Diplomacy & Defense Think Tank News

Export bans are frequently used as trade policy instruments to stabilise domestic prices, but they often generate unintended consequences. This study examines the effects of Indonesia's palm oil export ban, introduced in April 2022 and lifted in May 2022, on the performance of the global agricultural sector and the stock markets of palm oil–producing countries. Drawing on a conceptual framework, we develop hypotheses regarding how stock indices with different industrial compositions respond to such trade policy interventions. Using an event study methodology, we analyse daily stock market data from palm oil–producing countries as well as a global agriculture-specific MSCI index. The analysis encompasses both nationally diversified stock indices and a global sector–specific index. The results reveal statistically significant negative cumulative abnormal returns for the global agricultural index following the introduction of the export ban, whereas national cross-industry indices show statistically insignificant reactions. Following the lifting of the ban, the global agricultural index experienced statistically significant positive cumulative abnormal returns, whereas national indices exhibited mixed and statistically insignificant responses. These findings are consistent with the conceptual framework, suggesting that the policy primarily affected the global agricultural sector while leaving diversified national stock markets largely unaffected. Importantly, the effects associated with the introduction and subsequent lifting of the ban did not fully offset each other, resulting in an overall negative net effect on the global agricultural index.

Export bans are frequently used as trade policy instruments to stabilise domestic prices, but they often generate unintended consequences. This study examines the effects of Indonesia's palm oil export ban, introduced in April 2022 and lifted in May 2022, on the performance of the global agricultural sector and the stock markets of palm oil–producing countries. Drawing on a conceptual framework, we develop hypotheses regarding how stock indices with different industrial compositions respond to such trade policy interventions. Using an event study methodology, we analyse daily stock market data from palm oil–producing countries as well as a global agriculture-specific MSCI index. The analysis encompasses both nationally diversified stock indices and a global sector–specific index. The results reveal statistically significant negative cumulative abnormal returns for the global agricultural index following the introduction of the export ban, whereas national cross-industry indices show statistically insignificant reactions. Following the lifting of the ban, the global agricultural index experienced statistically significant positive cumulative abnormal returns, whereas national indices exhibited mixed and statistically insignificant responses. These findings are consistent with the conceptual framework, suggesting that the policy primarily affected the global agricultural sector while leaving diversified national stock markets largely unaffected. Importantly, the effects associated with the introduction and subsequent lifting of the ban did not fully offset each other, resulting in an overall negative net effect on the global agricultural index.

Ein Team der Abteilung Energie, Verkehr, Umwelt hat bei der europäischen Jahrestagung der International Association for Energy Economics (IAEE) zwischen dem 6. und 9. September 2026 in München aktuelle Forschung vorgestellt. Dabei hat Lukas Barner, Doktorand im DIW Graduate Center seit 2024, den ...

The Hamburg Sustainability Conference (HSC) 2026 convened around 1,600 participants from 112 countries at a moment of profound disruption and redistribution of power. The HSC confirmed sustainability as the organizing framework linking peace and stability, economic
resilience and competitiveness, planetary boundaries and international cooperation. The rules-based international order is under pressure, fiscal space is shrinking and conflicts are multiplying, while climate change, biodiversity loss and resource depletion are compounding
these pressures on economies and societies already stretched thin. Yet the world has never possessed greater technological capabilities, financial wealth and scientific knowledge. What matters now is how we use these resources to deliver shared prosperity and resilience.
The discussions at HSC 2026 can be summarized along three narrative arcs. The first asks why cooperation remains essential in a more multipolar world. Legitimacy depends on changing the terms of North–South relations, broadening representation and building a cooperative multipolar order in which a wider group of actors can shape and defend common rules. The launch of the South–North Commission on Development offers a structured process for rethinking international cooperation toward fairer partnerships and a  sustainability Agenda beyond 2030. The second narrative arc examines what needs to change. Sustainability transformation pays
off economically, but only where investment can flow, risks are shared and governance is predictable. Nature and human capital are systematically undervalued. The bottleneck is the institutional and financing structures needed to unlock them at scale. The third narrative arc addresses how cooperation can deliver. It shows that partnerships accelerate sustainable development when they move from one-off initiatives to systems that mobilize finance, create markets and build local capabilities. Scaling Capital for Sustainable Development (SCALED), Innovative Capital Mobilization in Africa (ICAMA), critical-mineral and hydrogen partnerships, and urban examples from Mombasa and eThekwini show how tangible forms of cooperation can turn commitments into sustainable investment, deeper value chains and locally owned action. This shift is especially visible in Africa, where partnerships are increasingly framed around opportunity, value creation and agency. The road ahead runs through the Triple COP+ year, the G20 under UK chairmanship, and the SDG Summit in September 2027. HSC 2026 demonstrated that navigating the new - disruptive - normal is possible. The task now is to prove it at scale. […]

