Notably, as the globe shifts toward a knowledge economy, the convergence of digital transformation and sustainable development has become a significant focus of scholarly inquiry. It broadly denotes the integration of digital technology into business, government, and society. Digital technologies are a crucial facilitator of the UN Sustainable Development Goals (World Bank, 2018), especially in the domains of education, health, agriculture, and energy. Research indicates that digital transformation serves as a significant motivator for sustainability (Sarasini et al., 2024) contended that emerging digital technologies present opportunities for circular economy models, capable of enhancing impact and efficiency by utilizing existing resources more intelligently, thereby reducing ecological footprints. Similarly, Yousaf et al. (2021) indicate that digital finance can enhance financial inclusion and wellbeing in emerging economies. Digital transformation is seen as a crucial element of sustainable economic diversification under Saudi Arabia’s Vision 2030, particularly within the context of the Gulf area. The UAE is making substantial advancements, shown by the Smart Dubai effort, which aims to utilize digital technology to facilitate sustainability objectives in urban planning and public service provision (Khanna et al., 2021). Moreover, digital transformation includes intelligent infrastructure, such as smart grids and smart cities, as a component of sustainability. According to Albitar et al. (2020), smart city efforts facilitate reduced CO₂ emissions, optimized utility usage, and enhanced urban resilience. However, the implementation of digital technologies has its own challenges, including data security, digital inequality, and the environmental impacts of the necessary ICT infrastructure (Lozano et al., 2016).

Nearly all energy types share a commonality: they facilitate sustainable growth. Energy consumption and sustainable development are closely interconnected, since energy serves as both a catalyst for economic advancement and a contributor to environmental deterioration, particularly when derived from fossil fuels. Emissions associated with energy from these sources account for approximately 73% of total global greenhouse gas emissions (IEA, 2022). A substantial body of literature explores the energy-sustainable development link. The research conducted by Apergis et al. (2020) revealed that renewable energy usage significantly impacts economic growth and environmental preservation (Omer, 2008; Dincer and Rosen, 1999) demonstrated a strong correlation between energy consumption and economic growth in emerging nations, while simultaneously confirming that over-reliance on non-renewable energy undermines environmental sustainability. Energy consumption in Saudi Arabia and the United Arab Emirates has predominantly favored fossil fuels. Recent policy modifications aim to diversify the energy portfolio. In 2021, Vision 2030 initiated the Saudi Green Initiative, aiming to derive 50% of the nation’s power from renewable sources by 2030. The UAE’s Energy Strategy 2050 similarly targets a comparable energy mix, consisting of 44% renewable energy (Zhe et al., 2024; Sarasini et al., 2024). Moreover, energy efficiency measures are integral to sustainable growth. According to Kaygusuz (2012), energy savings of 20–30% may be realized in industries, transportation, and buildings through enhancements in efficiency while maintaining equivalent functioning. Energy auditing and the implementation of sophisticated metering infrastructure have shown to be successful strategies for promoting energy saving and sustainability (Energy and Iea, 2020). Nonetheless, the rebound effect, where energy and carbon benefits from efficiency are diminished due to heightened consumption (higher throughput), remains a concern. Harmonizing energy consumption with sustainable development necessitates an equilibrium among energy accessibility, cost-effectiveness, and ecological preservation.

