Modeling formula one’s path to net zero by 2030 - a systems approach for evaluating decarbonization strategies

In 2019, Formula 1 (F1) committed to achieving net-zero carbon emissions by 2030 across its entire value chain; however, this target has drawn both support and skepticism from the academic community. We present an exploratory systems scenario model using F1’s official sustainability reporting data to evaluate seven decarbonization interventions through two complementary analytical modes: a static calculator that enables direct scenario comparisons, and a time-dynamic simulator that projects adoption trajectories from 2024 to 2030 using gap-closing stock dynamics and an emissions-pressure feedback loop. Static sensitivity analysis shows that a 50% absolute reduction threshold is achievable under moderate actions. However, achieving net zero requires simultaneously maximizing all levers and offsetting a residual emissions balance whose magnitude depends on factors such as the quality and supply of sustainable aviation fuel (SAF) across the lifecycle. Finally, we provide an interactive dashboard allowing users to test different intervention combinations in real time.

Six years into F1’s net-zero pledge, which covers Scopes 1, 2, and 3 emissions, the organization has successfully reduced its restated 2018 baseline of 228,793 tCO 2 e to 168,720 tCO 2 e in 2024 (Formula One World Championship Limited, 2025). Yet critics have characterized these efforts as greenwashing (Sturm et al., 2025) and doubt persists as to whether the remaining reductions can be achieved before the 2030 deadline. This uncertainty motivates the development of a tool to assess whether specific intervention combinations can deliver net zero under realistic assumptions.

Despite the 26% reduction achieved so far, reaching net-zero emissions by 2030 requires eliminating or offsetting the remaining 168,720 tCO 2 e, a challenge complicated by a 24-race calendar spanning five continents, thousands of personnel traveling to each event, and air freight operations on tight turnaround schedules. The sport’s expansion trajectory compounds this further. Cadillac’s arrival as an eleventh constructor in the 2026 season introduced an operational footprint across four facilities on two continents, three in the United States and one in Silverstone, United Kingdom (Cadillac Formula 1 Team, 2025), adding manufacturing, logistics, and travel emissions that were not present in the 2018 baseline. Moreover, these figures carry a caveat that applies to all carbon accounting: Scope 3 emissions are inherently difficult to measure. Unlike Scope 1 (direct emissions from sources owned or controlled by the organization) and Scope 2 (indirect emissions from purchased energy), Scope 3 covers upstream and downstream activities across the entire value chain. More simply, Scope 3 includes the Scope 1 and Scope 2 emissions of an organization’s suppliers, freight providers, and commercial partners. F1 acknowledged this in its 2025 report (Formula One World Championship Limited, 2025) by restating the 2018 baseline to 228,793 tCO 2 e, down from 256,551 tCO 2 e, after obtaining improved supplier data—a 10.82% downward revision that illustrates how sensitive such estimates are to measurement methodology. Because these activities fall outside the reporting organization’s direct control, they typically rely on spend-based estimates or industry-average emission factors rather than primary data, introducing substantial measurement uncertainty, which is permitted by the Greenhouse Gas Protocol’s Scope 3 standard (World Resources Institute, 2024).

F1 has stated that it will prioritize a minimum 50% reduction in absolute emissions against the 2018 baseline before resorting to offsetting (Camm, 2025) with the 2030 target depending heavily on the credibility of reported emissions changes. Existing literature (Al-rubaye and Naimi, 2025) notes that the Fédération Internationale de l’Automobile (FIA) shows year-on-year volatility in its reporting, partly driven by methodological updates rather than real emission changes. Our model therefore propagates uncertainty from these published figures.

Scholars have also questioned whether F1’s efforts are substantive, describing the initiative as largely superficial and symbolic (Sturm et al., 2025) and arguing that F1 uses greenwashing rhetoric to account for what are negligible transformations in relation to its very large carbon footprint. They claim that the sport selectively discloses its purportedly environmentally friendly developments while concealing, deflecting, or downplaying more significant issues surrounding travel and transport (Sturm et al., 2025). Some studies have adopted a more analytical approach to assess this matter by applying a triple bottom line framework to F1’s SDG alignment and found that although F1 is actively seeking to become a green sport, it is not yet at a sufficient level, and point to its reasoning being partly due to the country distances not being taken into consideration in the race calendar (Öztopcu, 2023). Other studies report a 13% emissions reduction by 2022 and 75% of circuits operating on renewable energy by 2023, but argue F1 must “persistently raise the bar” to survive greenwashing accusations (Chiles, 2025). FIA emissions in 2023 were 36% above the Science-Based Target Initiative (SBTi) target of 15,500 tCO 2 e, with their model projecting an 18% increase through 2030 under current trends.

