Serbia’s manufacturing sector continues to do the heavy lifting. In 2025 it absorbed shocks, sustained export volumes, and delivered the bulk of industrial growth. By any conventional reading of the statistics, manufacturing looks resilient. Yet that resilience now conceals a structural inversion that becomes decisive in 2026: manufacturing is no longer being supported by the energy system it depends on. Instead, it is compensating for the energy system’s weaknesses. This reversal is gradual, largely invisible in headline data, and therefore easy to underestimate—until it begins to cap growth outright.
The industrial figures from 2025 show manufacturing output rising by roughly 2.5–3.0%, driven by processing industries tied to EU supply chains. Metal products, rubber and plastics, machinery components, and selected chemical segments performed well. This confirms Serbia’s continuing role as a near-shore manufacturing base for Europe. It also explains why GDP growth held near 2.0% despite weakness elsewhere. But the same data show that electricity, gas, and steam supply stagnated or declined. Construction linked to industrial expansion weakened. Energy-intensive subsectors faced higher cost volatility. The result is an economy in which the load-bearing sector is quietly losing the structural support that made its expansion economical in the first place.
Manufacturing growth can persist for a time even as its foundations erode. Firms respond by squeezing margins, postponing upgrades, and intensifying use of existing assets. That is exactly what the 2025 pattern suggests. Output rose, but investment did not keep pace. Capacity utilisation increased, but capacity expansion slowed. In the short run, this looks like efficiency. In the medium run, it becomes fragility.
The reason energy matters so profoundly is not only that it is an input, but that it sets the marginal cost environment for every industrial decision. When energy supply is abundant, reliable, and improving in efficiency, manufacturing expansion lowers unit costs over time. When energy supply stagnates and carbon intensity remains high, each additional unit of output raises exposure to price volatility and regulatory repricing. The data indicate Serbia has moved closer to the second configuration.
Energy supply underperformance in 2025 was not a one-off. It reflects years of underinvestment, delayed grid reinforcement, and continued reliance on lignite-heavy generation. As a result, industrial producers face electricity costs that are volatile and increasingly misaligned with EU benchmarks, while embedded emissions rise relative to competitors. This combination matters more in 2026 than it did in 2022 or 2023, because the external environment has changed. EU buyers are no longer indifferent to how energy is sourced. CBAM and related procurement rules turn energy into a competitiveness filter.
Manufacturing firms feel this pressure first at the margin. Contracts shorten. Price negotiations tighten. Buyers ask for more documentation, more guarantees, more proof of transition. None of this shows up immediately in production indices. It shows up in investment decisions. When uncertainty around energy costs and carbon exposure rises, firms delay capacity upgrades, electrification projects, and automation. This delay feeds back into productivity, reinforcing the problem.
The statistical divergence between manufacturing growth and construction decline is therefore not accidental. It reflects a system where factories keep running, but fewer new ones are built, fewer lines are modernised, and fewer grids are reinforced. Manufacturing is drawing down the option value of past investments without replacing it. Over time, that reduces resilience.
This dynamic is particularly visible in energy-intensive manufacturing. Metals processing, chemicals, cement-linked segments, and heavy fabrication remain active, but their cost base has become more sensitive to energy volatility. Even where output volumes hold, margins thin. In export markets, thinner margins translate into weaker pricing power. Volume can grow while value stagnates. For a capital-intensive sector, that is a warning sign.
By contrast, less energy-intensive manufacturing segments perform better on a risk-adjusted basis. Assembly operations, component manufacturing, and processing with lower heat demand show greater resilience. This divergence within manufacturing mirrors the divergence between manufacturing and energy supply at the macro level. It is a signal of how the system is reallocating stress.
The temptation is to read manufacturing’s continued growth as evidence that energy issues are manageable. That interpretation mistakes lag for immunity. Manufacturing reacts to energy constraints with a delay because firms can absorb costs temporarily. They cannot do so indefinitely. The constraint becomes binding not when output falls, but when investment stops. The 2025 data suggest investment is already hesitating.
In 2026, this hesitation becomes more consequential. With GDP growth moderating to 1.5–2.5%, manufacturing remains the primary engine. But engines require fuel and maintenance. Energy supply that does not expand in capacity or improve in carbon intensity effectively taxes each additional unit of manufacturing output. The tax is not always explicit; it appears as higher financing costs, stricter buyer terms, and deferred capex.
This is where CBAM magnifies existing weaknesses rather than creating new ones. Manufacturing firms exporting into the EU do not face CBAM as a single charge. They face it as a bundle of expectations. Energy-intensive production without a credible transition path is treated as higher risk. Higher risk reduces contract duration and increases required returns. That, in turn, discourages the very investments needed to reduce risk. A feedback loop forms.
The energy system sits at the centre of that loop. Without accelerated investment in generation mix, grids, and flexibility, manufacturing’s growth ceiling lowers. Output can still expand at the margin, but productivity gains slow and competitiveness erodes. The economy appears stable until it is suddenly not.
The strategic implication is uncomfortable but clear. Manufacturing cannot continue to carry Serbia’s growth on its own if the energy system remains static. Energy is no longer a background condition; it is an active determinant of industrial viability. Treating energy policy as separate from industrial policy is no longer analytically defensible.
In practical terms, restoring alignment requires energy investment to move ahead of manufacturing demand, not behind it. Grid reinforcement must anticipate industrial clusters rather than react to congestion. Generation mix must shift in ways that reduce marginal emissions and price volatility, not merely add capacity. Without this shift, manufacturing’s apparent strength becomes a liability, because it masks the erosion of the system that supports it.
Serbia’s manufacturing sector is not failing. It is compensating. That distinction matters. Compensation can last for a few years; structural support determines whether it lasts a decade. As 2026 unfolds, the question is not whether manufacturing can still grow. It is whether the energy system will allow that growth to remain economically rational.
Manufacturing still carries Serbia. But unless energy begins to carry manufacturing again, that load will become unsustainable—quietly at first, and then all at once.