For decades, foreign direct investment has been a significant driver of sustainable development, particularly in developing and emerging nations. FDI is a vital source of finance, facilitates technological transfer and interchange, and is crucial in capacity building, which is essential for achieving the SDGs (Voica et al., 2015). As stated by the UN Conference on Trade and Development, FDI serves as a primary driver of sustainable economic growth globally, generating employment, enhancing productivity, and facilitating infrastructure development (Yousaf et al., 2021). The link between FDI and sustainable development is not unidirectional. FDI can either adversely affect or enhance the environment, contingent upon the attributes of the investment and the host nation. Foreign Direct Investment in pollution-intensive sectors will markedly exacerbate environmental pollution unless stringent environmental regulations are implemented (Samour et al., 2022). Foreign Direct Investment has been a fundamental component of economic diversification initiatives in the Gulf Cooperation Council area. As previously mentioned, Saudi Arabia’s Vision 2030 and the Emirati economic strategies focus on sustainable foreign direct investment in renewable energy, technology, and tourism. According to Imran et al. (2024), foreign direct investment in green technology has increased at an annual pace of 18% since the implementation of Vision 2030. The converse aspect of the award pertains to the quality of foreign direct investment and its origins. According to FDI Projection 03 (Samour et al., 2022; Sarfraz et al., 2022), the origin of FDI may influence their effect on sustainable development, with investments from nations adhering to rigorous environmental regulations favorably impacting sustainability. Conversely, foreign direct investment from nations with lax regulations may result in the emergence of a “pollution haven.’ Numerous empirical studies, such those by Abedi and Moeenian (2021) and Al Shammre and Alshahrani (2024), demonstrate that FDI exerts a dual influence on CO₂ emissions: it fosters economic growth while also contributing to environmental pollution, unless accompanied by environmental safeguards. Consequently, host nations have the difficulty of formulating policies that maximize developmental advantages from foreign direct investment while minimizing ecological impact (Deryag and Khalifa, 2024). The research indicates that sustainable development may be affected, either favorably or adversely, by the three primary factors studied in this paper: digital transformation, energy consumption, and foreign direct investment. Future research should pursue integrated frameworks that delineate the synergies and trade-offs among these factors, especially in resource-abundant nations like KSA and the UAE. The study comprises many components: the third section addresses the gap and novelty, the fourth section outlines the theoretical framework and methods, the fifth section presents the empirical analysis, and the final section concludes the study. Presented here are pertinent research enquiries:

The Kingdom of Saudi Arabia and the United Arab Emirates achieved rapid economic advancement mostly via the extraction of natural resources and international investments. However, blossoming expansion results in sustainability challenges, particularly around carbon emission control, diversity of energy sources, and equitable development. While various studies have individually investigated the roles and effects of FDI, energy consumption, and digital transformation, to our knowledge, there is a paucity of research examining the interactive synergy of FDI, energy consumption, and digital transformation on sustainable development, especially in oil-dependent economies undergoing structural reform. To our knowledge, no prior research provides a thorough assessment that incorporates all of these predominantly separate primary variables into a unified model. The potential for this is promising in mitigating energy inefficiencies and facilitating intelligent solutions for environmental management; nevertheless, the impact of digital transformation, significantly undervalued in the Gulf context, remains unexamined. Secondly, the majority of research fail to compare several Gulf nations, despite their frequently analogous economic frameworks and shared policy goals (e.g., Saudi Vision 2030; UAE Net Zero 2050) (Mishra et al., 2024). Most current work relies on outdated data and too simplistic models, yielding conclusions with little accuracy and policy significance. This study addresses the identified gaps by analyzing updated data from 2000 to 2023 and employing advanced econometric approaches to examine the combined effects of digital transformation, foreign direct investment, and energy consumption on sustainable development in KSA and the UAE. The innovation is in employing cost–benefit analysis in a multi-variable, comparative framework with methodological rigor, revealing pathways for the long-term viability of these economies independent of oil reliance.

This research examines the relationship between sustainable development (see Appendix Tables 1, 2) and the trinity of digital transformation, energy consumption, and foreign direct investment in Saudi Arabia and the United Arab Emirates from 2000 to 2023. In this research, sustainable development is the dependent variable, while digital transformation, energy consumption, and foreign direct investment are the independent factors. WDI database, a publicly accessible and extensively utilized global development data repository (Al Shammre and Alshahrani, 2024). Expanding upon prior research on this topic, alternative studies have employed various methodological techniques. Researchers employed structural equation modelling and partial least squares methodologies utilizing World Bank data to assess the impact of FDI on economic growth within the Omani setting (Naseem et al., 2024). For the analytical process, we used Eviews-10. Additional studies examined the labor market by investigating the long-term effects of FDI on male and female unemployment, utilizing panel data from 1990 to 2018 and employing econometric methodologies such as FMOLS, variance decomposition, impulse response functions, and Granger causality tests to analyze both long-term and short-term dynamics (Majid, 2020). Co-integration tests and long-run estimate methodologies (FMOLS, quantile regression, wavelet coherency analysis, etc.) have been employed in the domain of sustainable development and green energy. This research employs the FMOLS approach, which robustly and reliably estimates the long-run equilibrium connection among variables, rendering the results highly useful for policy formation in the Gulf. The index variables and PCA techniques employed to produce the digital transformation index and sustainable development are shown in Appendix Tables 1, 2, respectively (Table 1).