A shared limitation of this body of work is that the existing literature is largely descriptive, offering qualitative assessments and trend forecasts without providing a mechanism to test specific interventions. Consequently, stakeholders, fans, sponsors, and regulators cannot currently evaluate whether F1’s stated commitments are achievable under specific assumptions regarding technology, calendar design, and offset quality. We address this gap by developing an exploratory systems model calibrated to official F1 data, offering seven adjustable policy levers and two analytical modes: a static calculator for instant scenario comparisons, and a time-dynamic simulator that projects adoption trajectories to 2030 using a feedback loop that captures diminishing returns. This study identifies the combination of actions required to realistically achieve net-zero emissions by comparing the relative effectiveness of intervention combinations, moving beyond the question of whether F1 is greenwashing.

Systems approach

Systems dynamics (Forrester, 1968; Sterman, 2000) Modeling is well-suited to problems involving feedback loops, delays, and policy testing under uncertainty. We adopt a systems approach over other approaches such as lifecycle assessment (LCA) and econometric regression, as LCA is designed for attributional accounting, whereas the present study seeks to evaluate prospective interventions. Econometric regression can project trends, but our aim is to simulate counterfactual policy scenarios not yet implemented; thus, we chose a systems approach that allows stakeholders to adjust intervention levers and observe the consequences immediately, motivating the inclusion of static and dynamic analytical modes. Figure 1 shows the model’s causal loop diagram, illustrating the reinforcing and balancing loops that govern emissions trajectories. The reinforcing loop captures calendar-driven growth in freight and travel, while the balancing loops represent the decarbonization levers (SAF, renewable electricity, intermodal shift, and remote operations), whose strength is modulated by emissions-based pressure.

Data sources and baseline

The model draws baseline emissions from the F1 2025 Sustainability Update (Formula One World Championship Limited, 2025) which restated the 2018 baseline at 228,793 tCO 2 e (down from the originally reported 256,551 tCO 2 e, reflecting improved supplier data). Power unit emissions are derived from the F1 2019 Sustainability Strategy (Formula One World Championship Limited, 2019)which reports on-track fuel combustion at 0.7% of total annual emissions. Calendar regionalization potential is based on Struijk (2024), who noted that maximum calendar regionalization could achieve up to a 62% distance reduction (Struijk, 2024). Table 1 lists all baseline constants.

Modal emission intensities are drawn from the GLEC Framework (Smart Freight Centre, 2026) on a well-to-wake basis and cross-checked against the UK Government greenhouse gas conversion factors (DESNZ, 2025). We set the rail and sea emissions factor at 10% of the air freight factor. There is variation in estimates of published modal comparisons depending on many associated contexts, including type of vehicle, load, and fuel, among other things. Across averages, deep-sea container freight is estimated around 2% and rail freight between 4 and 8% of the per tonne-kilometer emissions of long-haul air freight (Smart Freight Centre, 2026; DESNZ, 2025). However, these modal factors do not fully represent a realistic intermodal freight alternative for Formula 1 in this case. The alternative we consider would theoretically combine deep-sea container shipping for intercontinental races with articulated road and rail transport across the European calendar. On average, road freight has an emissions factor of approximately 9% of air freight on average, while roll-on/roll-off ferry transport is approximately 6% on average (Smart Freight Centre, 2026; DESNZ, 2025). It is possible that some freight would also require temperature-controlled handling; any temperature-controlled handling would increase the respective modal emissions factor by several percentage points. All of these emission factors are before accounting for terminal and logistics hub emissions. We therefore adopt 10% as the central estimate to account for pre-carriage, on-carriage, and handling emissions not captured in the raw modal factors. This parameter is varied from 0 to 20% in the sensitivity analysis (Section 2.5).