The Hamburg Sustainability Conference (HSC) 2026 convened around 1,600 participants from 112 countries at a moment of profound disruption and redistribution of power. The HSC confirmed sustainability as the organizing framework linking peace and stability, economic
resilience and competitiveness, planetary boundaries and international cooperation. The rules-based international order is under pressure, fiscal space is shrinking and conflicts are multiplying, while climate change, biodiversity loss and resource depletion are compounding
these pressures on economies and societies already stretched thin. Yet the world has never possessed greater technological capabilities, financial wealth and scientific knowledge. What matters now is how we use these resources to deliver shared prosperity and resilience.
The discussions at HSC 2026 can be summarized along three narrative arcs. The first asks why cooperation remains essential in a more multipolar world. Legitimacy depends on changing the terms of North–South relations, broadening representation and building a cooperative multipolar order in which a wider group of actors can shape and defend common rules. The launch of the South–North Commission on Development offers a structured process for rethinking international cooperation toward fairer partnerships and a  sustainability Agenda beyond 2030. The second narrative arc examines what needs to change. Sustainability transformation pays
off economically, but only where investment can flow, risks are shared and governance is predictable. Nature and human capital are systematically undervalued. The bottleneck is the institutional and financing structures needed to unlock them at scale. The third narrative arc addresses how cooperation can deliver. It shows that partnerships accelerate sustainable development when they move from one-off initiatives to systems that mobilize finance, create markets and build local capabilities. Scaling Capital for Sustainable Development (SCALED), Innovative Capital Mobilization in Africa (ICAMA), critical-mineral and hydrogen partnerships, and urban examples from Mombasa and eThekwini show how tangible forms of cooperation can turn commitments into sustainable investment, deeper value chains and locally owned action. This shift is especially visible in Africa, where partnerships are increasingly framed around opportunity, value creation and agency. The road ahead runs through the Triple COP+ year, the G20 under UK chairmanship, and the SDG Summit in September 2027. HSC 2026 demonstrated that navigating the new - disruptive - normal is possible. The task now is to prove it at scale. […]

The Hamburg Sustainability Conference (HSC) 2026 convened around 1,600 participants from 112 countries at a moment of profound disruption and redistribution of power. The HSC confirmed sustainability as the organizing framework linking peace and stability, economic
resilience and competitiveness, planetary boundaries and international cooperation. The rules-based international order is under pressure, fiscal space is shrinking and conflicts are multiplying, while climate change, biodiversity loss and resource depletion are compounding
these pressures on economies and societies already stretched thin. Yet the world has never possessed greater technological capabilities, financial wealth and scientific knowledge. What matters now is how we use these resources to deliver shared prosperity and resilience.
The discussions at HSC 2026 can be summarized along three narrative arcs. The first asks why cooperation remains essential in a more multipolar world. Legitimacy depends on changing the terms of North–South relations, broadening representation and building a cooperative multipolar order in which a wider group of actors can shape and defend common rules. The launch of the South–North Commission on Development offers a structured process for rethinking international cooperation toward fairer partnerships and a  sustainability Agenda beyond 2030. The second narrative arc examines what needs to change. Sustainability transformation pays
off economically, but only where investment can flow, risks are shared and governance is predictable. Nature and human capital are systematically undervalued. The bottleneck is the institutional and financing structures needed to unlock them at scale. The third narrative arc addresses how cooperation can deliver. It shows that partnerships accelerate sustainable development when they move from one-off initiatives to systems that mobilize finance, create markets and build local capabilities. Scaling Capital for Sustainable Development (SCALED), Innovative Capital Mobilization in Africa (ICAMA), critical-mineral and hydrogen partnerships, and urban examples from Mombasa and eThekwini show how tangible forms of cooperation can turn commitments into sustainable investment, deeper value chains and locally owned action. This shift is especially visible in Africa, where partnerships are increasingly framed around opportunity, value creation and agency. The road ahead runs through the Triple COP+ year, the G20 under UK chairmanship, and the SDG Summit in September 2027. HSC 2026 demonstrated that navigating the new - disruptive - normal is possible. The task now is to prove it at scale. […]

• Turkey aspires to emerge as an AI-enabled middle power. Its main advantage is the ability to add selected AI tools to areas where it is already strong, especially defence manufacturing, unmanned systems and regional security cooperation.

