Russia’s Turn to China Under Sanctions: Sanctions Decay, Substitution Elasticity, and the Challenge of Sustaining Pressure

In May 2026, Ukraine’s sanctions commissioner, Vladyslav Vlasiuk, offered an apparently encouraging assessment of export controls. Standing beside components recovered from Russian missiles and drones, he described the forced shift from some Western components to Chinese alternatives as evidence that sanctions were working. In the same briefing, however, he noted that Russia’s use of missiles and drones had increased compared to the same period a year earlier. Both observations, reported by the Office of the President of Ukraine, can be true. Controls may displace Western components and impose real costs without sufficiently reducing Russia’s ability to produce and deploy weapons.

That tension exposes a central weakness in the way sanctions and export controls are often evaluated. A reduction in Western components is treated as success, while the appearance of Chinese substitutes is treated as a secondary problem. But if the targeted capability survives, production continues, and investigators lose visibility into the new supply chain, the result may be less decisive than the component count suggests. This does not mean that governments should lift component controls whenever a substitute appears; restrictions can still raise costs and delay purchases, while removing them could simply give Russia easier access to the original products. Governments need to track how long those costs and delays persist as Russia adapts, and how controls should change when their effect begins to weaken.

The political context makes the problem more urgent. A recent CSIS discussion highlighted how continuous U.S. sanctions implementation against Russia has largely stalled while the European Union and United Kingdom have continued adding designations. An accompanying CSIS analysis explained why static lists lose value as companies, jurisdictions, and routes change. The European Union’s 21st sanctions package, adopted in July 2026, illustrates the contrast. It included 37 listings directly connected to long-range drones and added 51 entities to tighter dual-use restrictions, including entities in China and Hong Kong linked to circumvention involving microelectronics, computer numerical control machine tools, and semiconductor equipment. Designations show that governments are acting but not whether the restrictions continue to constrain their targets. The decisive test comes after a company is listed, a component controlled, or a route disrupted: How quickly, at what cost, and with what loss of performance can the target replace it? Where adaptation moves faster than governments can identify and constrain the substitute, the legal restriction remains in place while its practical effect erodes.
 

Sanctions Decay and Substitution Elasticity

That progressive loss of practical effect is known as sanctions decay. Political inaction can accelerate it, but so can a control design that focuses on yesterday’s supplier while overlooking tomorrow’s substitute. The pace of decay depends in large part on substitution elasticity: the speed, cost, engineering burden, performance penalty, reliability loss, and visibility change involved in replacing a restricted product, supplier, or acquisition pathway. A highly elastic supply chain can absorb pressure quickly. A less elastic one contains dependencies that are difficult to replace, integrate, qualify, or scale.

Recent trade data shows both adaptation and friction. The Kyiv School of Economics (KSE) found that export controls introduced since 2022 reduced the value of Russia’s imports of Common High Priority goods—which covers components and equipment important to Russian weapons production—by more than half, but that China had come to account for approximately 75 percent of reported exports of those goods to Russia. A later KSE analysis found that the premium Russia paid China for such goods rose with its dependence on Chinese supply and reached approximately 300 percent in 2025.

Those findings should not be reduced to a verdict that sanctions either work or fail. Russia retained market access, but at a higher cost and through a more concentrated supply relationship. Further analysis should establish whether the premium delayed production, limited scale, reduced reliability, or created new leverage and investigative signatures. A price increase is evidence of friction, but it is not, by itself, evidence of capability disruption.

The distinction is especially important because Russia’s response is not a wholesale switch from Western to Chinese technology. Old stock, newly acquired Western components, Russian production, Chinese substitutes, and Western products embedded in Chinese-made modules can coexist within the same platform family. Ukraine’s current War and Sanctions record—a public platform run by Ukrainian military intelligence about recovered Russian weapons—for the Geran-4 drone, for example, identifies a Chinese turbojet engine, network switch, mesh modem, connectors, and transceiver alongside components attributed to U.S. manufacturers. The configuration demonstrates adaptive, mixed sourcing and does not support a simple story of one national supply base replacing another.