We chose to separate power unit emissions (1,602 tCO 2 e at baseline) from event operations because of the FIA’s 2026 mandate for 100% sustainable fuel, which allows us to model this source as zero from 2026 onward (Formula One World Championship Limited, 2019, 2026; Federation Internationale de l’Automobile, 2024). Notably, this treatment reflects an accounting convention and does not imply zero life cycle emissions. Under the Greenhouse Gas Protocol, biogenic and synthetic carbon from fully sustainable drop-in fuels is reported separately from Scope 1 combustion emissions. Life cycle emissions associated with fuel production remain in the upstream Scope 3 categories included in the model. Because on-track combustion accounts for only 0.7% of baseline emissions, alternative accounting treatments for this term do not materially affect the results.

We split business travel into VIP (15%) and staff (85%) categories; VIP personnel, including drivers, team principals, and senior sponsors, must attend every race, whereas staff roles such as broadcast operations can partially shift to remote work. F1 does not publish this breakdown by role; this parameter is an estimate based on our best knowledge and judgment.

We also include offset effectiveness as an uncertain model parameter. Previous studies have utilized a variety of ex post methods to estimate several offset types’ effectiveness, with broad variation in results. For example, Cames et al. (2016) estimated that for Clean Development Mechanism projects, approximately 85% of the projects had a low likelihood of additionality (Cames et al., 2016). Similarly, Calel et al. (2025) estimated that 52% of assessed reductions would have occurred without the credit; Badgley et al. (2022) reported approximately 30% over-crediting in California forest offsets; Gill-Wiehl et al. (2024) found cookstove credits to be over-issued by a factor of 9.2; and Probst et al. (2024) concluded that fewer than 16% of issued credits across project types correspond to real reductions (Calel et al., 2025; Badgley et al., 2022; Gill-Wiehl et al., 2024; Probst et al., 2024). However, much of this evidence reflects credits issued under standards that have since been updated. The Paris Agreement Crediting Mechanisms and the changes made by the Integrity Council for the Voluntary Carbon Market and CORSIA Phase I have both sought to increase requirements for credit quality. Given that our model covers purchases from 2024 to 2030, we selected 70% as the static calculator default to reflect improvements in the effectiveness of credits since these peer-reviewed studies were conducted. We vary this parameter from 20 to 100% in our sensitivity analysis as described in Section 2.5.

Model architecture

The dashboard consists of two tabs; both use the same emission equations and differ only in how lever values are determined.

Tab 1: Static what-if calculator

This tab consists of seven sliders that configure the policy levers directly: SAF adoption (0–100%), rail/sea share (0–60%), remote work (0–100%), renewable electricity (0–100%), carbon offsets (0–100%), calendar size (18–26 races), and regionalization (0–100%). This includes four preset scenarios: (a) 2018 Baseline, (b) 2023 Level, (c) 2024 Level, and (d) Net Zero Required, to enable quick comparison. The historical year presets override calculated values with actual F1-reported figures to match the published data exactly.

Tab 2: Time-dynamic simulation

This tab consists of adoption speeds (%/yr) instead of lever positions. This consists of the six stocks: SAF adoption, intermodal shift speed, calendar size, remote work, clean grid growth, and offset rate that evolve from 2024 (present) to 2030 via forward Euler integration (dt = 0.25 years, 25 steps), with regionalization set per scenario and held fixed. Each stock follows gap-closing dynamics:

Where S (t) denotes the current adoption level of a stock, S _ max its ceiling, and rate the user-set adoption speed (%/yr). P (t) is the emissions-pressure multiplier.

The rate of change is proportional to the remaining gap between the current value and the stock’s ceiling, producing rapid early gains that decelerate as the ceiling approaches. The offset stock is an exception, as it uses a linear ramp unmodulated by pressure, because offset purchasing is a financial decision with no natural saturation curve. Calendar size is held constant. The dynamic model enables the emissions-pressure feedback loop. As gross emissions fall, the political and economic incentive to further reduce emissions weakens, reflecting an empirical pattern in decarbonization where the last tons cost more than the first. Supplementary Section 3.2 details the theoretical basis for this mechanism. We express this formulation as:

With σ = 1.0, at baseline emissions, the pressure multiplier is unity P (t) = 1.0 and adoption proceeds at full speed. As emissions approach zero, P (t) declines to 0.30. We set this floor to capture the residual share of adoption pressure that remains based on regulatory, contractual, or reputational constraints. We do not have a basis on which to calibrate either the floor or the exponent, so we adopt unity as the simplest specification and vary both across multiple ranges (Supplementary Section 5.4). These assumptions affect the pace of decline but do not alter the relative rankings of the levers or the qualitative conclusions.

