After Donald Trump announced earlier this month that he would consider holding a spring summit with North Korean leader Kim Jong Un, there has been a flurry of debate over what the president should seek from the potential meeting. On one end of the spectrum is the popular notion of denuclearizing North Korea, which usually means complete, verifiable, and irreversible dismantlement of North Korea’s nuclear capabilities, or CVID. While professing nominal commitment to this goal, it appears Kim is conditioning it on such formidable requirements that it is extremely unlikely that the regime will actually pursue this in any meaningful timeframe, no matter how hard the United States sanctions, threatens, or incentivizes them. Kim believes it would be suicidal to give up his “existential” deterrent, so complete denuclearization is simply not on the table today.

Even if it was negotiable in the near-term, CVID is based on an outdated understanding of North Korea’s nuclear enterprise. When the U.S. government developed the CVID concept in the mid-2000s, North Korea had conducted just one nuclear explosion test and its long-range ballistic missile program was still in its infancy. North Korea’s technical progress over the intervening decade—five additional nuclear tests and dozens of missile flights—means that a more sophisticated and intrusive approach to rolling back its dangerous capabilities is needed.

On the other end of the spectrum, and what North Korea might accept following a summit, is a simple temporary suspension of nuclear and missile flight testing, as Russia has suggested, for which Kim would still demand some sanctions relief or other incentive. But the Trump administration would immediately reject such a minimalistic concession. After all, Pyongyang’s unchecked arsenal is already worrisome, and it can continue to grow and improve without full-scale tests.

So if CVID is non-negotiable and a suspension is not in itself a satisfactory waypoint, what would be an approach that would allow Washington to pursue a highly ambitious but feasible strategic objective should the Trump-Kim summit produce momentum for serious negotiations?

China’s more ambitious concept of a nuclear freeze could be a starting point. Such a freeze may help build confidence during negotiations that North Korea is indeed willing to contemplate deeper limitations on its nuclear arsenal and infrastructure to stabilize the situation, pending full denuclearization.

But what the Trump administration should set as a strategic objective for negotiations is a comprehensive and verified capping of North Korea’s threatening strategic capabilities and activities. A broad cap could serve the medium-term interests of the United States, and its two allies, Japan and South Korea, while also finding acceptance in China and North Korea.

Capping means imposing significant, verifiable qualitative and quantitative limits on further development of North Korea’s nuclear weapons and their delivery vehicles. This would include curtailing production of the key bomb fuels plutonium and enriched uranium. And it would cover development and construction of additional critical capabilities and activities, including delivery vehicles; long-range ballistic missiles and related components; and weapons research, development, and engineering.

Additionally, to minimize North Korea’s capacity to use nuclear or conventional force offensively, militarization of nuclear forces must be extremely tightly constrained. Under a capping agreement this means ceasing activities to upgrade, deploy, increase readiness, and improve survivability of nuclear forces.

The price of such capping is implicit acknowledgment of the reality that Pyongyang will retain nuclear weapons while the agreement is being implemented. But, this price is worth paying in order to inhibit North Korea’s further militarization into a fully-fledged, combat-ready arsenal that can target the United States, especially considering that that capability is very close at hand.

A successful capping deal must also occur alongside stricter implementation of existing broad-based UN Security Council sanctions. This means more effective monitoring and enforcement of the import of banned equipment and material of proliferation concern. One means of achieving this would be to channel North Korean trade through a limited number of agreed ports in the region. Just as important, the same arrangement could also serve to verify that North Korea is not conducting covert nuclear or missile testing offshore, exporting those items to other countries, or generating illicit hard currency to sustain the program.

Because full denuclearization of the Korean Peninsula is not possible in the near-term, this cap should be open ended,  designed to last from now until such a future is possible. It would serve the ultimate purpose of denuclearization by limiting the size and sophistication of North Korea’s nuclear arsenal through increasingly irreversible steps, and by providing transparency over a greater range of North Korean nuclear and missile infrastructure that would facilitate their eventual elimination. And over time it would create opportunities to redirect North Korean scientists and technicians away from weapons to peaceful work.

There are, of course, several ways in which Pyongyang could game a cap. It could claim to have more capabilities than it in fact does, in order to have them grandfathered under a cap. Or, as in line with its past behavior, it could try to conceal elements of its programs in order to keep options open in case the United States threatens to renege on the deal or to attack. In anticipation of such actions, Pyongyang would have to clarify and allow for verification of the capabilities it has achieved in the categories of activities mentioned earlier at the very outset of such a deal. This would provide a baseline for future monitoring and define additional capabilities that North Korea would have to agree not to pursue, while offering useful indicators of cheating if it occurs. This is why verification must be thorough and comprehensive. Capabilities that North Korea will neither admit to having nor subject to transparency and verification should not be accepted. Any undeclared activity would constitute a violation of the deal. Together, these diverse transparency requirements and the restrictions under a capping deal would increase the likelihood of detecting violations.

At this moment, the Trump administration must face the reality that no past administration has been able to prevent North Korea from becoming the nuclear state that it is now. Complete denuclearization is not possible in the meaningful future, and therefore, an immediate suspension of the most worrisome developments followed by a comprehensive verifiable cap is the best realistic option for negotiating a deal with North Korea. It would improve the security of the United States and its allies, deescalate tensions, and provide a tolerable arrangement with North Korea regardless whether full denuclearization is feasible in our lifetime. Ideally  both parties in Washington would embrace this serious, ambitious, but potentially realistic objective halt escalation toward confrontation, which would also reassure South Korea and Japan of America’s prudence and steadfastness. Pursuing this course would also deprive Kim Jung Un of opportunity to split the United States from its allies and weaken international enforcement of the sanctions that have brought him to the table.

Securing such an agreement is not going to be easy and will obviously require not just considerable sustained pressure on the DPRK but also some American concessions. But with North Korea rushing headlong to achieve the capability to mount a hydrogen bomb atop a long-range missile that can target the United States, it is better to try to seek now a realistic capping agreement at a reasonable cost, than to hold out for a denuclearization agreement that can’t be bought at any price short of a bloody war, if that.

This article was originally published in Foreign Affairs.