inspection schedule.
- A regional threat, even an alleged one, raises the cost of having no shared definitions.
- A demand to include a third nuclear power can be both sincere and a delay tactic.
- Expert talks measured in years cannot begin in the month they are needed if nobody opens them.
What “Started Yesterday” Actually Asks For
When officials say talks should have started yesterday, they are not asking for a signing ceremony. They are asking for a room, a mandate, and a stack of technical questions that have answers only if both sides still remember how to ask them. The old treaty’s machinery was specific. Notifications when a missile moved. Short-notice inspections at declared sites. Exhibitions so that a new variant could be looked at rather than guessed at. Conversion rules so that a bomber taken out of the nuclear role was not still counted, or was counted differently, under an agreed method.
None of that restarts with a press statement. It restarts with agreed definitions. What is deployed. What is stored. What a heavy bomber counts as when it can carry both conventional and nuclear loads. How submarine-launched missiles are attributed when the boat is at sea and the warheads are not visible. These are dull questions. Dull questions are how you avoid interesting surprises.
In my experience, the side that says the process will be long is often the side that has already done the internal homework and wants the other side to admit the calendar. Moscow is signaling that it will not treat a revived ceiling as a weekend project. That can be read as seriousness. It can also be read as leverage: the longer the gap, the more each side can adjust its forces and then negotiate from the new baseline. Both readings should make a risk manager sit up.
The China Problem Is Real, And It Is Also A Trap
Washington’s long-standing line is that future landmark deals should include China. On the merits, that instinct is hard to dismiss. A ceiling that binds two players while a third builds without a matching cap is a ceiling with an exit ramp. Beijing has shown little appetite for joining a format designed around American and Russian force structures, stockpile histories, and verification habits. Forcing a three-way table before a two-way table exists is a good way to have zero tables.
There is a workable sequence that diplomats discuss in private and rarely put in a communique. Restore a bilateral floor first, even a thin one, so the two largest arsenals are not unbounded. In parallel, build a separate conversation with Beijing about transparency, missile notifications, and crisis hotlines that does not pretend to be a warhead treaty on day one. Sequence is not surrender. Sequence is how you avoid using a third capital as an excuse to do nothing with the second.
I am not convinced either capital has accepted that sequence. The American side can say, with some justice, that a deal which freezes a duopoly while China modernizes is strategically sloppy. The Russian side can say, with some justice, that holding the bilateral file hostage to a Chinese signature is a way to run out the clock. Sloppy and cynical can coexist. The arsenal does not care which adjective you prefer.
A treaty that waits for every nuclear power to sit down is a treaty that never starts. A treaty that ignores every other nuclear power ages into a fiction. The adult move is a floor now and a wider table later.
Verification Is The Part Amateurs Skip
Anyone can announce a number. The old pact’s value was that the number could be checked by people who did not trust each other. Trust is a bad operating system for nuclear forces. Procedures are better. When procedures lapse, each side falls back on national technical means, the satellites and sensors that see a lot and still miss the things that happen under a roof. National technical means were always part of the picture. They were never the whole picture.
Without inspections, a launcher can be declared converted and still be convertible back. A warhead can sit in a storage site that used to be visited and now is not. A new missile can fly in a test that both sides see, while the production rate stays opaque. Opacity is not the same as cheating. Opacity is the condition in which cheating cannot be distinguished from caution. That distinction is the whole product arms control sells.
Restarting inspections after a gap is harder than maintaining them. Teams lose clearances, language skills, site familiarity. Host countries lose the habit of opening gates. Political supervisors rediscover how awkward it is to let the other side’s officers walk a base. If the expert talks the Kremlin described as lengthy have not even been scheduled, the inspection gap is already training both systems to live without each other. Habits formed in a gap do not unwind because a communique sounds hopeful.
How A Legal Vacuum Shows Up In A Crisis
Imagine a bad week in the Baltic, or a strike that lands closer to a strategic asset than either side intended, or a satellite anomaly that looks like a prelude. In the old framework, there was at least a shared vocabulary for what each side had deployed and a channel that was not invented during the panic. In the current framework, the vocabulary is national. Each military brief will use its own counting rules. Each political brief will treat the other’s silence as concealment.
That is how misread signals get expensive. A bomber exercise that used to be notified becomes a story about a changed alert status. A submarine leaving port on a normal cycle becomes a story about a surge. None of this requires either side to want a nuclear crisis. It requires both sides to lack a recent, agreed picture. Want is overrated as an explanation. Friction is underrated.
The targeting logic of the old treaty is the piece I wish more market notes mentioned. The pact was built, in part, to regulate how rivals thought about each other’s political and military centers. When that regulation lapses, worst-case targeting slides back toward assumptions from the 1980s, updated with better missiles and worse trust. You do not need a new doctrine document for that slide to happen. You need an empty calendar.