Emissions categories equations

Each emission category is calculated from its baseline value, adjusted by calendar multipliers, policy lever positions, and SAF life cycle reductions. Two calendar effects drive logistics and travel emissions: a calendar multiplier that scales linearly with the number of races relative to the 2018 baseline of 21, and a distance multiplier that captures regionalization benefits. Their product forms the effective calendar factor (ECF). We applied linear scaling to the calendar multiplier to simplify freight proportioning. We assume that a portion of freight moves as pre-positioned sea freight sets whose emissions are not fully proportional to race count. In addition, there will be some travel and facility activity that is fixed regardless of calendar length. Linear scaling therefore likely overestimates the impact of calendar size on total emissions. As with other parameters, we further investigate this assumption through the uncertainty analysis described in Section 2.5.

Logistics emissions are calculated by applying the ECF to the baseline freight value and splitting freight into two streams: the air-freight share is reduced by the SAF blend ratio and its life cycle factor, whereas the rail/sea share uses a lower emission factor. VIP travel scales the business travel baseline by the VIP share (15%) and ECF, with SAF reductions applied to the aviation component. Staff travel follows the same structure but is further reduced by the remote-work fraction, as staff who work remotely do not travel. We assume that reductions in staff travel are proportional to the share of work performed remotely. For example, a 30% remote-work share reduces staff travel emissions by 30%. This approach assumes equal travel intensity across remote-capable and non-remote roles. Consequently, the estimated reduction will be understated if remote-capable staff travel more than average and overstated if they travel less.

Event operations scale with the calendar multiplier and are reduced by the adoption of renewable electricity. Facility emissions depend only on renewable electricity, as they occur year-round and remain independent of the race calendar. Power unit emissions follow a hard cutoff: they equal the baseline scaled by the calendar multiplier before 2026, and zero thereafter, reflecting the FIA’s mandate for 100% sustainable fuel. Emissions are calculated as gross and net. Gross emissions reflect the reduction in emissions relative to the restated 2018 baseline of 228,793 tCO 2 e, achieved only through operational levers that avoid or remove the emissions at the source, for example, renewable energy, intermodal shift, reduced travel, etc. Net emissions are gross emissions after subtracting the share covered by carbon offsets, which we calculate as the product of the offset rate, offset effectiveness, and residual gross emissions. Therefore, we treat offsets as a distinct accounting term rather than as a form of abatement. This structure reflects the fact that an offset does not reduce the emissions the organization produces. It finances a claimed reduction elsewhere whose validity depends on leakage, additionality, permanence, and verification, none of which are usually managed directly by the purchasing organization.

Reductions are categorized as either gross reductions (abatement relative to the 2018 baseline) or offset-covered residuals. The offset-covered residual represents the portion of gross emissions that operational levers cannot eliminate and that must therefore be offset to achieve net zero.

This distinction is consequential for interpreting F1’s statement of achieving absolute 50% reductions against the 2018 baseline before resorting to offsets. The 50% absolute reduction is a claim about the gross reductions, whereas reaching net zero depends on offsetting a residual whose credibility rests on the verification and effectiveness of the credits used. We therefore report gross and net emissions separately throughout and treat offset-covered net zero as conditional on offset quality rather than equivalent to physical decarbonization. To further elaborate, two scenarios that reach the same net figure by different mixes of abatement and offsetting are therefore not equivalent. Where the text refers to net zero being reached, this denotes the accounting outcome under stated offset assumptions and not the elimination of physical emissions.

Full equation derivations are provided in Supplementary Table 2.

