Scenarios and framework
Our study examines two key technology scenarios for achieving CO2 and climate neutrality in the global aviation sector by 2050. In the DACCU scenario, synthetic fuels produced from green hydrogen and CO2 captured by DAC lead to a gradual substitution of fossil fuels, eventually replacing conventional jet fuels entirely by 2050. This substitution follows an S-shaped curve, according to technology diffusion theories 55–59. Conversely, the DACCS scenario focuses on the incremental DACCS-based offsetting of continued fossil jet fuel use. To ensure comparability, the share of emissions offsets follows the same S-shaped curve of DACCU deployment, reaching 100% by 2050.
Our analysis includes two different 2050 goals for the aviation sector. The first is to achieve CO2 neutrality, that is, to reduce CO2 emissions to net-zero by 2050. In the DACCS pathway, this means offsetting CO2 emissions only. In the DACCU pathway, fuel substitution is assumed to fully eliminate CO2 emissions (except for indirect emissions, cf. Methods). Since DACCU-based fuels are expected to burn cleaner 46,47, this pathway also achieves a partial mitigation of the non-CO2 effects. Therefore, the climate benefits of the two pathways are not equal under a CO2 neutrality target. The second target, climate neutrality, on the other hand, includes non-CO2 effects and thus enables a more balanced comparison of the two technology pathways. In fact, to achieve climate neutrality both pathways must neutralize any residual non-CO2 effect with the deployment of DACCS. A schematic of how the different pathways and a business-as-usual with fossil kerosene achieve different targets is shown in Fig. 1.
Our analysis combines these different technologies and climate target scenarios while assuming rising aviation demand (cf. Methods). This comprehensive framework enables a holistic comparison of DACCU, DACCS and conventional aviation based on fossil kerosene in terms of costs, energy use, and climate impacts.
Emit-and-offset is cheaper under a CO neutrality target, but not under a climate neutrality target
We first calculate the costs of the two technology pathways to achieve CO2 and climate neutrality under our standard input assumptions (see Methods and Supplementary Tables 1–3). For CO2 neutrality, the DACCS pathway is significantly less costly than the DACCU pathway, which it outperforms by about €200 billion in 2050 (Fig. 2) and €120 billion in 2060 (see Supplementary Fig. 4). This cost difference is mainly due to the high electricity and capital costs of electrolysis in the DACCU pathway, which is essential for synthetic fuel production. The cost comparison under CO2 neutrality does not capture the full benefits of DACCU-based fuels because the reduction in non-CO2 impacts due to cleaner synthetic fuels is not reflected in the cost (see Supplementary Fig. 1). Both the DACCS and DACCU pathways achieve substantially higher costs than a business-as-usual scenario with continued fossil jet fuels use, which is cheaper than the DACCU scenario by over €500 billion.
Under climate neutrality, where the climate impacts of the two pathways are identical, the DACCU pathway has significant cost advantages over DACCS, which it outperforms by over €280 billion in 2050. The higher cost of the DACCS pathway is mainly attributable to the higher carbon removal rates required to offset non-CO2 emissions, which are higher than in the DACCU pathway (see Supplementary Figs. 1–2). The large offset requirements are due to the sustained demand growth assumed in the analysis. However, assuming no growth of the sector still results in a competitive advantage of the DACCU pathway (see Fig. 5b). Despite its economic advantage, the DACCU pathway results in higher electricity consumption due to energy-intensive electrolysis (cf. Supplementary Fig. 3). This limits its scaling potential to regions with abundant and affordable renewable energy. Finally, both DACCS and DACCU pathways are more expensive alternatives compared to the continued use of fossil kerosene, highlighting the role of policy interventions to propel these pathways forward.
Emit-and-offset is more expensive than synthetic fuels on a cost-per-avoided-emissions basis, but is more efficient in scaling DAC
Looking at the total costs for abated emissions relative to the business-as-usual (Fig. 3a), the resulting picture is almost opposite than the one drawn when looking at absolute yearly costs (Fig. 2). Under the CO2 neutrality target, the DACCS pathway has the highest costs per emissions abated, reaching abatement costs over €500/tCO2e compared to less than €200/tCO2e for the DACCU pathway. This difference arises because DACCS only includes costs associated with reducing CO2 emissions. Conversely, in the DACCU pathway, the abatement extends to non-CO2 emissions, thereby increasing the total volume of abated emissions over which the costs are distributed. Under the climate neutrality target, where both technology pathways abate the same level of emissions, DACCU again emerges as more cost-effective because of the smaller amounts of carbon removals required to offset the remaining non-CO2 effects.
Apart from mitigating the aviation sector, both options could also serve as a means of scaling up DAC. This rationale is rooted in the potential role that the aviation sector could play as a niche for the initial deployment of DAC, as the sector is bound to face significant costs in mitigating its emissions due to the lack of affordable alternatives. This perspective results in a picture opposite to that of cost-effective abatement. We find that as the volume of DAC installations increases, the DACCS pathway consistently offers a lower cost per DAC unit than the DACCU pathway (Fig. 3b). DACCU incurs higher costs due to the production of green hydrogen. This has a significant impact on the cost per unit of DAC installed.
