Photo: Andrew Harnik/Getty Images
U.S. President Donald Trump is right. The war in Ukraine must end. The challenge is determining how to set conditions for a lasting peace deal in a way that doesn’t further destabilize Europe or create openings for additional Russian aggression.
This installment of Critical Questions reviews findings published in the CSIS Futures Lab Strategic Headwinds series over the last year on the opportunities and challenges to negotiating—and implementing—a peace deal in Ukraine. This series integrates expert surveys and analysis with an AI model trained on historic peace negotiations to provide perspectives on the prospects for peace and ending the war in Ukraine. Drawing on CSIS Futures Lab survey data across Ukraine war experts, this analysis maps where the plan’s provisions intersect with issues previously identified as plausible negotiation articles.
The only way Trump’s Ukraine peace plan can make any meaningful progress is if it serves as an opening gambit to start a larger peace negotiation with Russian President Vladimir Putin. The plan’s limited security guarantees should change to include third-party guarantees that extend beyond economic sanctions. Ukraine can survive not being in NATO. But it cannot survive as a sovereign state with a hollowed-out military and limits on the ability of European states to deter future Russian aggression.
Q1: Why do most peace deals tend to fail?
A1: Peace deals tend to fail because of commitment problems: Neither side can know if the other party won’t take advantage of the pause in fighting to regroup and attack them in the future. Kyiv cannot know if Moscow really wants peace or is just waiting for a more favorable future moment to seize additional land by force. Based on historical analysis, 31 percent of interstate wars end in a stalemate under ceasefire agreements, creating this exact type of commitment challenge. The fighting stops, but the underlying disputes remain, and trust is low.
Commitment problems are even more acute when one of the parties has a history of breaking deals. Russia has a track record of breaking past agreements, including the Minsk I Protocol in 2014, and has even used peace talks as a way to position forces on the battlefield rather than to seek long-term peace. From a Ukrainian perspective, the Minsk II agreement only slowed fighting to the contact line in the Donbas while setting the stage for the 2022 invasion.
Domestic politics and economic shocks common in post-conflict settings tend to compound fears. Bad deals tend to produce political crises and can even topple democratically elected governments, paving the way for instability and future conflict.
Q2: How might Trump’s Ukraine peace plan succeed?
A2: It’s all about economics. Human experts see the most promising parts of the Trump plan in the “pragmatic” space rather than the grand political trade. They think Trump’s ideas around reconstruction and investment in Ukraine have real traction, especially letting Western companies help rebuild infrastructure and energy systems and using Ukraine’s gas transit network as part of a postwar European energy regime, which all sides rate as fairly negotiable. By offering a clear pathway to economic reintegration and G8 membership, the proposal effectively removes the economic levers the West has used to constrain the Kremlin.
Strategically, this creates a high probability that Russia would agree to the initial terms, as the peace plan grants them a strategic victory without further military cost. Experts rate “using Ukraine’s gas transit infrastructure for European energy supplies from Russia” as both highly negotiable and reasonably satisfying for all four sides. Those are the parts of Trump’s package most likely to survive in any serious negotiation—because they sit in the narrow band where expert expectations for mutual gain actually converge.
Security guarantees are more complicated. Human experts see them as critical to any lasting deal. The Trump peace plan includes limited guarantees at best. Earlier expert surveys and studies in Strategic Headwinds call for robust commitments, including air as well as cyber and space support alongside foreign peacekeepers. Trump’s plan limits these measures and even calls for rolling back NATO aircraft in Poland. Rather, the “security guarantee” is tied to economic sanctions and removing de facto recognition of Russian-occupied parts of Ukraine.
Rerunning the same expert survey, albeit with a large language model as the judge, paints a similar “success core,” but is more bullish on some elements and more skeptical on others. AI models rate Trump’s humanitarian plank extremely highly, treating the return of deported Ukrainians and children as both very negotiable and very satisfying for Ukraine, Europe, and the United States. Algorithmic judges also see demilitarized zones and peacekeeping arrangements as one of the safest and most workable security devices in a peace package, which again contrasts with Trump’s peace plan. Like human experts, the AI survey is very positive about large-scale Western investment in Ukrainian resources and infrastructure.
Where AI differs from humans is that it is much colder on sanctions relief and returning frozen Russian assets, and much warmer on keeping sanctions tight and using them as enforcement, so it expects Trump’s peace plan to succeed most where it leans into reconstruction, gas and infrastructure cooperation, robust guarantees and humanitarian deals, rather than big gifts to Moscow on territory or sanctions.