Parameter uncertainty and sensitivity design

The model contains two kinds of uncertain quantities: policy levers (chosen by stakeholders) and structural parameters (fixed assumptions about physical or organizational relationships). We conducted a sensitivity analysis for all the levers. The structural parameters and their variable ranges are: SAF life cycle emission reduction: 50–85% (the Carbon Offsetting and Reduction Scheme for International Aviation (CORSIA) framework default is 80%, but actual reductions depend on feedstock, like waste-based, which can achieve up to 85%, whereas crop-based pathways may deliver only around 50%) (Prussi et al., 2021; International Civil Aviation Organization, 2025)

Rail and sea emission factors relative to air freight: 0–20% of the air emissions factor (we assume the central value at 10%, which may increase with older vessels or last-mile trucking).

VIP travel share of business travel: 10–25% (we assume the baseline to be 15%, as F1 does not publish this split).

Distance reduction as a result of calendar regionalization: 40–75% (literature (Struijk, 2024) notes that distance reduction of 62% is achieved at complete regionalization, but partial optimization or more radical consolidation shifts this window).

Offset effectiveness: 20–100%.

The policy levers in the dashboard include: SAF adoption (0–100%), rail/sea freight share (0–60%), remote work (0–100%), renewable electricity (0–100%), calendar size (18–26 races), regionalization (0–100%), and offset rate (0–100%).

We conducted two types of sensitivity analyses: a local one-at-a-time (OAT) analysis and a global Monte Carlo simulation. In the OAT analysis, we varied each structural parameter across its full range while holding all other parameters at their central values. We set the balanced mid-ambition scenarios at: SAF = 45%, rail/sea share = 5%, remote work = 35%, renewable electricity = 90%, calendar size = 22 races, and regionalization = 35%. In the Monte Carlo analysis, we generated 10,000 samples by drawing all policy levers and structural parameters simultaneously from uniform distributions spanning their respective ranges.

Model calibration

The model was calibrated against F1’s published data (Formula One World Championship Limited, 2019, 2024, 2025). Three historical presets were used to confirm internal consistency: the 2018 baseline, the 2023 reporting year, and the 2024 reporting year. The 2018 baseline preset reproduces the restated total of 228,793 tCO₂e, disaggregated as follows: logistics at 67,994 tCO₂e, business travel at 79,425 tCO₂e, event operations at 22,813 tCO₂e (of which 1,602 tCO₂e is attributable to power unit combustion), and facilities at 58,562 tCO₂e. The 2023 and 2024 presets reproduce the officially reported totals of 182,801 tCO₂e and 168,720 tCO₂e, respectively. As F1 does not publish year-specific lever values, these presets represent reverse-calibrated lever approximations and should not be interpreted as verified operational parameters. We assess the model out of sample against F1’s 2026 Carbon Update (Formula One World Championship Limited, 2026)published after this study was submitted, which reports 2025 emissions by category (Supplementary Section 6.1). The model reproduces F1’s reported total to within 4% (143,099 against 148,805 tCO 2 e) with the dominant categories in the same rank order, with logistics at 50,889 tCO 2 e against 48,400 tCO 2 e, and business travel at 52,180 tCO 2 e against 57,934 tCO 2 e. However, for facilities and event operations, the model overestimates facilities emissions at 27,016 tCO 2 e against 20,923 tCO 2 e and underestimates event operations at 13,014 tCO 2 e against 21,548 tCO 2 e. This is due to the renewable energy adoption lever that is shared between the two categories; although in practice, facilities might adopt renewable electricity faster than event sites, and one shared adoption rate cannot fit both, F1 does not report adoption rates by category; we retained the single lever rather than add a parameter we cannot calibrate. This supports the model’s exploratory nature by providing an accurate aggregate footprint and capturing dominant emission sources for decision support.

Static sensitivity analysis

To assess the relative influence of each policy lever, we tested each parameter at its maximum value while holding all others fixed at their 2024 levels, corresponding to a gross emissions baseline of 168,720 tCO₂e. Supplementary Figure 1 reflects the gross emissions reduction from each independent intervention in such a simulation.

