Launch on warning: How China is rewriting its nuclear doctrine

What did minimum deterrence mean?

For half a century, Chinese nuclear doctrine rested on an assumption that set it apart from the American and Soviet approaches. Beijing held that deterrence required no more than the ability to deliver a handful of warheads to an adversary’s territory, and that this made high readiness unnecessary. Warheads were stored separately from missiles, the arsenal numbered around two hundred devices, and no first use was declared without qualification. Early warning had no place in this arrangement: if the response were to come after an enemy strike rather than after its detection, a few extra minutes could change nothing.

That model is now ending, and the clearest evidence lies not in warhead numbers but in the warning and command infrastructure. The Pentagon’s 2019 report conceded that People’s Liberation Army writings saw interest in launching missiles upon warning; however, the Pentagon called this an aspiration. Four years later, the same document describes the implementation of an arrangement that Chinese terminology calls early warning counterstrike (预警反击). David Logan of the Fletcher School and Phillip Saunders of the National Defence University, in a July 2025 study for the Defence Threat Reduction Agency, stated plainly that China has assembled all three elements that such a concept requires: sensors able to detect and track incoming missiles, forces held at a readiness allowing launch within minutes, and a command chain capable of reaching a decision in time.

Sensors in space and on the ground

An intercontinental missile launch is detected in the infrared, by the exhaust plume standing out against the cold Earth. The pace of Chinese progress can be read from the Pentagon’s own reports: in 2019, China had at least one early warning satellite, three years later, at least three. They are attributed to a programme codenamed Huoyan (火眼), “fire eyes,” hidden behind the civilian designation of communication technology test satellites, with estimates running from three to five. They are positioned over the Indian Ocean, the South China Sea, and the Pacific. The ground segment rests on large, phased-array radar stations, erected since 2009. Logan and Saunders count at least seven such stations across six sites, among them Huanan, Yiyuan, Hangzhou and Korla. Chinese reporting puts their detection range at around 4,000 km, and their parameters approach those of the American PAVE PAWS stations.

However, these numbers indicate that the system is still limited. Three satellites can support only a rudimentary network of early detection, and with so few of them the loss of a single satellite could leave the whole system blind. Chinese analysts acknowledge this themselves, arguing that more sensors are required to confirm incoming missiles with confidence. The weakness is still there, and that is making the whole network prone to failure in a crisis and, at the same time, an attractive target for an adversary. Without redundancy, the system is at once fragile and tempting as a target.

Command and readiness

Sensors are useless without a command chain able to turn their readings into an order, and here the changes are equally marked. The Rocket Force now keeps part of the arsenal on continuous alert and has adapted crew training and certification accordingly. The shift comes through in the language: after the 20th Party Congress, in October and November 2022, the Rocket Force newspaper began equating ordinary peacetime duty with a wartime posture, using the formulations “missiles on the rack, on high alert” (弹在架上，高度戒备) and “duty is battle” (值班就是战斗). In one reported case, a brigade received warning of an incoming strike just as it finished preparations and managed to fire its missile and disperse its equipment a minute before the enemy warheads arrived. Yang Chengjun, a retired Rocket Force strategist, claimed unit reaction times had been cut “from days and hours to minutes.” A graduated alert system probably exists as well; John Lewis and Xue Litai described a four-stage alert scale as early as 2012.

What the Chinese doctrinal texts say

Officially, Beijing does not confirm any departure from its established doctrine, but the doctrinal texts have their own say. The Academy of Military Science’s Science of Military Strategy of 2013 states that once it has been established with certainty that the enemy has launched nuclear missiles, and before his warheads have reached their targets and detonated, a counterstrike may be launched rapidly.

One contradiction remains unresolved, and it touches the core of China’s declared policy. The 2013 edition assumes that launching before an enemy warhead detonates does not breach no first use, since an enemy strike remains the precondition. Yet Fu Cong, then director-general of the arms control department at the Chinese foreign ministry, declared in 2019 that the concept does violate the principle, and called on all nuclear powers to abandon it.

The Defence Intelligence Agency stated in 2024 that the newer silos are being folded into precisely this arrangement.

Consequences: a compressed decision window

The first consequence concerns time. Launch on warning was calculated for intercontinental missiles, which leave roughly half an hour between launch and impact; submarine-launched, hypersonic and depressed-trajectory missiles shorten that window drastically. Stationing shorter-range American nuclear missiles in the Indo-Pacific would therefore strain the Chinese system further and raise the odds of a decision taken on a misreading of intent.

The second is the use of the same sensors for nuclear and conventional roles. The same radars and other systems serve missile defence and non-nuclear operations; an attack on them during a conventional conflict would be regarded as a strike against the capability of nuclear retaliation. This results in a classic escalation mechanism created by entanglement.

Consequences: the gap in the safeguards

The third axis is the gravest and the least examined. The Americans and the Soviets both required a warning to be confirmed by two independent kinds of sensor, precisely to limit the risk of a false alarm. Chinese writing runs the other way: a 2018 review of hundreds of Chinese academic papers found that analysts fear missing an attack far more than an unfounded alarm, and in CNKI, China’s principal academic repository, the equivalent term does not appear at all. A 2017 report on Rocket Force work praises the finding that an excess of checkpoints slows command down. The paradox is that a system without redundancy is the one that most needs procedures guarding against false alarms, and those are exactly what the Chinese literature lacks.

Consequences: who can halt the order

The fourth axis is a concern about the peak of the mountain of nuclear forces that is China’s high command and political decision-making chain. An order for a nuclear strike reportedly must come from both the Central Military Commission and the Politburo Standing Committee, and summoning both within less than half an hour would be practically impossible. The most important observation lies in the usual reasoning of delegated authority; Xi Jinping matters not because he alone can order launch, but because he alone can forbid one, should the attack prove accidental or limited. The greater danger is not that an order arrives too late, but that a pre-planned response is set in motion and no one manages to stop it.

A complication: the decision predates the buildup

Work on the concept began earlier than assumed. As early as 2010, the Second Artillery Force ordered research on early warning counterstrike to be strengthened, and a year later its service newspaper revealed the creation of a research group at the Second Artillery Academy in which the subject was named, alongside deterrence, as a pillar of operational theory. The decision therefore came before the arsenal expansion of 2015–2016. A conclusion relevant to arms control follows. Beijing reached for the concept when its arsenal was small and vulnerable to destruction in a single disarming strike, and when rapid response compensated for weakness. Today the arsenal has grown and would survive a strike better than ever, which largely voids the original justification.

What we do not know

The limits of this analysis should be stated clearly. China’s lack of transparency obscures where its arsenal is heading, and this has pushed some American officials towards assuming the worst. A study by the Centre for Global Security Research, however, concludes that Beijing has not abandoned no first use and is most likely not seeking parity with the United States. Logan and Saunders themselves concede that the exact rules governing nuclear use cannot be reconstructed from open sources. Since the 2013 Science of Military Strategy recommends measured ambiguity towards an adversary, some of the Chinese material may also have been written with Western readers in mind. The direction of travel is nonetheless clear. The satellites, radars, continuous alert duty and doctrinal language described above make sense only if China intends to respond within minutes, a break with the minimum deterrence that set it apart from the other nuclear powers for half a century. The effect on strategic stability will therefore depend less on the number of Chinese warheads than on three things that open sources cannot reveal:

1.      How much of the force is kept on high alert.

2.      How much certainty Beijing will demand before ordering a counterstrike.

3.      Who will be able to stop that order once it has been given.