The FMOLS is a strong panel data model with little room for mixed stationarity variables. Considerably, it integrates both long-term and short-term dynamics, hence, especially suitable for analyzing the effects of digital transformation, energy consumption, and foreign direct investment on sustainable development. FMOLS is used to measure the impact of changes in digitalization, energy consumption, and capital influx on the growth paths of sustainable growth in KSA and the UAE. The addition of variance decomposition into the model also provides more information regarding causality and explains the percentage of variation in prediction errors that can be contributed to each predictor over time. In all countries, particularly those seeking to diversify from oil, FDI is relevant to long-term sustainable development and its apparent short-term environmental and economic impacts. This model could then be used to narrow further the analysis of FDI driven industrialization by examining the extent to which energy consumption and/or digital transformation is dispersed, as well as the resultant worsening or mitigation of environmental pressure. Also, the data are usually checked for stationarity and long-run correlation using unit root tests and Co-integration tests respectively, which adds to the robustness of the results. These diagnostic techniques are essential for assessing the dynamics of digital transformation, energy consumption, and foreign direct investment in relation to modelled sustainable development. This study utilizes FMOLS to provide empirical data about the temporal interaction among these factors, while also supporting KSA Vision 2030 and UAE Sustainability objectives through pertinent policy implications. This work aims to replicate such couplings in a dynamic context to yield insights for strategic approaches that would attain an economically progressive and ecologically sustainable result.

The unit root of each variable was validated using the stationarity tests conducted. The stationarity tests identified in the existing literature are already approximatively. Several broad tests include the Augmented Dickey-Fuller test (Dickey and Fuller, 1981; Im et al., 2003; Shin and Schmidt, 1992). The unit root test posits non-stationarity as the null hypothesis, with failure to reject the unit root signifying non-stationarity as the alternative hypothesis; hence, the unit root equation is presented below: (3) y it = ρ y i ( t − 1 ) + δ X it + ε it

Where I = 1, N be a country-level panel; t = 1, T the period X it : capturing variables in the theoretical model,  is the coefficient of auto-regression and is the error term. Condition >1, represent a stationary behavior, whereas ρ i = 1, results in a unit root in y t . The study of Im et al. (2003), also provides variables with different serial correlation orders and applies the Augmented Dicky Fuller test, so-called, the average: (4) ε it = ∑ j = 1 ρi ρ ij ε it − j + μ it

If Equation 3 takes the place of Equation 4, we obtain Equation 6.

While the ADF can calculate by Equation 5: (5) y it = ρ i y it − 1 + ∑ j = 1 ρi ρ ij ε it − j + δ i X it + ε it

Where ρi represents the number of lags in the ADF regression. T-bar statistics identified by Im et al. (2003), the ADF average of individual statistics as shown below: (6) t NT ( ρi ) ¯ = 1 N ∑ i − 1 N t iT ( ρ i )

The statistical alternative “t” allows all individuals to test a null hypothesis for the existence of a unit root. t i ( ρ i ) represents the estimation of ADF; t denotes the number of observations, and N is the number of individuals.

The last step of an empirical analysis is the causality test that measures the intensity of the causal link between the variables in the model. Causation can be one-way (unidirectional), two-way (bidirectional), or non-existent. We employed the Dumitrescu Causality test by Dumitrescu and Hurlin (2012). What is a more effective tool for directional causality than the standard method (Dumitrescu and Hurlin, 2012). It includes the two main types of heterogeneity, i.e., the heterogeneity of the regression model and the heterogeneity of the causal effect. (7) y it = a i + ∑ j − 1 J ⋋ i j y i ( T − J ) + ∑ j − 1 J β i j X i ( T − J ) + e it

The measurable cross-sections are denoted by x, y, and ⋋ i j , β i j , in Equation 7 respectively. Each observable has its own set of auto-regressive parameters and regression coefficients, expected to change over time. Behavior of one may be attributable to another panel list. Wald statistics can be used to test the following hypothesis based on Equation 8: (8) W N . T HNC = N − 1 ∑ i = 1 N W i , t

For each analyzed cross-sectional unit the Wald statistics is denoted by W i , t .

