China’s 15th Five-Year Plan (2026–2030): Global Footprint

Executive summary

China’s global footprint through 2030 is likely to be shaped by the combination of technology-led industrial upgrading at home, selective manufacturing abroad, and continuing exports of Chinese equipment and intermediate goods. The strategic objective is greater Chinese resilience while foreign customers become more closely integrated with Chinese production and technical ecosystems. This is an analytical judgment, not evidence that Beijing directs every overseas corporate decision.

The strongest demonstrated overseas deployment is in green manufacturing. AI expands the footprint through cloud services, models and industrial software; semiconductor influence extends through packaging, mature chips and the products containing them. These channels reinforce one another, but their overseas maturity differs.

Ports and adjacent industrial parks form a logistics layer beneath these sectors. CFR’s July 2026 tracker records 145 overseas port projects with varied Chinese equity or operational stakes, 132 active and 13 inactive. Multiple projects can occupy one port, and these are inherited investments rather than demonstrated products of the 15th Plan. Assessment: terminal access and shipping links can reinforce exports and overseas factories; the tracker does not establish control of 145 distinct ports, host-state policy, or naval access. CFR: Tracking China’s Control of Overseas Ports, July 2026

The footprint extends beyond vehicles: the IEA’s 2026 assessment identifies Chinese firms as substantial owners of solar-PV production capacity outside China. Assessment: moving a plant across a border can diversify geography without diversifying corporate control. Solar panels, batteries and wind equipment can make decarbonisation cheaper while increasing exposure to particular upstream suppliers. IEA: Energy Technology Perspectives 2026, executive summary

Sovereignty is best assessed as practical freedom to regulate, obtain supplies and change vendors. Investment can increase that freedom by adding factories and infrastructure, or reduce it when vital inputs, maintenance, software or finance become difficult to replace. Ownership alone does not establish political control.

What the 15th Plan establishes—and what it does not

The National People’s Congress approved the 2026–2030 outline on 12 March 2026. “New Quality Productive Forces” describes innovation-led productivity and industrial upgrading, rather than a separately measurable overseas investment category. The outline prioritises core technologies including integrated circuits; the government’s AI Plus programme stresses widespread commercial application, AI agents and intelligent terminals. Plan approved, State Council, 13 March 2026 15th Plan outline, official Chinese text State Council: AI Plus, 5 March 2026

Assessment: the international transmission mechanism is domestic research, manufacturing scale and competition → exports and selective localisation → installed systems and supplier relationships → greater switching costs. That sequence can generate leverage without a formal political condition attached to each transaction.

Attribution discipline: the Thailand, Indonesia and Brazil examples below began before 2026. They are inherited capabilities consistent with the new plan, not projects proved to have been caused by it. The plan supplies strategic direction; implementation depends on firms, provincial authorities, financiers and host governments.

ODI stock: the 7.2% benchmark, correctly dated

China’s National Bureau of Statistics reports approximately US$3.1 trillion of outward direct investment stock at end-2024, representing 7.2% of the global stock. The newer 2025 statistical bulletin, released on 23 September 2026, reports stock of US$3.38407 trillion and annual net ODI flows of US$213.58 billion, up 11.1%. It records overseas enterprises across 189 countries and territories. The ministry’s release reports a 7.4% share of global investment stock for 2025; the 7.2% figure refers to 2024. MOFCOM: 2025 ODI bulletin release NBS: 14th Plan external economic baseline, June 2026 MOFCOM/NBS/SAFE: 2025 ODI statistical bulletin, 23 September 2026

The 7.2% figure is therefore a historical global-share benchmark, not a 2030 target and not a verified end-2025 share. No updated global denominator is inferred here. A rise in annual flows cannot be mechanically applied to the old share.

Measurement matters: stock is an accumulated position, while flows cover a period. Reinvested earnings and intercompany debt count toward ODI; ordinary exports, construction contracts, sovereign loans and central-bank swaps are different categories. Financial routing through Hong Kong or offshore centres can obscure ultimate operating exposure. Project commitments and nameplate capacity must not be added to recorded ODI as though they were realised investment.