The sensitivity analysis shows SAF adoption exerts the greatest single-lever influence, reducing gross emissions by 58.32% from the 2024 level at 100% adoption. This result reflects the fact that SAF simultaneously governs logistics and business travel emissions, which together accounted for 71.95% of 2024 emissions. Calendar regionalization at 100%, which implies a 62% reduction in overall travel distance, produces the second-largest reduction, eliminating 69,796 tCO₂e (41.36% of the 2024 emissions). Remote work, renewable electricity, calendar size reduction, and modal shift to rail and sea follow in descending order, reducing gross emissions by 30.70, 25.52, 20.33, and 18.81% of the 2024 emissions, respectively, at maximum lever deployment.

When all levers are maximized simultaneously, 100% SAF adoption, 100% calendar regionalization, 100% remote work, 100% renewable electricity, an 18-race calendar, and 60% of logistics shifted to rail and sea, gross emissions are not fully eliminated. A residual of 3,877 tCO₂e remains, confirming that carbon offsets are required to reach a net-zero scenario, even under the most aggressive feasible combination of operational interventions. No combination of the modeled operational levers eliminates gross emissions within the tested ranges, so any net-zero outcome in this model is necessarily offset-mediated. The magnitude of residual depends on factors such as the quality of the adopted lever. Even at 100% SAF adoption, emission reductions depend on the fuel’s life cycle factor; under low (crop-based) life cycle assumptions or constrained SAF supply, residual emissions might rise (Supplementary Figure 8).

Fifty percent reduction threshold

Formula 1 stated that it will prioritize a minimum 50% reduction in absolute emissions against the 2018 baseline before resorting to offsetting (Camm, 2025). We therefore simulated a scenario in which different lever combinations achieve this threshold of 114,396 tCO 2 e or below without any offset credits.

A balanced, realistic configuration specifying SAF adoption at 45%, rail/sea share of 5%, remote work rate of 35%, renewable electricity at 90%, a calendar size of 22 races, and calendar regionalization of 35% yields a gross emission of 71,560 tCO 2 e, a 68.72% reduction. These lever values follow the feasibility criteria anchored to F1’s 2024 progress and extended within the available limits. SAF is held to a partial blend because its supply remains constrained even though the technology is available. The intermodal shift, calendar reduction, and regionalization are kept conservative because they are bound by existing freight and race hosting. A less aggressive scenario with SAF at 40%, rail/sea share of 0%, remote work rate of 20%, renewable electricity of 70%, calendar size of 22 races, and calendar regionalization of 25% yields gross emissions of 104,847 tCO 2 e, a 54.17% reduction. The 50% threshold is thus achievable without extreme assumptions. Achieving reductions beyond the 50% threshold becomes substantially more demanding, requiring every lever to be adopted near-maximum values, and would still rely on offsets of uncertain quality.

Time dynamic trajectories

The dynamic tab of the dashboard maps emissions over time from 2024 to 2030 under four adoption speed scenarios, as shown in Figure 2. The causal loop diagram, as shown in Figure 1, illustrates the feedback structure driving these trajectories.

The 2018 baseline scenario—characterized by minimal adoption speed, 21 races, and no offsets—results in a negligible decline in net emissions from 169,368 tCO 2 e (2024) to 151,990 tCO 2 e (2030). The 2023 scenario reaches 93,269 tCO 2 e, demonstrating a substantial decline and achieving the 50% reduction mark by 2030. The 2024 scenario reduces the net emissions to 57,708 tCO 2 e, but the slope decreases gradually after 2028 as the feedback loop erodes adoption pressure. The net zero scenario reaches zero but requires aggressive adoption rates, including SAF adoption at 50%/yr., remote work at 20%/yr., renewable electricity at 30%/yr., offset ramp at 17%/yr., 20 races, and 100% regionalization. The gap-closing dynamic is visible in all curves, showing steep early declines that flatten as stocks approach their ceilings and P(t) weakens. The adoption speeds represent scenario assumptions. The renewable energy speed is classified as technically feasible because Formula 1 already reported a shift to renewable energy sources to power Formula 1 and F1 Team sites. The intermodal shift is consistent with F1’s stated air-to-sea freight strategy. The SAF ramp is aspirational and is based on F1’s aim to expand their SAF framework (Formula One World Championship Limited, 2024). The offset ramp reflects a financial decision with no natural adoption rate, and the calendar and regionalization settings are scenario choices, with regionalization bounded by empirical estimates and the selected speeds being optimistic in the near term.