The Variables Correlation Matrix is a table that illustrates the strength and direction of the linear relationship among numerous pairs of variables within a single dataset. The values in the matrix vary from −1 to +1, where +1 signifies perfect positive correlation, −1 signifies perfect negative correlation, and 0 denotes the absence of linear correlation. A correlation approaching 1 or −1 denotes a robust and direct link, whereas values around 0 suggest the lack of a linear relationship. It offers a method for identifying possible multicollinearity issues and assists in the subsequent model definition.

Table 3 correlation coefficients among SD, DT, EU, and FDI for the models of KSA and UAE A substantial positive connection exists between SD and DT (r = 0.737), indicating that digital technologies influence and advance sustainable development in alignment with national visions established for their respective areas, such as Saudi Vision 2030 and UAE Vision 2021. Likewise, SD has a substantial positive association with FDI (r = 0.626), suggesting that foreign capital supports sustainable initiatives, either through green investments or technological spillover. Despite a weak correlation between SD and EU (r = 0.486), it suggests that energy consumption, seen as a metric of economic and industrial advancement, might potentially facilitate sustainable development, provided that efficiency rules govern these initiatives. DT exhibits a poor correlation with EU (r = 0.101); digitalization has not yet directly influenced energy consumption (Alam et al., 2020), or this impact may manifest through efficiency improvements. In summary, the matrix elucidates the correlation among these factors and substantiates their collective impact on sustainability outcomes in the two nations.

Unit root tests, sometimes referred to as individual variable stationary tests, are a crucial first phase in time series analysis. The ADF, IPS, and KPSS tests assess whether a time series variable is stationary or requires differencing to achieve stationarity. Stationarity is crucial, as non-stationary data may lead to the identification of spurious correlations between variables (Rehman et al., 2021).

Table 4 presents stationarity diagnostics for the variables SD, DT, EU, and FDI, employing several unit root specifications, including ADF, DF-GLS, KPSS, and PP tests. The variables SD, DT, EU, and FDI are non-stationary at level I (0) but attain stationarity at the first difference level I (1), indicating that they are integrated of order one. In other words, they are not stationary at level; but, at first difference, ADF and PP tests produce extremely significant p-values (p < 0.01), so demonstrating stationarity at I (1). Similarly, the DF-GLS test corroborates this finding, indicating that all variables are stable following initial differencing (p < 0.01). Conversely, stationarity at the level is denied (with values over the crucial limit), although it is affirmed at the first difference (with values significantly below the threshold), corroborated by the KPSS test with SD (0.1519–0.0992) and EU (0.1529–0.65039) (Xie et al., 2024). These findings confirm that all series are integrated of the same order, hence validating the use of co-integration approaches from a long-run viewpoint and enhancing econometric modeling in the analysis of sustainable development drivers in KSA and the UAE.

Co-Integration Test The initial and most critical stage in time-series analysis is to ascertain the existence of a long-term link among the variables. The co-integration test (Drake, 1993; Johansen, 1992), is a prevalent approach employing maximum likelihood estimation to ascertain the existence of a co-integrated connection among various time series variables under analysis. The VAR model of vector k is characterized by an integration order of I (1) (Johansen, 1992).

The outcomes of the Johansen co-integration test for SD, DTE, Energy Use (EU), and FDI in KSA and UAE are presented in Table 5. The Trace and Max-Eigenvalue statistics for evaluating the maximum rank of co-integrating equations, for both, at the 5% critical value. Utilizing the Trace test, we see test statistics of 71.45 and 33.31, which surpass the essential values of 47.85 and 29.80, respectively. Consequently, the null hypotheses of “no co-integration” and “at most one” are rejected, with p-values of 0.0001 and 0.0189, respectively and 0.0001 and 0.0189, respectively. Similarly, the Max-Eigenvalue test corroborates these findings by rejecting the null hypotheses of “none” and “at most one,” both associated with considerably greater p-values (0.0015 and 0.0102, respectively), results that align with those of Zaheer and Rashid (2014) and Liu et al. (2022). The findings indicate the existence of a minimum of two long-term equilibria among the variables. It already carries the assumption that digital transformation is not only dynamically submerged in a set of interconnections with sustainable development, but also dynamically coincide in terms of adaptive optimization with sustainable digital development in the long run, reaching an optimal state for both. The results confirm the robustness of the VECM framework and highlight that coordinated policies promoting simultaneous digitalization, sustainable energy consumption, and foreign direct investment might contribute positively to the long-run sustainability performance in both regions.