• Turkey’s position in AI is uneven. It is stronger in applications and systems integration than in advanced chips, computing infrastructure, cloud capacity or specialised talent. For Ankara, strategic autonomy is therefore less about becoming self-sufficient and more about managing dependence.

• Defence offers Turkey the clearest way to turn AI into geopolitical influence. As military systems become more software-driven, defence exports can create longer relationships through training, software updates, data management and cybersecurity.

• At the same time, AI creates a new dependency problem. More advanced systems require more energy, better chips, greater computing capacity and continued access to trusted NATO and European technology networks. Strategic autonomy may therefore depend as much on reliable partnerships as on domestic production.

• Turkey’s biggest limits may be political and institutional rather than technical. Brain drain weakens the talent base, economic instability makes long-term investment harder, and an unpredictable foreign policy can reduce the trust needed for sensitive cooperation in AI, defence and data.

Read here in pdf the Policy paper by Ahmet Erdi Öztürk, Non-Resident Senior Scholar, Turkey Programme.

ARTIFICIAL INTELLIGENCE IS BECOMING part of international power, not simply another technology sector. Much of the debate still focuses on which states can build frontier models, produce advanced chips or command huge computing resources. That makes sense when discussing the United States and China, while the European Union has tried to shape the field through regulation and standards (Landwehr-Matlé, Oppermann and Lambach, 2026). But this framework is less useful for middle powers. Turkey does not need to reproduce the whole AI ecosystem of the leading powers. The more relevant question is whether it can acquire and use selected technologies in areas where it already has an advantage. For this reason, Turkey is better understood as a possible AI-enabled middle power: a state that can turn selective technological capabilities into military, economic and diplomatic leverage.

This argument fits closely with the debate on strategic autonomy. In Turkish foreign policy, the term usually describes Ankara’s efforts to gain more room for manoeuvre while remaining inside Western security and economic structures. Öniş and Kutlay (2021) situate this strategy in the context of a more post-Western international order, where Turkey tries to diversify its partnerships and reduce excessive dependence on traditional centres of power. AI makes this harder. It relies on advanced chips, computing power, cloud services, electricity, data, software and highly skilled people. For Turkey, technological autonomy cannot realistically mean self-sufficiency; it means retaining access to critical technologies, limiting vulnerability and preventing outside actors from using technological bottlenecks as political pressure.

Recent work on digital sovereignty points in the same direction. Weymouth (2025) argues that competition over AI infrastructure is encouraging the formation of techno-blocs, as states try to secure key resources and reduce dependence on foreign suppliers. Brown and Ozluk (2026) make a similar point about Turkey. They describe its AI strategy as a search for sovereignty that still requires cooperation with larger regulatory and technological systems. The key point is simple: technological sovereignty is not a choice between dependence and independence. For a middle power, it is about controlling some critical areas while managing dependence in others.

The recent report by the National Intelligence Academy (Milli İstihbarat Akademisi, MİA) shows that this way of thinking is also emerging inside Turkey’s strategic community. The report does not reduce AI competition to software and algorithms; it also highlights chips, critical minerals, energy, data infrastructure and robotics (Milli İstihbarat Akademisi, 2026). This is useful, but the next question is how Turkey’s strengths and weaknesses are distributed across these areas. One way to see this is as an AI power stack. At the top of the supply chain are semiconductors, high-performance computing, energy and cloud infrastructure. In the middle are data, software and AI models. At the downstream end are sector-specific applications, systems integration and the ability to turn technology into practical results.

Turkey’s position is highly uneven. Its main weaknesses are upstream. It does not have a globally competitive advanced-chip industry, it does not control major cloud or GPU infrastructure, and it remains dependent on international supply chains. Talent is another problem. 

Seen this way, Turkey’s position is highly uneven. Its main weaknesses are upstream. It does not have a globally competitive advanced-chip industry, it does not control major cloud or GPU infrastructure, and it remains dependent on international supply chains. Talent is another problem. AI requires specialised researchers, data scientists and technical managers who can connect research with industrial and military use. Brown and Ozluk (2026) are therefore right to stress human capital and institutional capacity. These weaknesses matter even more as access to critical technologies becomes increasingly political.

Strategic technological power does not come only from producing the best algorithm. It also stems from connecting software to platforms, sensors, communications, doctrine and operational practice.