CFR’s project counts cannot be added to ODI stock or converted into a share of global ports. Its definition of ownership requires equity in a specified port, wharf or terminal; leases and operating concessions convey different rights. The tracker’s aggregate includes varied equity and operational stakes, so each case needs its own contract and beneficial-owner check. CFR: overseas-port tracker, data notes

Assessment: the headline stock demonstrates breadth, but strategic influence is concentrated in particular nodes. A relatively small investment controlling a hard-to-replace process or service can matter more than a larger, easily substitutable financial holding.

Trade: less direct US exposure, continuing Chinese supply-chain centrality

For calendar 2025, nominal Chinese goods exports increased 5.5% overall, while exports to the US fell 20%. Exports rose 13% to ASEAN, 8% to the euro area and 7% to Latin America. The US decline was US$104 billion. These are export growth rates in US dollars—not bilateral total-trade growth, market shares or real volumes. Euro-area figures should not be relabelled as EU figures. ECB: Global trade redirection, Economic Bulletin 1/2026

The ECB finds only limited evidence that US tariffs caused broad trade diversion: positive effects were significant for ASEAN and Africa, but not the euro area. ASEAN intermediate-goods trade also raises the possibility of processing or rerouting before products reach the US. Less direct bilateral trade does not prove that US final demand has ceased to matter.

CFR’s October 2025 analysis puts heavier emphasis on Southeast Asian rerouting and warns that direct-shipment declines can overstate China’s separation from US consumers. That is a useful competing explanation, but its claims about the scale of rerouting are analytical estimates, not a measured allocation of the 2025 regional export changes above. Preserve the ECB’s narrower causal finding until product-level origin and final-demand data resolve the difference. CFR: China’s trade-war and growth-reform assessment, October 2025

Assessment: separate three mechanisms when tracking the footprint: new non-US demand; genuine overseas production serving regional customers; and indirect US exposure through third-country supply chains. Only the first two clearly diversify final-market dependence. Rerouting is not itself proof of unlawful circumvention.

For January–September 2026, the official October release reports goods exports of RMB23.2 trillion, up 16.4% year on year. That is a newer aggregate in renminbi, not a directly comparable destination series in US dollars. The report therefore retains 2025 for its consistent regional comparison and does not label those rates as 2026 performance. MOFCOM overseas platform: January–September 2026 trade

European Union: local factories, imported dependencies

Green technology is the clearest channel. MERICS/Rhodium estimates Chinese FDI in Europe at €16.8 billion in 2025, including €8.9 billion in greenfield investment. Automotive investment reached €7.6 billion, with 93% in EV supply chains; Hungary received €3.9 billion across sectors. The geographical coverage is EU plus UK, not EU alone, and this transaction-based dataset is not directly interchangeable with China’s official ODI statistics. MERICS/Rhodium: Chinese FDI in Europe, 2025 update, May 2026

Assessment: factories can give Chinese firms proximity to customers and help them respond to trade barriers, while host countries gain jobs and capacity. However, local assembly does not establish domestic control of battery chemistry, machinery, design rights or upstream materials. Competition among member states for projects can complicate collective bargaining; an investment announcement does not prove a political concession.

Trade defence is an active counterweight. The Commission’s January 2026 guidance addressed price undertakings for Chinese BEVs, alongside countervailing duties introduced in 2024 ranging from 7.8% to 35.3%. Guidance is not blanket repeal of the duties. Host leverage lies in market access, procurement requirements and enforceable investment conditions. European Commission: BEV price undertaking guidance, January 2026

Ports show a spectrum of bargaining outcomes. COSCO’s Hong Kong unit holds 67% of the listed Piraeus Port Authority, according to the authority’s share register. At Hamburg’s Tollerort container terminal, COSCO Shipping Ports instead acquired a non-controlling 24.99% stake after German investment screening; HHLA says the transaction did not sell Hamburg port or HHLA. The first is corporate majority ownership in a port authority; the second is a restricted terminal-company minority position. Neither percentage alone establishes sovereign control or naval access. Piraeus Port Authority: shareholding structure HHLA: COSCO investment fact check