Through China, From China, and With China

Russia’s growing dependence on China tells policymakers less than it may seem. Chinese actors may support Russia in different ways: (1) moving Western technology through concealed supply chains, (2) supplying Chinese-made products that Russia can use in place of Western technology, or (3) helping Russia integrate these technologies into its systems and develop domestic production capacity. Each role changes what investigators can see, which governments have jurisdiction to act, and where policy leverage remains.

1. Where China functions as a diversion hub, a Chinese or Hong Kong intermediary procures, sells, or transships Western-origin technology. Manufacturer records, export jurisdiction, and potential licensing authority still exist, but the pathway is concealed through additional entities and transactions. The enforcement challenge is familiar: Identify the intermediary, reconstruct the route, engage the manufacturer, and take action against the network.
2. Where Chinese-made products substitute for Western ones, the function may remain the same, but the original manufacturer, trace records, and export authority disappear from the chain. Investigators can no longer assume a cooperative trace response or access to transaction records. Controls designed around the original product may produce component displacement while also reducing visibility.
3. Where China begins to produce the capability itself, Chinese actors provide not only components but also integration, production, or a complete weapon system. In 2024, the U.S. Treasury Department described its action against entities involved in the Garpiya long-range attack drone as the first U.S. sanctions on Chinese entities directly developing and producing complete weapons systems with Russian firms. At that point, the policy problem is no longer limited to diversion of Western technology.

These categories overlap. KSE has identified three parallel pathways sustaining Russia’s military industry: direct Chinese substitutes, products made by Western companies in China, and transshipment through Chinese intermediaries. Western components may remain inside a Chinese module even when the external label and immediate supplier are Chinese, or a Chinese intermediary may sell both Chinese and Western products. The investigative task is to establish what actually changed: the product, the channel, the integrator, the jurisdiction, or all four.

China is also changing the legal environment surrounding efforts to identify and disrupt these activities. Under State Council Order No. 834, adopted in 2026, Chinese authorities may take countermeasures against foreign organizations that interrupt normal transactions with Chinese entities or impose discriminatory restrictions that cause or may cause substantial harm to China’s industrial and supply chains. The order also expressly subjects supply chain investigations and information gathering conducted in China to Chinese laws and regulations. Although it does not demonstrate a particular Russian procurement route, the measure shows how China’s expanding counter-sanctions framework may raise the cost of sanctions-driven disengagement and reduce visibility into substitute products, suppliers, and routes.

The speed of corporate adaptation also matters. Avinex Global registered in Hong Kong, operating from a Shenzhen address associated with Avtex Semiconductor, one week before the United States added Avtex to the Entity List. The U.S. Treasury Department later reported that Avinex had sent thousands of microelectronics shipments to Russia, including items critical to precision-guided weapons. The chronology illustrates one route to sanctions decay: A name-based restriction can remain legally valid while its practical coverage erodes because a related entity is already positioned to preserve the same supply function.

From Component Control to Capability Disruption

Capability-first analysis can help to close this loophole. Instead of beginning with the nationality or brand of a recovered component, it begins with the required function: navigation, flight control, communications, propulsion, manufacturing, testing, or another capability. It then asks which interfaces must remain stable, what performance is acceptable, which substitutes are technically compatible, what redesign or qualification is required, and whether an alternative can be obtained at the necessary scale.

A substitution assessment should examine function, performance, physical format, electrical and software interfaces, availability, price, integration burden, reliability, and production capacity. It should also identify the commercial and technical indicators that would reveal the switch. The Common High Priority Items List already recognizes that dependencies differ: Its highest tier includes items critical to advanced Russian weapons for which domestic production is lacking and global manufacturers are limited, while other tiers cover more widely available electronics, mechanical items, production equipment, and machine tools. That differentiation should become the starting point for assessing substitution, not the end of the exercise.

A practical case for testing substitution elasticity is the anti-jamming satellite navigation antenna used in Russia’s Geran-2 attack drone. The antenna helps the drone continue navigating when defenders interfere with satellite signals. Ukraine’s Defence Intelligence reported that a Geran-2 used in 2025 carried a new antenna containing 15 identified components; only two were American, while most of the rest were Chinese, including a chip made by the Beijing Microelectronics Technology Institute that analyzes incoming signals and determines which ones the antenna should ignore.