Uncertainty analysis

We conducted two types of uncertainty analyses, with the local (OAT) analysis revealing which structural parameters exert the greatest influence on a given scenario. Supplementary Figure 2 shows the results for the balanced realistic configuration discussed in Section 3.3. Each bar spans the gross emissions range that results when one structural parameter is swept across its plausible bounds while all others remain at their central values. SAF life cycle reduction (50–85%) produces the widest range, from roughly 69,000 to nearly 85,000 tCO 2 e, as it is the parameter governing the two largest emission categories carrying the most weight. Maximum regionalization (40–75%) spans the second highest, whereas the VIP travel share and rail/sea share emission factors matter less, showing shorter bars. The max regionalization denotes the distance reduction due to calendar regionalization. Literature (Struijk, 2024) noted that for a 23-race calendar, at 100% calendar regionalization, the maximum distance reduction would be 62%. We extended this range modestly, under the assumption that shorter race calendars affect the overall distance, thus stretching it to 75% maximum distance reduction (Supplementary Figure 2).

The second sensitivity analysis was a global Monte Carlo simulation sampling all 12 parameters simultaneously (n = 10,000) over the 2024 baseline values. The levers were bounded by F1’s 2024 range (SAF adoption at 5%, Renewable electricity at 48%, rail and sea freight at 5%, and remote work at 19%) since capacity already in place is not realistically reversed within the target window, whereas the discretionary levers of regionalization and offsetting were allowed to span the full 0–100% range. Figure 3 ranks each parameter by its variance contribution (Pearson r 2) to net emissions. The carbon offset rate has the highest r 2 value of 0.273, followed by SAF adoption (0.170), renewable electricity (0.130), offset effectiveness (0.121), and regionalization (0.118). Together, these five parameters account for nearly 81% of the outcome variance. The calendar size is moderate at 0.032, whereas remote work, SAF life cycle reduction, rail/sea share, VIP share, the rail/sea emission factor, and the maximum regionalization distance reduction each fall below 0.02. Figure 3 demonstrates that outcome uncertainty is highly influenced by questions of policy choices rather than emission factors alone. We validated the ranking with a global sensitivity analysis using Sobol indices under independence and Shapley (SHAP) values under correlation, in both cases the offset rate and SAF remain the dominant levers (Supplementary Section 5.3).

Supplementary Figure 3 shows the resulting distribution of net emissions. The distribution concentrates most outcomes between 20,000 and 80,000 tCO 2 e, with the 5th–95th percentile range of approximately 15,000 to 95,000 tCO 2 e. The 50% reduction target of 114,396 tCO 2 e falls well within reach, as nearly all sampled outcomes fall below it. The net-zero scenario, on the other hand, lies at the far left of the distribution, and few sampled outcomes reach net zero without simultaneously pushing every lever to its maximum. The median outcome leaves a substantial residual gap, confirming that moderate or even moderately aggressive parameter combinations are unlikely to meet the 2030 target. This independent sampling maps the spread of outcomes across the decision space. Supplementary Figure 5 represents the independent Monte Carlo overlapped by a second run in which the levers are correlated through a latent ambition factor representing a dependency that advances many levers together, widening the distribution and lowering the share of combinations meeting the 50% target from about 99% to about 88%. As the true degree of coordination in F1’s decisions is unobservable, we report both cases in Supplementary Section 5.3.

The results of this study offer several findings that are directly relevant to ongoing debates regarding the credibility of Formula 1’s net-zero commitment. Most notably, the static sensitivity analysis demonstrates that achieving a 50% absolute emissions reduction, F1’s stated precondition for offsetting, is feasible under moderate, operationally plausible assumptions. This finding partially corroborates (Chiles, 2025), who argued that meaningful progress is within reach, and stands in contrast to the more skeptical assessments of (Sturm et al., 2025), who characterized F1’s decarbonization trajectory as largely symbolic. Our estimates suggest that the 50% threshold is not symbolic in principle; rather, the credibility question shifts to whether the required adoption rates can be politically and operationally sustained.