Semiconductors show reciprocal vulnerability. In 2025 the Dutch government intervened at Nexperia, and China imposed company-specific export controls affecting its Chinese locations. The Dutch order was subsequently suspended in November. This historical episode illustrates how fabrication, ownership and downstream processing can sit under different jurisdictions; it should not be read as a statement of the company’s complete October 2026 legal position. Dutch government: Nexperia supply restrictions, November 2025 Nexperia: suspension of Dutch order, 19 November 2025

Mineral processing is another source of leverage. The IEA’s 2026 outlook identifies China as the top refiner with over 90% of supply for gallium, graphite, manganese and rare earths in its assessment. EU leaders had already raised the impact of rare-earth and magnet export controls at the July 2025 summit. The relevant exposure is continuity of licensed supply, not merely the location of a European factory. IEA: Global Critical Minerals Outlook 2026, Outlook EU Council: 25th EU-China summit, July 2025

AI assessment: compared with the documented EV pipeline, the evidence assembled here does not establish a similarly large Chinese AI infrastructure footprint inside the EU. Industrial software, cloud services and embedded vehicle systems remain important procurement questions. Examine administrator access, model hosting, data portability and update rights case by case; Chinese branding alone establishes neither data transfer nor dependence.

Latin America: industrialisation, resources and logistical control

Brazil provides evidence of localisation beyond imports. Bahia’s government reported almost 20,000 vehicles produced at BYD’s Camaçari operation in 2025. That is a reported output measure, not the much larger advertised annual capacity or a measurement of Brazilian component content. Bahia government: BYD production, January 2026

Assessment: the bargain is access to a large consumer market in exchange for industrial activity. Brazil’s leverage is strongest when it converts market access into supplier development, worker skills and verifiable domestic value added. Across mineral-producing economies, the corresponding test is whether contracts build local processing and transferable expertise or reinforce commodity exports and imported finished technology.

Grid infrastructure deepens the footprint beyond product sales. Brazil’s public energy research company EPE documented a July 2025 financing-cooperation memorandum between the New Development Bank and State Grid’s Brazilian subsidiary for the Graça Aranha–Silvânia transmission project, described as under construction. This is evidence of institutional cooperation and project development, not a completed asset or a verified loan drawdown. Assessment: oversight, service continuity and technical access become central when the foreign investment supports essential electricity infrastructure. Brazil EPE: State Grid transmission finance cooperation, July 2025

AI enters through commercial infrastructure and services. Huawei’s Brazil Cloud Summit 2025 promoted inference services, AI agents and cloud-stack upgrades. This establishes a vendor offering, not its market share, public-sector adoption or control of Brazilian data. The sovereignty questions are contractual: encryption keys, audit access, hosting jurisdiction and the cost of migration. Huawei: Brazil Cloud Summit 2025 offerings

Peru shows why logistics belongs in the same analysis. Its trade ministry reported that Chancay handled 62% of Peru’s agricultural exports sent to China in January–September 2025. The port also became a test of regulatory authority: after Ositrán appealed a January 2026 judgment limiting its oversight, a Lima appellate court reversed that judgment on 1 July and affirmed the regulator’s statutory competence. COSCO said it would seek further constitutional review, according to Ositrán, so the July ruling should not be treated as the end of all litigation. Commercial use and a live legal contest coexist; Chinese investment did not eliminate Peru’s regulator. Peru trade ministry: Chancay agro-export use Ositrán: Chancay appellate ruling, July 2026 Ositrán: COSCO’s stated further appeal

Panama is a different kind of host-state test. Its Supreme Court declared the Panama Ports Company concessions for Balboa and Cristóbal unconstitutional in January 2026; the government then granted temporary operating concessions to other companies in February. PPC is a subsidiary of Hong Kong-based CK Hutchison, and a concession is not ownership of the canal. CFR reports retaliatory pressure from Beijing and shipping disruption after the ruling. The verified judgment and transition demonstrate Panama’s legal capacity to change operators; the reported retaliation shows that exercising that capacity can carry commercial costs. It does not prove Beijing controlled the canal before the ruling. Panama Supreme Court: January 2026 decision Panama presidency: temporary port concessions, February 2026 CFR: China in Latin America, March 2026