Turkey looks stronger further downstream. Over the past two decades it has built a substantial defence-industrial ecosystem linking state procurement, large defence firms, newer technology companies and export markets. AI can be added to this existing base in areas such as unmanned systems, intelligence processing, sensors, radar, electronic warfare and command-and-control. The key distinction is between invention and integration. Strategic technological power does not come only from producing the best algorithm. It also stems from connecting software to platforms, sensors, communications, doctrine and operational practice. Gormus (2025) shows this in the Turkish case, arguing that while AI-enabled military capabilities may support strategic autonomy, they can also create new interoperability problems within NATO.

This broader viewpoint also explains why Turkey should not be judged against the global technology leaders alone. Regional powers have very different strengths. Israel has a deeper research, cybersecurity and high-tech ecosystem. Saudi Arabia and the UAE have more capital, abundant energy and strong partnerships with major technology companies. Turkey’s advantage is different: it combines a sizeable defence industry, manufacturing capacity, operational experience with unmanned systems and a broad export network. Technological sophistication and strategic effect do not always move together. A state can remain dependent on advanced inputs while still producing major strategic effects, if it can successfully integrate imported and domestic components into usable systems.

This dynamic fits a wider debate on middle-power agency. Middle powers are often described as states that hedge between larger powers. Mishra (2026), however, uses the idea of positional power to show how they can benefit from rivalry and institutional competition. Viewed through this lens, Turkey’s technological position does depend in part on this kind of networked position: NATO membership, economic links with Europe, partnerships in the Gulf, ties with Asian suppliers and a widening defence-export geography. Now AI is poised to make these connections more valuable; unlike one-off hardware sales, modern military systems require ongoing relationships around software, training, data, cybersecurity, and interoperability.

An AI-enabled middle power can therefore be defined as a state that does not control the whole AI ecosystem, but has enough industrial and institutional capacity to turn selected technologies into strategic effects. For Turkey, three questions matter most. First, how exposed is it to external restrictions? Second, how well can it integrate AI into sectors where it already has advantages? Third, do these capabilities improve Ankara’s bargaining power with allies, rivals and customers?

More AI may strengthen Turkey’s defence and security autonomy, but it may also deepen dependence on foreign chips, computing infrastructure and specialised software. 

This leads to a central paradox. More AI may strengthen Turkey’s defence and security autonomy, but it may also deepen dependence on foreign chips, computing infrastructure and specialised software. Strategic autonomy in the AI era is therefore not about escaping interdependence; it is about organising it. The MİA report is right to focus on the material foundations of technological competition (Milli İstihbarat Akademisi, 2026). The middle-power perspective adds another point: these capabilities matter geopolitically only when they create leverage. The challenge facing Turkey is to reduce upstream vulnerability while using its downstream strengths to improve military capacity and foreign-policy influence.

From Drones to Algorithms: Defence, Regional Competition and Foreign-Policy Leverage

Turkey is not building military AI from scratch; it is adding new functions to an existing network of drones, sensors, electronic warfare, communications and command systems.

If Turkey is to become an AI-enabled middle power, defence is the most obvious space in which to test the idea. Because Turkey’s defence transformation began well before the current AI boom, AI is entering a sector that already has experience of localisation, procurement reform, battlefield use and exports. This matters. Turkey is not building military AI from scratch; it is adding new functions to an existing network of drones, sensors, electronic warfare, communications and command systems. The shift from heavy import dependence towards greater domestic production has already tied industrial policy more closely to foreign policy (Ciftci, 2023; Baysal, 2025). AI may deepen this link, because success increasingly depends on connecting sensors, data, software and weapons into faster and more effective systems.Turkey

Automation, autonomy and AI should not be treated as the same thing. An automated system follows predefined rules. An autonomous system can respond to its environment and carry out some tasks with less human input. An AI-enabled military system uses tools such as machine learning, computer vision or predictive analytics to process information and support decisions. This distinction matters because drones are often described as AI systems, when many are mainly automated or remotely controlled. The more important military uses of AI are likely to be in sensor fusion, intelligence analysis, target recognition, mission planning, electronic warfare, logistics and human-machine cooperation (Johnson, 2026; Omer, 2026). The main advantage may not be removing humans from the process, but helping them make sense of large amounts of battlefield information more quickly.