That finding shows what the newer design contains, but it does not by itself establish how easily substitution occurred. A proper test would compare earlier and later antennas: which components used in earlier antennas disappeared, what replaced them, how quickly the change followed restrictions or enforcement, and whether it required redesign or imposed additional costs, delays, reliability problems, or reduced performance. A fast, inexpensive switch with little loss of capability would indicate high substitution elasticity, while a slow, costly, or technically inferior switch would indicate lower elasticity and reveal where controls continue to impose useful friction. Repeating the comparison over time would show whether that friction persists or begins to erode.

Substitution can remove one point of control without eliminating leverage elsewhere in the supply chain, which may lie in software, production equipment, materials, testing, carrier boards, system integration, qualification, finance, logistics, insurance, or international market access. It may also shift from the component manufacturer to the module maker, contract manufacturer, freight forwarder, bank, or systems integrator needed to scale the substitute.

Where substitution elasticity is high, Russia can replace a restricted item relatively quickly and with little additional cost or loss of performance, meaning that imposing controls may not cause hoped-for disruption. Where elasticity is low, a narrow intervention against an integration bottleneck or scarce production dependency may impose meaningful delay. Effective controls should therefore target the points at which proportionate pressure can impose meaningful friction on acquisition by adding delay and cost, reducing reliability, constraining scale, or increasing investigative visibility.

Preserving Visibility: From Field Evidence to Anticipation

The China challenge concerns both leverage and visibility. As products, suppliers, and acquisition channels move beyond Western jurisdictions, authorities must determine which function Russia is preserving, what replaced the restricted item, and which dependencies remain. Recovered material, compared across cases and combined with commercial data, can distinguish diversion through Chinese intermediaries from Chinese substitution or deeper participation in production. It can also show whether adaptation imposed meaningful cost or merely changed component origin.

An earlier CSIS piece argued that tracing and attribution can turn the conclusion of an investigation into the starting point for anticipation. Applied to sanctions, that evidence cycle serves two connected but distinct purposes:

1. Longitudinal field, technical, and commercial evidence tests substitution elasticity by showing what changed and assessing the resulting cost, delay, performance loss, or new visibility.
2. Mass tracing, triangulation, and supply chain investigation identify acquisition channels that can be investigated and disrupted before they become durable workarounds.

The cycle begins with field documentation. For sanctions analysis, a recovered component becomes useful only when it is tied to the weapon, subsystem, and function in which it was found. A reproducible record captures its identity, condition, configuration, markings and production codes, associated components, recovery context, and information gaps. Comparing such records over time can show whether a restricted component remains in use, coexists with alternatives, or has been replaced. It can also help bound the period in which the switch occurred.

Tracing then asks how the relevant components moved through commerce. A uniquely serialized item can often be followed through a chain of transactions. Semiconductors and other mass-produced commodities present a different problem: Their markings may identify a product and production lot, but not one device within that lot. A manufacturer may therefore identify dozens of distributors or customers that received units from the relevant lot without being able to determine which one handled the recovered device.

Mass tracing addresses that ambiguity by submitting trace requests for many recovered components and retaining the complete set of potential custodians returned for each one. Responses may be inconclusive individually, but collectively they map the possible acquisition channels sustaining the targeted capability, including Chinese distributors and customers associated with the relevant production lots. When compared over time, they can also show whether the channels changed following a control, designation, or enforcement action.

Triangulation begins after mass tracing. If one trace response identifies a single distributor as the last known custodian, investigators can look for that company in broader responses for other recovered components. Repeated appearances isolate a company from dozens of possible distributors and customers, creating a pattern worth investigating. Supply chain investigation then tests that pattern against corporate, trade, financial, logistics, and field evidence. It asks whether the recurring entity helped replace a controlled product or route, and whether the switch imposed delay, cost, performance loss, reliability problems, new exposure, or constraints on scale.

The immediate value of this work is operational: Recurring trace results identify acquisition channels that manufacturers and authorities can investigate and disrupt before they become durable workarounds. Substitution elasticity is assessed separately by observing what replaces a restricted or disrupted component, supplier, or route and measuring the time, cost, technical burden, and performance consequences. Repeating that assessment shows whether controls continue to impose friction or whether their effect is decaying.