Crisis friction, stripped down: No shared count -> competing briefings No notifications -> exercises look like surges No inspections -> storage looks like uploading No hotline habit -> the first call is the hard call
Sanctions, Prisoners, And The Wrong File
There is a temptation, especially in Washington, to fold nuclear restraint into a wider package: relief on a sanctions line, movement on detainees, a quieter channel about the war. Packages can unlock things that pure arms talks cannot. They can also bury the arms talks under issues that move on a different clock. A prisoner release can be arranged in weeks. A warhead definition takes seasons. If the package is judged a success because the visible trade cleared, the invisible file can stay blank and still be called progress.
I would rather see the nuclear channel separated enough that it cannot be held hostage to the week’s bargaining, and linked enough that neither side treats it as optional charity. That is a narrow path. It is also the path that survived, in weaker form, through colder decades than this one. The alternative is a world in which every nuclear conversation waits for a goodwill gesture that the war keeps postponing.
Europe sits in an awkward chair in all of this. European governments do not hold the American or Russian strategic arsenals, yet European cities sit inside the flight times those arsenals imply. Hardened rhetoric from Moscow toward the alliance, and hardened posture talk in return, raises the premium on a US-Russia ceiling that European capitals cannot sign. They can, however, stop treating the ceiling as an American hobby. A continent that talks daily about deterrence and rarely about the missing treaty is arguing with one hand.
What A Thin Restart Could Look Like
Nobody serious expects the 2010 text to be photocopied and re-signed. The politics will not allow it, and the force structures have drifted. A thin restart is a different animal. It would freeze deployed strategic warheads near the old ceiling for a defined period, restore a minimum of data exchanges, and schedule a small number of inspections so the muscle memory returns. It would punt the China question into a parallel track with its own metrics, so that track can fail without killing the floor.
A thin deal would also have to say something, even briefly, about new systems that did not dominate the 2010 text: novel delivery vehicles, dual-capable missiles that blur the nuclear line, and the temptation to upload stored warheads onto existing launchers once the cap is gone. Uploading is the sleeper risk. You do not need new factories to raise the deployed number if the launchers are already built and the warheads are already in the warehouse. Caps existed largely to make uploading a violation rather than a Tuesday.
- Authorize expert teams with a written mandate, not a summit adjective.
- Freeze deployed strategic totals near the last agreed ceilings while talks run.
- Restore data exchanges before arguing about the perfect inspection quota.
- Separate a bilateral floor from any trilateral ambition.
- Write an explicit rule on uploading stored warheads onto existing launchers.
- Keep a crisis channel staffed by people who already know the other side’s officers.
Would Moscow accept a freeze that locks in a baseline it might prefer to revise? Would Washington accept a bilateral floor while still saying China must eventually sit down? Those are political questions, and they are the reason a spokesman can say the work takes a long time without being cynical. Long is fine. Unstarted is not. The phrase that should have begun yesterday is really a claim about sequence: the calendar was already late when the treaty died, and it is later now.
Markets, Insurance, And The Risk Nobody Puts On A Slide
Financial markets do not trade warhead ceilings directly. They trade the consequences of a world in which those ceilings are political again. Energy routes, shipping insurance, defense procurement, currency safe havens, the premium investors demand for European assets when rhetoric jumps. Most of the time the premium is a rumor tax. Occasionally it is a repricing. The absence of a treaty does not cause a selloff on a quiet Thursday. It removes one reason to believe a bad Thursday stays conventional.
Defense budgets in Europe have already moved for reasons that have nothing to do with New START’s obituary. Still, an unbounded strategic relationship between Washington and Moscow changes the story those budgets are buying. More conventional depth makes sense if the nuclear background is stable. If the nuclear background is a argument, conventional spending is racing a different risk. I have found that investors who only model the visible war, the front line, the sanction waiver, the grain corridor, miss the layer that makes all of those local.
There is also a quieter balance-sheet issue inside the nuclear enterprises themselves. When caps lapse, industrial planning shifts from maintenance under a ceiling to options under no ceiling. Options cost money. They show up as longer production runs, delayed retirements, and modernization programs that no longer have to justify themselves against a treaty number. Those programs are slow. Their political half-life is long. A negotiation that starts two years late is negotiating with hardware that has already been ordered.
Rhetoric Versus Posture, And Why Both Matter
It is fashionable to dismiss nuclear statements as noise. Some of them are. A threat issued for a domestic audience is not a launch order. The trouble is that posture follows rhetoric with a lag, and posture is what the other side’s satellites see. If officials describe isolation of a detached region as a scenario for the entire arsenal, planners on the other side will ask what forces would make that description credible. Credible descriptions attract counter-plans. Counter-plans attract budgets. Budgets outlive the press cycle that sparked them.