Financial instruments widen the relationship. China announced an RMB66 billion credit line for Latin American and Caribbean countries in May 2025; a separate Brazil–China central-bank swap was renewed at RMB190 billion/BRL157 billion. A credit commitment is not disbursement, and a swap ceiling is neither ODI nor evidence of full drawing or replacement of dollar finance. China MFA: RMB66bn LAC credit line, May 2025 Chinese government: Brazil currency swap renewal, May 2025

Assessment: these instruments can broaden financing options and reduce exclusive dependence on Western providers. They can also create concentrated creditor or supplier exposure. Claims of automatic debt-driven loss of sovereignty require contract-level evidence of security, default remedies and actual enforcement; this report does not establish such a general mechanism.

Semiconductors are a weaker demonstrated channel here than green manufacturing or digital services. The evidence assembled does not support a regional wave of Chinese leading-edge fabs. ECLAC frames the broader challenge as connecting FDI in minerals and digital transformation to productive development; the relevant policy test is local capability creation rather than investment announcements alone. ECLAC: Latin American FDI 2025, July 2025

Southeast Asia: production networks and the AI infrastructure bridge

Indonesia illustrates vertical integration. CATL and partners broke ground in June 2025 on an approximately US$6 billion battery integration project spanning parts of the minerals-to-battery chain, corroborated by Indonesia’s energy ministry. This is a construction milestone and project value, not proof that the entire complex is operational or the full amount invested. CATL: Indonesia battery project groundbreaking, June 2025 Indonesia energy ministry: integrated battery project, June 2025

Thailand illustrates regional manufacturing. BYD inaugurated its Rayong factory in July 2024 with stated annual capacity of 150,000 vehicles. Capacity is not actual production. Assessment: regional assembly can support Asian customers and exports while retaining Chinese machinery, components and technical relationships. Local employment and local strategic control need separate measurement. BYD: Thailand factory inauguration, July 2024

Malaysia’s Kuantan shows how ports and industrial sites can be joined without exclusive Chinese ownership. The port operator reports a 60:40 split between Malaysian IJM and Beibu Gulf Holding; it also describes the nearby Malaysia–China Kuantan Industrial Park as a linked industrial area. A Kuantan–Beibu cold-chain route launched in April 2025. CFR’s companion tracker identifies 194 overseas industrial-park projects worldwide, but these are not all ports, not all operational and not all state-owned. Assessment: such corridors can lower logistics costs and support local industry while making shipping continuity and minority-investor rights important diligence points. Kuantan Port: ownership and industrial links Kuantan Port: cold-chain route, April 2025 CFR: overseas industrial-park tracker, July 2026

AI deployment is directly visible in Malaysia: Alibaba Cloud announced a Johor public-cloud region in June 2026 with two new data centres, taking its stated Malaysian total to five. This is a company-reported operational footprint; it does not establish national market dominance or that the facilities use exclusively Chinese chips. Alibaba Cloud: Johor cloud region launch, June 2026

Semiconductor participation is more substantial than simple assembly relocation suggests. MIDA’s 2025 project material identifies nearly RM5 billion of investment associated with TF AMD Microelectronics in Penang, covering wafer bumping and advanced packaging, and classifies the ultimate source as China. This is a cross-border corporate production network, not evidence of a self-contained Chinese leading-edge fabrication ecosystem. MIDA: notable 2025 projects, TF AMD Penang

Host governments retain agency. Malaysia’s Arm partnership targets design capabilities and engineer training, while its July 2025 rules imposed permits on exports, transhipments and transits of high-performance US-origin AI chips. Together, these show capability building alongside exposure to external technology controls. The rule is cited as dated evidence of that pressure, not a comprehensive statement of all current licensing requirements. MIDA: national Arm partnership, March 2025 Malaysia MITI: permits for US-origin AI chips, July 2025

Assessment: Southeast Asia can gain from multiple competing investors, yet its room for manoeuvre narrows if regional production depends simultaneously on Chinese inputs, US-controlled advanced computing and overseas customers. Track beneficial ownership, actual local transformation and end users. Do not treat every Chinese-backed factory as a conduit for tariff or export-control evasion.