Figure 1 shows experts’ assessment of positions of parties across negotiability and satisfaction for the subset of negotiation articles that resembles most of Trump’s peace plan. With this subset of articles, Ukraine’s negotiability space appears to be very low. More detailed discussions of methodology can be found at Strategic Headwinds.
Q3: How might Trump’s Ukraine peace plan fail?
A3: The real killer for the peace plan’s success, however, is the cluster of provisions on territory, Ukrainian military power, and amnesty. CSIS expert data identifies the specific mechanisms where the peace plan is statistically most likely to collapse: the imposition of severe military restrictions and territorial concessions on Ukraine. The proposal to cap the Armed Forces of Ukraine personnel size is viewed by experts as a nonstarter, with Ukrainian satisfaction ratings plummeting to a near-absolute zero and negotiability hovering at a very low level. However, the peace plan refers to a number of 600,000 Ukrainian troops. If this is the number of active forces, it could still put Ukraine in the top 10 militaries by size in the world.
Furthermore, the requirement to recognize Crimea and the Donbas as Russian territory—while freezing the lines in Zaporizhzhia and Kherson—demands a significant political concession from the Ukrainian leadership that the current government is unlikely to survive. Trump’s plan appears to systematically choose the Russian-maximizing version of each of these trade-offs. These are effectively impossible for a sovereign Ukraine to accept voluntarily. In probabilistic terms, this is where the package is least likely to be accepted absent a dramatic Ukrainian military collapse or a major fracture in the Western coalition. Trump’s plan creates both an international and domestic commitment issue for Ukrainian President Volodymyr Zelensky that makes it untenable.
Finally, the plan’s justice and governance provisions cut against expert expectations for a sustainable settlement. The Futures Lab’s survey indicates strong support—especially in Ukraine and Europe—for “demanding accountability for Russian war crimes,” and very high satisfaction with the “return of forcibly deported Ukrainians, including kidnapped children.” Trump’s plan adopts the latter but explicitly rejects the former by granting full amnesty and waiving future claims for all parties, landing squarely on the side that experts view as most satisfying for Russia and least satisfying for Ukraine. Early elections within 100 days, another Trump demand, show low satisfaction and only middling negotiability across the board in the data, suggesting they are more likely to destabilize than legitimize the settlement.
The AI-enabled survey largely converges with expert assessments on where Trump’s plan is likely to fail. It rates recognition of Russian control over occupied territories, including freezing the line of contact in Kherson and Zaporizhzhia, as deeply unacceptable in Europe, the United States, and especially Ukraine, making these clauses prime candidates to break any deal. The models also see strict limits on the Ukrainian military and a blanket amnesty as planting the seeds of renewed violence: capping Ukraine’s forces and constraining its weapons lock in a power imbalance that rewards maximalist aims in Moscow, while erasing accountability runs directly against the strong AI preference for war‑crimes trials and undermines the legitimacy of any settlement. Finally, although the survey finds more room for bargaining over NATO status than human experts do, it still treats a security architecture built mainly on economic sanctions and political denunciations, rather than robust defense commitments and a capable Ukrainian army, as an unstable foundation that is unlikely to deter future aggression.
Q4: Is this an end plan?
A4: Trump’s plan is the start, not the end, of negotiations. The president sees himself as a dealmaker and likely views ending the war in Ukraine as part of a larger grand bargain. Therefore, the peace plan is the first step in bringing Russia to the table. What should happen now is that diplomats from across Europe engage with the White House to map out how best to compromise with Putin, as morally repugnant as that sounds. The fact is, 60 percent of all wars conclude through some form of compromise, especially when there is no clear path to battlefield victory. Unless there is a fundamental change on the battlefield, Ukraine is likely to lose territory.
Benjamin Jensen is director of the Futures Lab and a senior fellow for the Defense and Security Department at the Center for Strategic and International Studies (CSIS) in Washington, D.C. Yasir Atalan is the deputy director and data fellow in the Futures Lab at CSIS.
Critical Questions by Benjamin Jensen and Yasir Atalan — October 17, 2025
Critical Questions by Benjamin Jensen — August 20, 2025
Commentary by Yasir Atalan, Erik Tiersten-Nyman, and Benjamin Jensen — November 14, 2025
