The description of Belgrade’s Expo 2027 programme as a Chinese-financed, Chinese-built development is only partly accurate. Chinese state-owned contractors have secured a central position in construction and equipment supply, but the disclosed financing structure is more Serbian than Chinese. The principal Expo buildings, National Stadium, supporting infrastructure and several transport packages are being delivered by Chinese engineering, procurement and construction groups, while much of the capital is coming from the Serbian budget, domestic bank borrowing and Serbia’s general sovereign funding programme.
The distinction matters. Expo 2027 is not currently structured as one large China Exim Bank loan tied to a Chinese contractor. It is closer to a Serbian sovereign capital-expenditure programme executed through Chinese EPC companies under bilateral agreements and direct-award procedures. China therefore has substantial influence over delivery, technology and contractor economics, but Serbia retains most of the financing, refinancing, cost-overrun and post-Expo utilisation risk.
The relevant fiscal envelope is approximately €3.4bn
The Serbian Fiscal Council’s latest estimate puts expenditure on Expo 2027 and directly connected projects at approximately €3.4bn. By the end of 2025, around €1.25bn had been spent on Expo in the narrower sense and approximately €750mn on associated infrastructure, including the National Stadium, linear infrastructure and the railway connection. That gives cumulative expenditure of roughly €2bn through 2025. Current plans add approximately €550mn for the Expo itself and another €850mn for related infrastructure during 2026 and 2027.
This figure is broader than the Expo operating company’s earlier estimate of €1.29bn, which covered the exhibition area and a defined group of supporting assets. The difference reflects scope: the Fiscal Council includes a wider set of investments and associated current expenditure, including subsidies and major connected projects. The two figures should therefore not be treated as competing estimates of precisely the same programme.
At €3.4bn, the programme is large enough to influence Serbia’s public-investment composition but not large enough, on its own, to destabilise sovereign finances. The risk lies in its interaction with other capital programmes, energy-sector guarantees, defence expenditure and Serbia’s relatively expensive cost of borrowing. The Fiscal Council expects the fiscal deficit to remain around 3% of GDP in 2026 and 2027, with central-government debt broadly stable in the mid-40% range of GDP. Yet it also notes that Serbia spends around 1.9% of GDP on interest, close to the EU average of 2.1%, despite carrying a much lower debt ratio. Serbian debt is therefore moderate in quantity but comparatively expensive to service.
PowerChina is the core construction platform
Power Construction Corporation of China, generally known as PowerChina, has been given a particularly broad position. Official Ministry of Finance disclosures identify PowerChina as the directly contracted main contractor for the linear infrastructure serving the new Surčin development, the National Football Stadium, Expo preparatory works and the construction of Expo facilities. The contracts were registered between April 2023 and March 2024, with the Republic of Serbia acting as financier or investor through the Ministry of Finance. The disclosed project register names the contractor and procurement method but does not provide the full values of these central PowerChina packages.
The Chinese role extends beyond the construction site. PowerChina is building the railway from Zemun Polje through Nikola Tesla Airport to the National Stadium and Expo zone under a contract worth €188.1mn. The contract was awarded directly pursuant to the Serbia–China intergovernmental infrastructure agreement, with financing identified as coming from the Serbian state budget.
CRRC Changchun Railway Vehicles is supplying nine electric multiple-unit trains for the Expo and Belgrade suburban rail network for €73.98mn. This contract was also concluded directly under the bilateral agreement and is funded from the Serbian budget. China Shandong International Economic is responsible for the reconstruction and expansion of the Novi Beograd railway station complex under another directly awarded contract, worth €138mn and again financed from the Serbian budget.
Those three publicly valued transport packages alone total €400.08mn. They do not include the substantially larger PowerChina contracts for the Expo complex, stadium and linear infrastructure, whose complete contract values have not been published in the same disclosure. It is consequently possible to establish that Chinese companies control a major share of Expo delivery, but not to calculate a reliable overall Chinese EPC percentage from currently available contract-level information.