The price difference for a CO2-neutral flight with DACCS and DACCU is small
We further assess the increase in price per flight per passenger to achieve CO2 and climate neutrality via the DACCS and DACCU pathways. In the context of CO2 neutrality, offsetting aviation CO2 emissions with DACCS proves to be more economical than fueling the same flight with DACCU-based synthetic fuels. However, the cost difference per passenger is modest, ranging from approximately €20–55 for long-haul flights (London-New York and London-Perth) to only €4 for a short-haul flight from London to Berlin. While the overall cost per passenger increases to achieve climate neutrality, DACCU becomes cheaper than DACCS, saving about €35–100 per passenger on long-haul flights and €6 on short-haul flights.
We also assessed the impact on the cost of flying relative to the expected future cost of flying in a business-as-usual scenario with continued use of fossil fuels. The projected increase in ticket prices for flights in 2050 ranges between 15–30% for DACCU and 8–20% for DACCS to achieve CO2 neutrality, rising to up to 40% (DACCU) and 60% (DACCS) to achieve climate neutrality. However, the increase in price is not the same for all flights, since the contribution of fuel costs to ticket prices varies for different routes, as the price is adjusted to demand and to endure competition. While the increases in price due to a complete neutralization of the climate effects of a flight may seem substantial, they lie well below the range of current variance in prices. Indeed, the difference in price between buying a ticket two weeks or two months in advance is, on average, 400% for the London-Berlin route, over 100% for the London-New York route, and 70% for the London-Perth route 60.
Cheaper electricity and high fossil jet fuel prices can make DACCU cheaper than DACCS (and even business-as-usual) even under CO 2 neutrality
To understand the conditions under which DACCU-based fuels could be economically competitive in the less-advantageous CO2-neutrality scenario with an emit-and-offset strategy via DACCS and even with the business-as-usual with continued use of fossil jet fuel, we perform local sensitivity analyses on the most influential parameters (see Supplementary Table 1–3).
Figure 5a shows that DACCU can become more cost effective than DACCS when electricity prices fall below 0.015 €/kWh. This threshold is well below the 2023 price of the cheapest renewable energy sources, onshore wind 61, but not unachievable in the future through technology learning, optimal siting, or in moments of excess production of renewable electricity, for example on sunny summer days in grids with a high share of solar PV62,63. In contrast, even when powered by free electricity, DACCU is still not competitive with the business-as-usual.
Conversely, rising fossil fuel prices prove transformative: DACCU becomes cost-competitive with DACCS at a fossil fuel price of €0.9/L and with the business-as-usual scenario at €1.8/L. Such high costs would not only make DACCU a more economical option, but would also discourage demand. However, doubling the current price of fossil jet fuel would require dedicated political ambition.
Accelerated technological learning and steeper learning curves benefit both DACCU and DACCS scenarios. Thus, even a learning rate of 50% - higher than has been observed historically for fast-learning technologies such as solar PV - cannot close the gap between the DACCU and DACCS pathways.
In summary, extremely optimistic changes in fossil fuel or electricity prices are required to make DACCU cost-competitive with DACCS or business-as-usual by varying a single parameter. However, a synergy of lower electricity prices with either rising fossil fuel costs or higher technological learning could accelerate a scenario where DACCU outperforms DACCS or even fossil jet fuels under optimistic but possible conditions (see Supplementary Figs. 6–8).
Pricing aviation climate impacts or limiting DACCU operation to times when excess electricity is available is sufficient to make DACCU cheaper than DACCS
Given the observed sensitivity of DACCS and DACCU performance to highly uncertain input assumptions, we examine the potential impact of different policies that affect these assumptions. Figure 6 shows the cost difference of DACCU compared to DACCS (Fig. 6a) and fossil jet fuel (6b) under different policies affecting some of the key input variables (see Supplementary Table 4).
Pricing emissions internalizes the impact of continued fossil jet fuel emissions and thus acts similarly to increasing the price of fossil fuels, while also internalizing the environmental costs of life-cycle emissions for both DACCS and DACCU. Conversely, pricing CO2 emissions alone cannot make DACCU cost-competitive with DACCS since, under CO2 neutrality, it applies only to indirect emissions, which are higher in the DACCU pathway (see Supplementary Fig. 1). On the other hand, pricing all aviation-related climate impacts can significantly favor the DACCU pathway, which already becomes more cost-effective than the DACCS pathways at €30/tCO2e*. Pricing emissions is also crucial to make DACCU economically competitive with fossil jet fuels. However, the prices on emissions need to be extremely high, starting at €500/tCO2 for CO2 emissions alone and at least €100/tCO2e* for all aviation-related impacts.
In contrast to direct subsidies based on synthetic fuel production, which are not sufficient to make DACCU competitive with DACCS even at €500/tfuel (which corresponds to about €1600/tCO2 DACCU-based fuels), a strategic approach is to leverage cheap electricity (below €0.01/kWh). Policies, such as seasonal restrictions aligned with periods of electricity surplus, could achieve this by limiting DACCU-based synthetic fuel production to periods of significantly cheaper surplus electricity. However, this approach comes with the constraint of limiting the volume of DACCU-based synthetic fuels that can be produced. While limiting the number of operating hours could increase the weight of capital expenditures per DACCU output, and thus lead to potential cost increases not accounted for in our modelling64, it could also reduce the deterioration, and thus extend the lifetime, of costly components of the electrolyzers and DAC, namely the stack and adsorbent.