The dominance of sustainable aviation fuel (SAF) in the sensitivity analysis, accounting for approximately 59% of gross emissions reductions at full adoption, underscores the degree to which F1’s net-zero pathway is structurally dependent on a single technology whose life cycle emissions reduction factor remains contested. The 50–85% uncertainty range applied here follows CORSIA guidance (Prussi et al., 2021; International Civil Aviation Organization, 2025). A threshold analysis in Supplementary Section 5.5 confirms that SAF’s dominance is conditional. SAF is the most effective lever only when its life cycle reduction exceeds 57%. At the central 80% value, it remains the most effective single abatement lever and, after the offset rate, the largest contributor to outcome variance. This finding carries a direct policy implication: F1 and its logistics partners should prioritize waste-based SAF feedstocks with verified life cycle assessments rather than defaulting to crop-based alternatives that may deliver reductions at the lower bound of that range.

The emissions-pressure feedback loop in the dynamic simulation captures a behavioral mechanism that is absent from prior descriptive analyses of F1 sustainability (Öztopcu, 2023; Al-rubaye and Naimi, 2025): the tendency for decarbonization momentum to erode as gross emissions fall and the urgency of further action weakens. The visible flattening of all trajectories after 2027–2028 suggests that F1 faces a structural risk of stalling in the final years before its 2030 deadline, even under otherwise aggressive adoption scenarios. This reinforces the argument made by (Chiles, 2025) that F1 must institutionalize sustained pressure mechanisms, such as binding contractual commitments with logistics providers and mandatory SAF blending requirements, rather than relying on voluntary momentum alone.

The Monte Carlo simulation further reveals that outcome uncertainty is primarily a function of policy lever positions rather than structural emission factors. The five highest-variance parameters, offset rate, SAF adoption, renewable electricity, offset effectiveness, and calendar regionalization, are all within F1’s sphere of influence or negotiation. This finding directly counters the narrative that measurement uncertainty or exogenous factors explain the risk of missing the 2030 target. The residual gap is, by this model’s accounting, a governance problem rather than a data problem.

Several limitations warrant acknowledgment. First, the VIP-to-staff travel split (15%/85%) is an internal estimate, as F1 does not publish this disaggregation; the sensitivity analysis shows that this parameter contributes minimally to outcome variance, but the assumption should be revisited if granular data become available. Second, the model does not endogenize cost. It does not account for the financial feasibility of high SAF adoption or of aggressive offset purchasing at the volumes implied by the net-zero scenario, which may constrain real-world implementation. In particular, certified low-carbon SAF trades at a premium to conventional jet fuel and remains supply-constrained, while high-integrity offsets raise questions regarding permanence and additionality (Bergero et al., 2023; Romm et al., 2025). Third, the forward Euler integration scheme (dt = 0.25 years) introduces numerical approximation error in the dynamic simulation. While the step size is sufficient for indicative trajectory analysis, higher-resolution integration would improve precision for near-deadline projections. Fourth, this model accounts for F1 as a reporting entity and does not separately model FIA emissions, which Al-Rubaye and Naimi (2025) identify as a distinct source of year-on-year reporting volatility. Finally, the expansion effect of Cadillac’s entry in 2026 is not dynamically modeled; its emissions are implicitly absorbed into the 2024 baseline trajectory rather than treated as an additive shock.

This study presents a systems analysis of Formula 1’s emissions trajectory calibrated to official sustainability reporting data, offering a tool for stakeholders to evaluate the feasibility of specific intervention combinations under realistic assumptions. We find that F1’s 50% absolute reduction target is achievable without extreme assumptions, but that achieving net zero by 2030 requires simultaneously pushing every major policy lever toward its operational ceiling while maintaining offset rates sufficient to cover a residual of approximately 3,900 tCO₂e. The dominant levers are SAF adoption and calendar regionalization; however, the quality and life cycle performance of SAF remain the greatest sources of outcome uncertainty. The emissions-pressure feedback mechanism identified here suggests that F1 faces a structural deceleration risk in the final years of its target window, making early and rapid adoption critical. Taken together, the findings indicate that F1’s 2030 commitment is physically achievable but institutionally demanding, and that the central risk is not a lack of available interventions but a failure to sustain implementation pressure through to completion.