There is a need for a method (FMOLS) which enables estimating long-term relationships inside Co-integrated variables restricting inessential statistical assumptions. This approach increases the reliability of the inference from the large amount of equation measurements because the errors between the equations are correlated. Employing this semi-parametric framework with FMOLS restricts the complexity of the estimation and leads to faster convergence rates, especially in the case of multi-co-integrated systems. FMOLS has many advantages; nevertheless, it is still limited to some constraints. When testing hypotheses in particular contexts, it may be conservative, making conclusions based on the results of the analysis less generalizable (Abid et al., 2022; Ur Rehman et al., 2023; Naseem et al., 2023).

The long-term relationship effects of DT, EU, and FDI on SD in KSA and UAE are assessed using the FMOLS estimate in steps 2 and 3, as demonstrated in Table 6. The FMOLS technique accounts for endogeneity and serial correlation, consistent long-term coefficient estimates. The results reveal that the coefficient for DT is positive and statistically significant (β = 0.4157, p = 0.0017), underscoring a substantial contribution of digital transformation to sustainable development in the region. FDI demonstrates a positive and substantial effect (β = 0.2659, p = 0.0068), indicating that the influx of foreign capital fosters sustainability, maybe by transferring cleaner technology and innovative practices (Kalam, 2020; Mohsin et al., 2024). The EU exhibits a negative but minimal significant coefficient (β = −0.0000626, p = 0.0558), albeit weakly indicating a potential unfavorable association between increased energy consumption and sustainable development. The model demonstrates a robust fit, evidenced by an R-squared value of 0.659, signifying that 66% of the variation in sustainable development is accounted for by the independent variables. These findings underscore the strategic significance of both digitalization and foreign investment in achieving sustainable development objectives in the Gulf region.

Table 7 illustrates the Dumitrescu-Hurlin causality test, a panel causality test designed to ascertain if one indicator may predict another within a panel dataset. Unlike classic causality tests, it considers the variety of dynamics within the interaction system across cross-sectional units, making it more suitable for panel data where correlations may differ among units. The DH test consolidates the results of individual cross-section causality, so offering a robust methodology for examining directional correlations in heterogeneous panels (Dumitrescu and Hurlin, 2012).

Results from the DH panel causality test reveal that nearly all variables are connected to each other with statistically significant bidirectional causal relationships, denoting persistent and strong dynamic interdependencies. In particular, information technology as a DT is found to Granger cause SD and vice-versa, pointing toward a mutual feedback relationship. Also, the EU Granger causes SD; on the other hand, SD Granger causes EU, thus suggesting a bounce feedback loop between environmental policy and energy behavior (Feedback loop interacts between environmental policy and energy behavior). Likewise, FDI directly causes SD but it is also affected by SD, revealing their mutual interdependence in sustainability. Thus, DTS could be Granger-caused by energy use, signifying the other side of the coin. On the other hand, FDI and DTS do not display a causal relationship, in either direction, with no significant causality between them (Khurshid et al., 2020; Imran et al., 2024). Moreover, FDI Granger causes EU, while the reverse does not occur, indicating that there is a one-way impact of capital inflow onto energy dynamics but not the other way around. The results show that the system remains tightly coupled with energy, digital and investment systems jointly contributing to sustainable development pathways (Salahuddin and Alam, 2016).

Constraints of the research about the impact of Foreign Direct Investment on sustainable economic growth in the Kingdom of Saudi Arabia. The primary constraint is its reliance solely on secondary data sources (e.g., the World Bank), which may constrain the depth and specificity of the research. Robustness may be enhanced if further research integrates more granular sectorial data or does micro-level analysis to examine the effects of FDI on specific industries. The current study overlooks significant factors influencing growth, such as institutional quality and domestic economic reforms, which may not have remained consistent or easily measurable during this period or among different nations, unlike economic growth. Integrating these aspects in the future may enhance the capacity to discern the ramifications of internal governance and external market signals for FDI. Consequently, the environmental aspect of sustainability, a crucial component of the KSA Vision 2030 policy, remains inadequately explored. Subsequent studies ought to investigate how foreign direct investment affects environmental outcomes, including case studies, qualitative analysis, or firm-level surveys. These methodologies would improve the comprehension of FDI and its impacts, especially regarding the transformation of spillover effects at critical periods and the transition of employment and economic restructuring from non-oil industries.