By late August 2022, the West’s focus on Russia’s war in Ukraine was diminishing. The two sides were bogged down in an extended stalemate, freeing Western leaders from making difficult choices or thinking too hard about the future of the conflict. Events since early September—dramatic Ukrainian gains, followed by Russian mobilization, annexations, missile attacks on civilian areas, and nuclear threats—have shattered that illusion, pushing the war into a new and more dangerous phase.

Since the start of the war, the Biden administration has effectively maintained a balanced realpolitik approach: arming and funding Ukraine yet continuing to make clear that the United States will not engage directly in the conflict. But the administration has avoided talking about one crucial area of war strategy altogether: how it might end. Experts and policymakers who have suggested that the United States should also support diplomatic efforts aimed at a negotiated settlement have been treated as naïve or borderline treasonous. Driving the administration’s skittishness about endgames, then, are questions of morality: many argue that it is immoral to push Ukraine toward a settlement.

But nearly all wars end in negotiations. Moscow’s escalation this fall raises the twin specters of a broader war with NATO and of the use of nuclear weapons. The global economic costs of the conflict are already enormous and will almost certainly increase with the onset of winter. Even if a negotiated end to the war seems impossible today, the Biden administration should begin to raise—both publicly and to its partners—the difficult questions that such an approach would entail. It must think through the right timing to push for negotiations and at what point the costs of continuing to fight will outweigh the benefits. In seeking a sustainable settlement, the administration must also figure out how to capitalize on Ukraine’s successes without setting the stage for further conflict. To prepare for the best deal, American policymakers must maintain a common front between the West and Ukraine, take account of Ukrainian and Russian domestic politics, and embrace flexibility, particularly in working out which sanctions against Russia can be lifted without strengthening Putin’s regime. If the administration does not prepare soon, it may find its carefully calibrated response to the war being overtaken by a dangerous fantasy of absolute victory.

NOT IF BUT HOW

In the eight months since the Russian invasion, the Biden administration’s support has allowed Ukraine to retake territory and inflict heavy damage on Russian forces while keeping the risk of large-scale escalation relatively low. The administration has also carefully avoided talking about what comes next, claiming that it is up to Ukrainians to decide what is in their best interest. But maintaining that position is becoming more difficult now that Russian President Vladimir Putin has doubled down on the war and made blatant nuclear threats against the West. Putin has chosen to take significant new risks rather than to back down, suggesting that this war will not end through simple Russian capitulation. Though these risks seem manageable for now, the time may come when negotiations are necessary to forestall catastrophe.

At the same time, the economic fallout of the war is rapidly growing. In Ukraine, public finances have been ravaged; the country is running out of cash. As the economic historian Adam Tooze put it in September, “Unless Ukraine’s allies step up their financial assistance, there is every reason to fear both a social and a political crisis on the home front.” Europe, meanwhile, is trapped in its own tightening noose, as surging energy prices exacerbate inflation and raise the prospect of a deep recession. All this makes the administration’s position—that Kyiv alone will decide when the war ends—increasingly untenable.

In reality, the question is not whether negotiations are needed to end the war but when and how they should unfold. Yet policymakers must contend with a Catch-22: the better Ukrainian forces perform on the battlefield, the more difficult it is to discuss a negotiated settlement, even though it is to Ukraine’s advantage to negotiate from a position of strength. As the risk of Russian escalation grows, so does the prospect that any Western leader who talks about ending the war will be portrayed as unrealistic, immoral, or caving to “nuclear blackmail.” But internal discussions on acceptable settlement terms now would better position all parties when the opportunity for such a deal does arise.

DURABLE, NOT MAXIMAL

To lay the groundwork for a settlement, American policymakers must act to ensure that American, European, and Ukrainian interests do not diverge. Ukrainian interests are not necessarily identical to those of its Western partners. For Kyiv, the stakes are higher, and—with the Ukrainian economy already in a shambles—it may determine that it has little to lose in risking escalation or continuing the war. But Ukraine’s efforts are made possible by Western arms, funding, and intelligence. European states are bearing substantial economic costs from the war. And any risk of escalation or nuclear exchange poses a direct threat to the West itself. Ukraine’s Western backers have a strong stake in the war; they should have a say in how it ends.

This does not mean that the West should push Ukraine to concede, as some have argued. But it does suggest that the United States and its partners should provide future aid with an eye to putting Ukraine in the best negotiating position, not simply continuing the war. For example, Ukraine and its allies must focus on core interests, such as preserving Ukraine’s sovereignty and protecting its population. These goals should be narrow by design: rather than trying to retake all its pre-2014 territory or to punish Russian leaders, Ukraine should pursue objectives that are less likely to produce dramatic escalation and more likely to result in a durable peace. Washington should encourage pursuing those objectives, and should also make clear to Kyiv, at least in private, where the limits of American support lie and what the White House perceives as unacceptable escalation risks. Setting clear expectations now reduces the risk of misperception in Kyiv.

American policymakers must also consider Ukrainian and Russian domestic politics, since internal support in both countries will be vital to making any settlement last. History suggests that a power transition in Moscow is possible but by no means likely or inevitable. Thus, policymakers need to focus on Putin and the small group of elites around him and consider what settlement they might be willing to accept. Given Putin’s mobilization of several hundred thousand additional frontline troops, it seems increasingly clear that he will seek to avoid a complete, devastating loss at any cost. But like many other authoritarians before him, he can sell a poor result as a win. This means that it may be possible to find some face-saving deal in which de facto realities, such as Russian legal control of Crimea, could be recognized, and which the Kremlin could portray to the Russian public as genuine concessions by the West.