Arms control, at its least romantic, is a way to interrupt that loop with a number both sides have checked. Without the number, the loop runs on national estimates. National estimates are not neutral. They are briefed by institutions that get resources when the estimate is grim. That incentive does not require villains. It requires a vacuum. The vacuum opened when the treaty lapsed and stayed open when the follow-on talks failed to start.
So when Moscow says the nuclear issue is completely unaddressed, the useful response is not a competing slogan. It is a date. A venue. A scope that is smaller than a new world order and larger than a photo. Anything else is commentary on a gap.
What Each Capital Is Probably Optimizing
Washington wants a deal it can defend at home, which means a deal that does not look like a gift to a wartime adversary and does not ignore Beijing. That political constraint is real. It also produces a bias toward waiting for a cleaner moment. Cleaner moments are rare in the middle of a European war. Waiting for one is how treaties die of old age and do not get replaced.
Moscow wants relief from the sense that it is negotiating from under a sanctions boot, and it wants any new ceiling to reflect systems it has fielded since 2010. It also wants the political win of saying the other side let the last treaty lapse. That win is already in hand. Spending it on a delay is a choice, not a fate. A spokesman who says the talks should have started yesterday is, whether he intends it or not, conceding that the delay has a cost for Russia too. Unbounded American uploading is not a Russian victory.
Beijing, absent from the table, optimizes by not joining a format that would cap a buildup it still sees as catch-up. The longer Washington insists on a three-way signature as the price of a two-way floor, the more room Beijing has to modernize outside the argument. That is the trap inside the principle. Principles that cannot be sequenced become cover.
A Practical Reading For Anyone Who Allocates Risk
You do not need to become a treaty lawyer to use this. Treat the absence of a cap as a structural input, not a headline. Ask, in any scenario note about Europe or energy or defense names, whether the scenario still holds if deployed strategic numbers drift up without a shared count. Ask whether a crisis channel exists that was exercised this year, or only inherited from a treaty that no longer binds. Ask whether a sanctions concession is being mistaken for a stability concession.
The answers will often be unsatisfying. Unsatisfying is information. A world in which the two biggest arsenals are managed by press statements and national satellites is a world with fatter tails. Fatter tails do not show up in a quarterly earnings line. They show up in the price of insurance, in the bid for safe collateral, and in the day a routine exercise is briefed as something else.
Stability check: shared ceiling + recent data + practiced channel = manageable friction. Remove any one term and the product is no longer stability. It is hope.The Calendar Is The Argument
Experts on both sides already know the outline of a successor. They know which systems are easy to count and which are political. They know that inspections take months to restart even after a political yes. They know a Chinese seat will not appear because an American statement demands it. What they do not have is permission to spend the months. Permission is the scarce asset. Everything else is craft.
That is why the yesterday line sticks with me. It is impatient in a file that is supposed to be patient. Impatience from the side that usually plays for time is a tell. It suggests the gap is no longer convenient, or no longer only convenient. Convenient gaps get extended. Inconvenient gaps get speeches. Speeches are not teams. Teams are what yesterday was asking for.
If a thin freeze and a restored data exchange began this season, the first inspections would still land late. Definitions would still be fought over. Domestic critics on both sides would still call it naive or one-sided. All of that is normal. Normal is available. What is not available is more free calendar. The treaty died in February. The summit that was supposed to sketch a sequel happened the summer before. The sequel was not sketched into a mandate. The spokesman is now describing a late start as if lateness were the news. Lateness is the record.
Where This Leaves The Rest Of Us
None of this requires you to pick a side in the war to see the hole in the floor. A cap is not a favor to Moscow. A cap is not a favor to Washington. A cap is a constraint both sides accept because unbounded competition in strategic forces is a bad business for both, and a worse business for everyone who lives inside the flight corridors. Europe hears the rhetoric. Markets hear the insurance quotes. The arsenals hear neither. They follow orders and budgets.
I keep the old numbers handy not out of nostalgia. Fifteen hundred and fifty. Seven hundred. They were compromises, and compromises age. They were also checked. An unchecked compromise is just a claim. We are living, for the moment, on claims. The creative diplomacy around prisoners and sanctions may yet produce a headline worth having. It will not, by itself, put a lock back on the warehouse. That lock is a separate job, it is late, and the people who would do it have been told, in public, that they should already be in the room.
Perhaps the talks open next month and this note ages into a complaint about timing. I would take that outcome. A late room is better than no room. Until the mandate exists, though, the accurate description is the one Moscow already used. The nuclear issue is unaddressed. The work is long. The start date was yesterday. Everything after that is a choice about how wide the gap is allowed to get, and who pays the premium while it does.