This concentration has two competing effects. Giving PowerChina responsibility for several connected packages can reduce disputes between designers, infrastructure contractors and site builders. It may also help Serbia manage the immovable opening deadline of 15 May 2027. At the same time, contractor concentration reduces the state’s ability to compare prices across competing bidders and increases dependency on one corporate group for the coordination of several critical paths. The closer the deadline comes, the weaker the client’s negotiating position becomes regarding variations, acceleration payments and acceptance of incomplete secondary works. The exhibition itself is scheduled to run for only 93 days, until 15 August 2027, making construction delay costs highly asymmetric.
The Chinese-loan narrative is overstated
China is an important sovereign creditor to Serbia, but it does not dominate the debt portfolio. At 31 May 2026, Serbia owed the Export-Import Bank of China approximately €2.765bn, against total central-government public debt of €41.14bn. China Exim therefore represented around 6.7% of total central-government debt and approximately 9.6% of direct external debt. It was the fourth-largest creditor category after eurobonds, dinar government securities and commercial-bank loans.
This is significant exposure, particularly because China Exim financing is normally project-specific and often connected to Chinese contractors. It is not, however, evidence that China is financing the whole Expo programme. In the official project descriptions, debt reports and loan legislation reviewed for this analysis, there is no identifiable dedicated China Exim Bank or China Development Bank facility covering the central Expo complex, the National Stadium, the Expo railway, the nine trains or the Novi Beograd station reconstruction. The available project disclosures instead identify the Serbian budget or a domestic Serbian loan as the funding source.
The contrast with other Serbian infrastructure projects is clear. The $850mn Fruška Gora Corridor follows the classic Chinese financing-and-construction structure: China Exim Bank finances 85% of the accepted contract amount, Serbia supplies 15% from its budget and China Road and Bridge Corporation is the directly selected contractor. The official documentation explicitly identifies both the Chinese loan and the Chinese contractor. No comparable policy-bank financing package is disclosed for the main Expo assets.
The Expo model is therefore more accurately described as Chinese execution supported by fungible Serbian sovereign financing. Serbia can raise money through eurobonds, domestic securities and commercial-bank or multilateral loans and then allocate budget resources to contracts with Chinese EPCs. The absence of a project-specific Chinese loan does not mean the projects are being built without borrowing. It means the borrowing risk sits more broadly on Serbia’s balance sheet rather than inside a clearly ring-fenced Chinese facility.
The stadium demonstrates the real structure
The National Stadium provides the clearest example. The Republic of Serbia contracted a RSD42bn long-term investment loan from Banka Poštanska štedionica to finance stadium construction. At the official mid-August 2026 exchange rate, that is approximately €358mn. The construction contractor is PowerChina, but the lender is a Serbian bank and the loan is denominated in Serbian dinars.
The maturity is up to 132 months, including an availability and grace period of up to 36 months, followed by 32 equal quarterly principal instalments. The interest rate is floating at three-month BELIBOR plus 2.90 percentage points. There is also an upfront fee of 0.30%, equivalent to RSD126mn, before funds are disbursed.
Three-month BELIBOR stood at 4.70% in mid-August 2026, implying an indicative current interest rate of approximately 7.60%. At full utilisation of RSD42bn, that would represent gross annual interest of about RSD3.19bn, or approximately €27.2mn, before principal repayments. Actual interest will depend on the timing and size of drawdowns and future BELIBOR resets, so this is an exposure illustration rather than a forecast of the realised cash cost.
The dinar denomination eliminates direct foreign-exchange risk for the stadium loan, which is positive from a sovereign-liability perspective. However, it replaces currency risk with domestic floating-rate risk and starts from a relatively high nominal funding cost. It also illustrates the central economic feature of the Expo model: Serbia borrows, Serbia carries the interest-rate risk, a Chinese EPC receives the construction payments, and Serbia remains responsible for operating the asset after completion.
Procurement transparency is the more serious structural concern
The principal concern is not that the Chinese creditor can exercise excessive leverage over Expo assets. It is that direct contracting and special procedures weaken price discovery and make it difficult to reconcile contract commitments, budget expenditure and final project costs.
Serbia’s Fiscal Council estimates that the public sector concluded procurement contracts worth around €15bn in 2025, of which nearly €7bn was contracted outside the regular Public Procurement Law. Intergovernmental agreements accounted for approximately one quarter of the value of these exempted contracts. The Council states that the values and execution of individual bilateral agreements remain insufficiently disclosed, limiting effective assessment of whether public money is being used efficiently.