In Kyiv, Ukrainian President Volodymyr Zelensky faces a more open, contentious political environment, as the divisions of Ukrainian politics begin to reemerge. Nonetheless, he faces a similar dilemma. Ukraine’s population has become more unified since February, rallying around a national struggle against invasion. After asking so much of Ukrainians, the Ukrainian government will find it difficult to compromise in any way that seems to reward the enemy. If Zelensky accepts an unpopular settlement, it could lead to his defeat at the ballot box. In these circumstances, a deal in which Ukrainians feel that they have largely triumphed is more likely to succeed. This makes it all the more important to manage expectations now. Washington should encourage Kyiv to take a more moderate stance on issues, such as Crimea, that are likely to figure in a future settlement; to tone down triumphalist talk; and to emphasize the economic rewards that Ukraine stands to receive through international reconstruction aid and European economic integration under a settlement.

Policymakers should set clear basic parameters for a settlement but have significant flexibility in many details. A few points are nonnegotiable. Paramount among them are Ukraine’s sovereignty and the protection of Ukrainian citizens, particularly those who wish to leave Russian occupied territory. But there are other issues on which flexibility is possible. Final territorial borders, for example, may be determined partly by military gains on the ground. Policymakers should not be irrevocably devoted to the pre-February 24, or even pre-2014, status quo. A more territorially compact Ukraine, shorn of Crimea and some of the Donbas—both of which retain some pro-Russian populations—might be more stable and defensible.

And in general, policymakers should seek to prioritize practical outcomes over abstract principles. An independent sovereign Ukraine that can defend itself and integrate economically with Europe, for example, would be far preferable to a Ukraine with permanent territorial disputes within its borders. The situation in Ukraine remains dynamic; U.S. policymakers should avoid tying their own hands now with declarations that may be hard to achieve in practice.

Meanwhile, sanctions relief is likely to be one of the most important but politically fraught parts of any negotiation for Western policymakers. Sanctions tend to turn into permanent features of international politics, even though their economic and political impact weakens over time. They are therefore often more useful as bargaining chips than as permanent punishment. Policymakers should think carefully now about how to use sanctions relief to obtain Russian concessions. Throughout the war, Western sanctions have served two goals: short-term punishment for Russia’s invasion of Ukraine and long-term weakening of Russia’s military machine.

Some sanctions relief for Russia will likely be a prerequisite for a successful peace deal, but policymakers should think carefully about which sanctions are worth lifting. Allowing Russia to repatriate some of its foreign exchange reserves, for example, might be helpful as part of a deal because it is attractive to the Kremlin for short-term economic stabilization and because keeping these reserves frozen does little to weaken Russia’s economy over the long term. In contrast, export controls imposed on Russia should serve to limit the country’s defense industrial base in the longer term; they should be maintained, if possible. Policymakers should also embrace thorough plans for phasing, in which Russia concretely concedes or withdraws in exchange for gradual sanctions relief—something that was notably absent from the failed Minsk agreements.

KNOW WHEN TO HOLD ’EM

There are three circumstances in which it may make sense for the United States to push for a settlement. The first is if Ukrainian forces continue to achieve significant success and leadership in Kyiv begins to talk about liberating Crimea. Given the importance of Crimea to Russian leaders, such a goal substantially raises the risk that Putin will resort to the use of nuclear weapons, damaging norms against nuclear use and directly endangering the United States or—more likely—its NATO allies. The second is if Russian forces regain the initiative and retake significant territory, particularly if they begin to advance out of the Donbas. This would suggest that Russian mobilization has worked and that a settlement may be necessary to maintain Ukraine’s sovereignty. The third is if the two sides become locked in another stalemate, with neither able to regain the advantage. In such a situation, Europe and the United States and even Russia and Ukraine may conclude that it is no longer worth bearing the substantial costs of continuing the war.

At first glance, it may seem strange that U.S. policymakers should consider a settlement when Ukraine is winning, when it is losing, and when it is doing neither. And each of the situations outlined above would likely produce wildly different settlements. But what connects all three is that, in each, battlefield outcomes point to a relative consensus around which a settlement could be built. Today, the battlefield is still dynamic; both parties think they are going to triumph. A settlement will become possible only when the outcome on the battlefield becomes more apparent. Until then, robust Western support can help ensure that the first of these scenarios is the most likely.

Recent airstrikes against Kyiv and other major Ukrainian cities suggest that Russia may be contemplating greater escalation. There are substantial risks and costs to a widening conflict; even if it is not yet the time to negotiate, policymakers need to explore now the circumstances under which the United States would push for an end to the war. They should think through how to effectively leverage sanctions and battlefield gains to put Ukraine in the best position at the negotiating table. And perhaps most important, policymakers in Washington should communicate the results of these discussions to Kyiv and to European capitals in order to avoid potentially dangerous divergences in national interest among Ukraine and its Western partners.

All wars end. By raising now the crucial questions that will need to be addressed in the case of Russia’s war in Ukraine, policymakers can guard against unwanted escalation and ensure a more robust and stable settlement when the time finally comes. Though a settlement may seem unpalatable now, only by astute and careful negotiation are Ukraine’s core interests—and the security of the region—likely to be protected for the long term.