How economic instruments change practical sovereignty

Capital and market access: greenfield plants, joint ventures and acquisitions create jobs and domestic constituencies for stable relations. Host markets can demand skills and supplier development in return. Evidence of influence is strongest when documented policy decisions or contract terms accompany the investment.

Operational dependence: proprietary software, maintenance, cloud interfaces, manufacturing recipes and model tools can make switching expensive. The decisive rights are access to keys and data, independent repair, interoperability and the ability to operate through a supplier interruption. These are due-diligence questions, not assumed deficiencies in the named projects.

Supply and finance: refining concentration, export licences, long-term offtake, credit and currency facilities can affect who obtains inputs and liquidity during stress. Currency diversification can strengthen autonomy; concentrated financing can constrain it. Both effects can coexist.

Maritime logistics: equity, leases, concessions, shipping services and port IT confer different powers. Inspect berth-allocation rights, terminal data, customs access, maintenance and interruption clauses, concession renewal and host enforcement before judging leverage. CFR rates physical potential for naval use from berth, depth and support characteristics; it explicitly distinguishes that from political feasibility. A port capable of receiving a warship is not evidence of basing rights. CFR: overseas-port tracker, methods and limitations

Rules and standards: product certification, procurement specifications, trade agreements and data rules shape which suppliers can compete. Track actual adoption and enforceable requirements, rather than equating participation in a Chinese initiative with wholesale adoption of Chinese governance.

The EU’s principal contest is over industrial competitiveness and regulatory coordination; Latin America’s is over upgrading, infrastructure oversight and financing terms; Southeast Asia’s is over capturing supply-chain value while navigating competing technology jurisdictions. These are analytical comparisons, not a ranking of national sovereignty.

2030 outlook and indicators to watch

Base case—selective localisation, persistent Chinese centrality (medium confidence): overseas EV, battery and cloud capacity expands where market size and policy allow, while important machinery, materials and know-how remain connected to China. This judgment would weaken if sustained local sourcing, independent R&D and alternative suppliers grow faster than Chinese imports.

Alternative—stronger host bargaining: governments use competing investors and conditional incentives to secure skills, local processing, open interfaces and resilient supply. Watch measurable supplier spending and technical employment, not memoranda of understanding.

Alternative—fragmentation: trade restrictions, licensing disputes or project economics stall investment and split technology ecosystems. Watch cancellations, utilisation, export-licence delays and the cost of duplicated infrastructure. No numerical scenario probabilities are assigned.

Tracking priorities: (1) ODI stock and annual flows, with the year and global denominator; (2) ultimate operating destination and beneficial owner; (3) factory completions, output, utilisation and domestic value added; (4) direct versus indirect US final-demand exposure; (5) cloud customers, hosting and switching rights; (6) semiconductor production stage and controlled inputs; (7) loan disbursements versus commitments and swap drawings versus ceilings; (8) port equity versus concession, throughput and diversion, contractual access rights, regulator access and enforcement outcomes.

Decision implication: evaluate whether a project adds locally usable capability and credible alternatives. A new factory or data centre strengthens autonomy most when the host can regulate it, maintain it and keep it operating through a supplier or diplomatic shock.

Evidence, confidence and limits

Research cut-off: 19 October 2026. This public-source study draws on official statistics, government and regulator statements, corporate project disclosures, ECB analysis, IEA research and the MERICS/Rhodium transaction study. Official plans establish intent; company disclosures establish claims about their projects; neither independently proves geopolitical outcomes.

Quality confidence: 85/100 — coverage and sourcing of this report. The key ODI dates and categories, trade comparison and regional examples were checked against named sources; the port cases were cross-checked against operator and host-government records. The largest uncertainties are project-level contract terms and beneficial ownership, realised local value added, private AI deployment, and the causal link between commercial dependence and political concessions.

Not checked: CFR’s underlying project-level dataset against every port, confidential contracts, site inspections, every post-announcement operational update, or the outcome of any further Chancay constitutional review. CFR’s tracker is a secondary, non-exhaustive project census, not an audit of sovereign rights. Forecasts are judgments; no 2030 investment totals, AI market shares, semiconductor self-sufficiency rates or sovereignty scores are invented.