Analysis can then assess which requirements will persist, which substitutes are realistic, who could provide them, which acquisition signatures should appear, and what evidence would require the assessment to change. Anticipation enables smart friction: proportionate action against the dependencies, entities, or channels most likely to sustain the next adaptation before a substitute or successor becomes established. Later recoveries show what changed and improve the next assessment. The recommendations that follow translate this evidence cycle into sanctions design, monitoring, coordination, and evaluation.

Designing Controls for the Switch

1. Require substitution assessments before major controls. U.S., European, and UK authorities should include a substitution assessment when proposing significant technology controls or designations. This assessment should identify likely Chinese and other alternatives, integration time, performance penalties, availability, scale, likely acquisition channels, residual dependencies, and observable indicators. A control should be designed with the expected response in view.
2. Target capability ecosystems and inelastic dependencies. Governments should look beyond individual part numbers toward the inputs required to integrate, qualify, produce, and scale the targeted capability, including software, machine tools, test equipment, materials, carrier boards, technical expertise, or logistics. KSE’s work on Russian military production identifies logistics firms, nominally civilian suppliers, and concentrated import dependencies that may offer more durable leverage than another interchangeable component.
3. Build successor monitoring into every designation. At the time of designation, authorities should capture the network features most likely to survive a change of company name: beneficial owners, directors, employees, addresses, telephone numbers, domains, banks, freight forwarders, suppliers, customers, product bundles, and payment patterns. These indicators should feed scheduled reviews rather than waiting for a successor to accumulate an obvious trade history.
4. Create a shared adaptation and substitution function. The United States, European Union, United Kingdom, Ukraine, and other partners should maintain a joint forecast register for priority capabilities that records likely substitutes and routes, evidence, confidence, warning indicators, disconfirming evidence, review dates, actions taken, and outcomes. Even when the pace of designation differs politically, shared analysis can preserve continuity and shorten the time between detection and action.
5. Build an interoperable evidence architecture. Governments should establish common standards and interfaces through which field observations can be ingested, compared, audited, and connected with customs, corporate, commercial, financial, and logistics data. The system should support data portability, transparent methodologies, explicit confidence assessments, and durable government access to supporting records and audit trails. It should allow evidence collected by different agencies, investigators, manufacturers, and partner governments to be compared across cases. It must remain useful when manufacturer cooperation or downstream transaction records are unavailable, while ensuring that recurring patterns are not mistaken for proof of a specific diversion pathway.
6. Measure capability disruption, not component displacement. Evaluation should track redesign and qualification time, price premiums, production delays, performance loss, failure rates, scaling constraints, procurement complexity, and new investigative exposure. Changes in weapon output or attack tempo should be examined through time-series evidence and compared with other possible causes. Neither correlation nor a declining Western component count should be treated as proof of causation.
   
   For Chinese intermediaries handling Western-origin goods, familiar tools such as manufacturer engagement, end-use scrutiny, entity designations, and pressure on finance and logistics remain relevant. For genuine Chinese substitutes or Chinese-centered capabilities, allied leverage is narrower. Policymakers should identify exposure to global finance, insurance, transportation, production inputs, and international markets, while acknowledging honestly when allied jurisdiction is weak. The answer is not to pretend that every supply chain contains a Western switch that can be turned off.
    

Keeping Sanctions Effective

The appearance of a Chinese substitute in a recovered weapon is not, by itself, proof that sanctions have failed, nor is the disappearance of a Western component proof that they have succeeded. Controls should be judged by whether the adversary’s capability became slower, costlier, less reliable, harder to scale, or more visible.

Sanctions decay occurs when adaptation outpaces policy. Substitution elasticity helps explain how quickly that decay may happen and where pressure might still matter. If governments assess likely substitutes before imposing the control, preserve independent visibility as supply chains migrate, and measure the effect on capability rather than the nationality of its components, sanctions can do more than displace yesterday’s component; they can shape tomorrow’s acquisition decision.

Damien Spleeters is a senior associate (non-resident) with the Economics Program and Scholl Chair at the Center for Strategic and International Studies in Washington, D.C.

This report is made possible by general support to CSIS. No direct sponsorship contributed to this report.