Three industries, three different routes to decarbonisation. All of them arrive at the same two energy carriers.

Glass, pulp and paper, and cement in the Basque Country have almost nothing in common technically, and new site-level modelling for the TRANSIENCE project finds each needs its own transformation route. Every route ends up drawing on electricity and biomass. How much of that biomass can be sustainably supplied is a question these results leave open.

TRANSIENCE project · 31 July 2026 · ~7 min read

Status: results from Deliverable D8.1, not yet approved by the European Commission and subject to revision.

Industrial decarbonisation is usually discussed in terms of technologies. But industrial plants combine high energy demand, long investment cycles, site-specific process configurations and exposure to international competition. Pathways therefore cannot be assessed from national energy balances or generic technology lists alone.

What is needed instead is a representation of individual sites: the production capacities that already exist, the process options available to each, and the interaction between energy prices, CO₂ prices and the timing of investment. A technology that looks cheap initially may be unavailable when a plant next reinvests, or may not fit the process it would serve.

The Basque Country is a useful place to test that proposition, with a strong industrial structure containing several energy-intensive activities relevant to decarbonisation. This study covers three sectors — glass, pulp and paper, and cement — identified in consultation with stakeholders as the most relevant for transformation and as having the highest impact on the energy system, including infrastructure requirements.

They differ strongly in process characteristics, and the differences matter for what follows. Glass is dominated by high-temperature furnace technologies with demanding quality requirements and growing scope for electrification. Paper is defined by its substantial steam demand. Cement emits CO₂ both from fuel combustion and from the calcination reaction itself, and therefore needs a broader portfolio: fuel switching, alternative calcination technologies, and carbon capture and storage.

What we modelled

The analysis runs in two steps, and the two are deliberately different in kind.

The first is an abatement-cost comparison — a screening exercise that takes the extra cost of a low-carbon technology relative to a conventional reference, covering investment, operation, maintenance and energy, and divides it by the emissions avoided. A negative value means the low-carbon option is cheaper than the reference under the assumed prices. The indicator is easy to communicate and makes clear which assumptions drive the answer, but it is marginal and static: it cannot represent plant vintages, technology availability, reinvestment timing, interactions between process units, or infrastructure constraints.

The second step addresses those gaps. FORECAST-Sites is a site-level optimisation model that minimises the total cost of investment and operation across each site, discounting future operating, energy and CO₂ costs at 8%. It represents process categories, technology options, energy-carrier requirements, the years in which technologies become available, lifetimes, costs, and emission or capture rates. Where the screening step ranks technologies, the model produces pathways — when reinvestment occurs, and how the energy-carrier mix changes.

The sequencing is the point. The screening step supplies intuition; the model tests whether that intuition survives once plant vintages, changing prices and technology availability are considered together.

Prices drive much of the result. Energy-carrier prices are based on Eurostat statistics for energy-intensive consumers together with additional projections. Electricity is held constant at 70 €/MWh across the whole model horizon and natural gas at 46.1 €/MWh; biomass stays around 38–41 €/MWh, and hydrogen falls from 165 €/MWh to 120 €/MWh in 2030 and around 90 €/MWh from 2035. The CO₂ price rises linearly from 70 €/t in 2025 to 200 €/t in 2050 — the assumption that most directly determines when fuel switching, electrification and capture become attractive. Sensitivities of ±30% on electricity and gas prices over ten years are included.

Each sector is represented through its own process structure. Glass is modelled as furnace processes across flat, container, and special glass, with regenerative, recuperative, hybrid (80% electric), fully electric and post-combustion capture options. Pulp and paper separates pulp production, paper production and steam generation, so heat supply can be analysed independently; steam options include combined heat and power, biomass boilers, gas, hydrogen or electric boilers, and heat pumps. Cement is broken into calcination, rotary kiln, the burners serving each, and a proxy process representing the choice between conventional and low-carbon cement.

One framing point matters throughout: this case study was built to demonstrate what a site-level model can do, not to produce a regional roadmap. The results are first model-based insights, not investment recommendations.

Glass: hybrid furnaces arrive early, full electrification later

The screening step is unambiguous for glass. Hybrid electrification is the most competitive option, at abatement costs below 50 €/t CO₂ — and under the assumed prices, combining it with natural gas can produce negative abatement costs, meaning it is cheaper than the conventional furnace. The advantage comes not only from substituting electricity for gas but from the efficiency gain of the hybrid furnace design itself.

The optimisation agrees, and adds timing. Furnace lifetimes are short at 15 years, so the model captures reinvestment in 2030 and 2035 and again in 2045 and 2050. Full electrification becomes more attractive in the long run, but only above a CO₂ price of roughly 100 €/t, and depends on assumed technology maturity, capital cost and lifetime. Carbon capture enters from 2040 to address both energy-related and process emissions, bringing the modelled glass sites close to carbon-neutral production.

The energy mix shifts accordingly. Electricity demand rises from around 0.14 million GJ a year in 2025 to around 0.86 million GJ from 2040, while natural gas falls from around 1.59 million GJ to around 0.17 million GJ over the same period.

Pulp and paper: decarbonisation is a steam question

Because paper production is defined by steam demand, the technology question is how that steam is raised. Screening and optimisation both identify heat pumps as the most competitive technology for low-temperature steam, with biomass boilers becoming more cost-effective than biomass combined heat and power after 2035 — a ranking that can change with the electricity price.