The reporting for Expo itself demonstrates the problem. A Ministry of Finance report identified approximately RSD14.9bn of Expo procurement undertaken under special procedures in 2025, while budget-execution information recorded approximately RSD66bn for the wider Expo programme. The difference is not proof of undisclosed or improper expenditure because the two figures can cover different projects and accounting categories. It does show that published information cannot yet be reconciled into a single contract-to-budget picture.
This also creates an EU-accession issue. The European Commission’s 2025 report stated that Serbia continues to circumvent ordinary procurement legislation through intergovernmental agreements and special laws, explicitly including the legislation for Expo 2027. It noted concerns involving legal remedies, tender deadlines, competition and the limited transparency of projects contracted under intergovernmental agreements.
The commercial trade-off is straightforward. Direct awards can accelerate mobilisation and reduce the risk that lengthy tender challenges derail a fixed event timetable. They also remove the competitive reference point that would establish whether the agreed EPC price, risk allocation, escalation mechanism and contingency provisions represent market value. Chinese companies may deliver efficiently, but without competing bids it is difficult to distinguish contractor efficiency from contractor pricing power.
Fiscal risk remains manageable, but overruns would be expensive
Serbia’s debt ratio does not suggest an imminent sovereign-debt problem. Central-government debt was around 43.7% of GDP in May 2026, and the Fiscal Council considers the medium-term debt trajectory sustainable. China Exim exposure of €2.765bn is material but too small to support a credible argument that China controls Serbia’s sovereign financing.
The more relevant risk is incremental. A 15% cost overrun on the Fiscal Council’s €3.4bn envelope would add approximately €510mn. A 25% overrun would add about €850mn. At an illustrative marginal financing cost of 5–6%, those overruns would generate an additional €25.5mn–€51mn of annual interest before amortisation. They would also arrive when Serbia is financing other infrastructure, possible energy-sector guarantees and broader public-investment commitments.
The debt structure adds sensitivity. Almost 79.7% of Serbia’s public debt was denominated in foreign currencies at the end of May 2026, principally euros and dollars. The dinar stadium loan is therefore useful diversification, but it does not fundamentally insulate the sovereign portfolio from external financing conditions.
Expo’s hardest fiscal test will come after the exhibition. The railway connecting the city, airport and Surčin district has a credible long-term transport function. The new trains and Novi Beograd station can also serve the wider metropolitan network. These assets have potential economic value independent of Expo attendance.
The legacy case is less automatic for the National Stadium, exhibition halls, aquatic facilities and new urban district. Their economic return depends on post-2027 event bookings, commercial occupancy, housing use, transport integration, maintenance requirements and the ability to avoid continuing operating subsidies. Construction completion does not transfer this demand risk to PowerChina or the other Chinese contractors. Once the EPC contracts are settled, the revenue and operating-cost exposure remains with Serbian public entities.
The correct risk classification
Expo 2027 should not presently be classified as a conventional Chinese debt-trap project. Chinese policy-bank debt represents only about 6.7% of Serbia’s central-government debt, and the disclosed main Expo financing is not a unified Chinese loan. Serbia retains access to eurobonds, domestic securities, commercial banks and multilateral lenders.
It should instead be classified as a Chinese-contractor-concentrated, Serbian-sovereign-financed investment programme with elevated procurement and legacy risk. Chinese EPC companies benefit from direct access to major contracts and a strategically important position across several interconnected assets. Serbia obtains rapid mobilisation and integrated project delivery but assumes the cost of debt, variations, schedule acceleration and long-term asset utilisation.
The immediate construction risk is moderate but rising as the fixed 2027 deadline approaches. The procurement and cost-transparency risk is high because the largest PowerChina contract values and amendments are not presented in a fully reconcilable public contract register. Sovereign refinancing risk remains manageable, although Serbia’s interest burden is relatively high for its debt ratio. The largest unresolved risk is the post-Expo commercial model: the infrastructure can create durable value, but only the transport components have an immediately evident demand base independent of the three-month event.
The defining feature of the programme is therefore not Chinese lending. It is the separation between who builds and who ultimately pays. China’s companies are deeply embedded in delivery. Serbia’s budget, domestic banking system and sovereign creditors carry the financial consequences.