The model concentrates investment in two waves, around 2030 and 2035, with heat pumps serving low-temperature steam while biomass CHP or boilers cover higher-temperature demand. Emissions fall primarily through the phase-out of natural gas in favour of electricity and biomass, reaching an overall carbon-neutral balance from 2030 once biogenic CO₂ is accounted as previously captured.

The energy shift is the largest of the three sectors. Natural gas falls from around 7.21 million GJ a year in 2025 to about 2.54 million GJ in 2030 and to zero from 2035. Electricity rises from about 4.17 million GJ to about 7.81 million GJ from 2030. Biomass dips initially and then climbs to around 4.46 million GJ from 2035.

The sector also illustrates the limits of the exercise. Heat pumps can be cost-attractive while remaining hard to deploy, because performance depends on achievable coefficients of performance, integration into existing process heat systems, technology readiness at higher temperatures, and a skilled workforce to install them. These local conditions could in principle be modelled; the information needed is scarce.

Cement: no single technology is sufficient

Cement is where the screening step is least able to answer on its own, because fuel switching alone cannot eliminate process emissions and the necessary technology combinations are hard to represent in a marginal cost comparison. What it does show is a clear ranking between two capture strategies: separating calcination so that process CO₂ is captured as a nearly pure stream — the Leilac calciner with carbon capture — outperforms post-combustion capture on a conventional cyclone calciner, regardless of the electricity and CO₂ price.

The optimisation sets out a staged pathway. Around 2030 the model invests in a new burner and switches the fuel mix towards biomass and natural gas. Around 2035 it adds Leilac calcination. Around 2040 post-combustion capture is added to catch remaining process and energy-related emissions. A burner renewal follows around 2050, keeping the established fuel mix.

Emissions fall in three stages: fuel switching to biogenic sources removes energy-related emissions while process emissions remain; Leilac captures a major share of those process emissions; and capture applied to biomass combustion then makes overall negative emissions possible. The fuel mix changes substantially — non-renewable waste, other renewables and petroleum coke disappear after 2025, biomass becomes dominant from 2030 and grows further from 2040, natural gas serves as a complement, and electricity rises in 2040 as capture comes online.

Three routes converge on two energy carriers

The three sectors follow genuinely different logics. Glass is a matter of furnace substitution. Paper is a matter of steam supply. Cement requires a portfolio because fuel switching cannot reach the calcination reaction. Yet electricity becomes more important in all three, for three different reasons — substituting part of the furnace heat input in glass, powering heat pumps and electric steam technologies in paper, and supporting new process configurations and carbon capture in cement.

That convergence has a direct implication: regional electricity infrastructure and prices become decisive boundary conditions for industrial transformation — not a sectoral concern but a shared one, arriving from three directions at once.

Biomass is the second recurring option, particularly in paper and cement, and the model finds it highly attractive on cost. This is an important result and also a warning sign. The model contains no explicit sustainability, availability or competing-use constraints on biomass, so a cost-minimising solution may overuse it. Future applications should test biomass potential limits or scenario-based restrictions, because unavailability of biomass would close off decarbonisation options the model currently treats as open.

The pairing of the two methods is itself part of the finding. Screening explains why hybrid electrification in glass and heat pumps in paper look attractive — their efficiency and cost structure make the case directly. But industrial sites do not invest continuously. They invest when equipment reaches the end of its life, when technologies become available, and when price signals make alternatives worthwhile. Turning intuition into a pathway requires the second step.

Limitations

A demonstration, not a roadmap. The case study exists to show how a site-level framework can structure technology comparison, identify cost-optimal pathways and support stakeholder discussion. The results are first model-based insights rather than final investment recommendations.

Unconstrained inputs. Neither biomass availability nor infrastructure capacity is bounded. Representation of infrastructure capacities and costs for CO₂, electricity and hydrogen needs further development, as does flexible fuel switching between hydrogen and natural gas, and flexible operation under electricity markets at high temporal resolution.

Sector-specific sensitivities. In paper, heat pump deployment is sensitive to the coefficient of performance and to local process integration. In cement, biomass availability and sustainable fuel-mix constraints are central. In glass, the long-term result depends strongly on the future maturity, cost and lifetime of fully electric furnaces.

Data and citation

• Scenario results are available on IAM PARIS.
• Model: FORECAST-Sites · Region: Basque Country · Sectors: glass, pulp and paper, cement · Years: 2025–2050
• Industrial sites: Fraunhofer ISI industrial site database, adapted in cooperation with TECNALIA.
• Full methodology, sector parameterisation and price assumptions: TRANSIENCE Deliverable D8.1 (will be soon available on the TRANSIENCE website)
• Licence: CC BY 4.0 — charts and text may be reused with attribution.

This explainer accompanies work carried out under the TRANSIENCE project, funded by the European Union. The underlying deliverable has not yet been approved by the European Commission and its findings may be revised.

The AI assistant Claude (Opus 5, Anthropic) was used to support the writing of this explainer. All findings, figures and interpretations derive from Deliverable D8.1 and have been reviewed and approved by the authors, who take full responsibility for the content.