Turkey’s defence sector is already moving in this direction. Baykar receives most of the international attention, but the picture is much broader. TUSAŞ produces larger unmanned systems including ANKA, AKSUNGUR and ANKA III. ASELSAN provides much of the radar, electro-optics, communications, electronic warfare and command infrastructure needed for networked operations. HAVELSAN focuses heavily on software, simulation and command systems, while STM works on tactical UAVs, loitering munitions, maritime systems, swarms and AI-supported decision tools. STM, for example, has highlighted deep-learning image processing, sensor fusion and autonomous mission allocation, and in May 2026 demonstrated a 20-unit KARGU swarm with live ammunition during EFES-2026. None of this proves that Turkey has achieved military AI superiority. What it does show is that AI-related functions are spreading across several firms and operational areas, rather than being confined to one flagship project.

Turkey’s earlier drone experience helps explain why this matters. Turkish UAVs had effects well beyond their cost in Syria, Libya and the 2020 Nagorno-Karabakh war, although drones alone did not determine the outcomes of hose conflicts. Rossiter and Cannon (2022) show how Turkey learned and adapted after early operational use, while Hwang and Song (2022) link drones to Ankara’s ability to intervene at relatively low political and financial cost and then convert military visibility into diplomatic influence. More recent work also warns against technological determinism: drones matter because of how they interact with doctrine, other military capabilities and enemy weaknesses (Hutto and Rogers, 2026). The same caution applies to AI. Turkey’s next advantage is unlikely to come from a single autonomous platform; it will depend on whether AI improves the way intelligence, electronic warfare, targeting, communications and command systems work together.

These capabilities also matter for foreign policy. Defence exports can create influence that goes beyond battlefield performance. Boyle and Wolfe (2026) show that drone exporters use sales not only to achieve commercial reasons, but also to build political relationships and longer-term dependencies. Turkish deals may include training, joint production, technology transfer, end-use arrangements and continued software support. AI can make these links deeper: as systems become more software-dependent, suppliers and customers remain connected through updates, data management, cybersecurity, maintenance and interoperability. Defence diplomacy can therefore become less about selling a platform and more about providing an ecosystem. This does not guarantee political alignment, but it can make the relationship longer and more difficult to replace.

Turkey is also part of broader regional competition over military technology. Israel remains ahead in advanced research, cybersecurity, sensors and military-technological integration. Saudi Arabia and the UAE are following a different model, using capital, energy and partnerships with global firms to develop their AI and defence sectors. The UAE’s EDGE and Saudi Arabia’s SAMI are examples of this state-backed approach, although both still rely heavily on foreign technology and partnerships (Rózsa, 2025). Iran follows another path, one shaped by sanctions and pressure for self-reliance but constrained by limited resources. Atwood (2025) shows the gap between Iran’s official claims of AI sovereignty and the more fragmented reality of its computing infrastructure, data centres and digital labour. These differences make a simple regional ranking misleading. Israel’s model is research-intensive, the Gulf model capital-intensive, Iran’s sanction-driven, and Turkey’s more focused on integration and manufacturing.

Turkey’s potential advantage is therefore not that it has the region’s most advanced AI ecosystem; it is that it may be able to combine sufficient technological sophistication with defence products that are scalable, relatively affordable and operationally tested. 

Turkey’s potential advantage is therefore not that it has the region’s most advanced AI ecosystem; it is that it may be able to combine sufficient technological sophistication with defence products that are scalable, relatively affordable and operationally tested. This matters because strategic utility is not the same as technological quality. Turkish drones expanded partly because they filled a market gap between advanced but tightly restricted Western systems and cheaper alternatives with different capabilities. AI-enabled systems could follow a similar path if Turkey can improve intelligence processing, autonomy and electronic warfare without making its products too expensive for existing customers. The wider growth of the defence sector has already become part of Ankara’s more assertive foreign-policy toolkit (Ciftci, 2023; Baysal, 2025).

The regional effects will be broad but uneven. In the Gulf, AI-related defence cooperation could support Turkey’s rapprochement with Saudi Arabia and the UAE, but Turkish firms will also face strong local competitors. In North Africa, defence ties can reinforce Ankara’s political and security presence, especially where affordable unmanned and electronic systems are attractive. The South Caucasus offers the clearest example of defence technology creating political visibility, while Central Asia connects defence cooperation with wider Turkic and connectivity agendas. The Black Sea is more constrained by Russia, NATO and the war in Ukraine. The Middle East is also highly competitive, with Israel, Iran, Gulf states and outside powers all active there. Turkey’s technological reach should therefore not be confused with regional dominance. Its value lies in giving Ankara more ways to enter security relationships, strengthen partnership and remain part of regional strategic calculations.

The broader point is that AI may change the political economy of Turkish defence influence before it changes the regional military balance. Turkey’s drone experience showed that a middle power could turn a specialised technology into operational reputation, exports and diplomatic access (Rossiter and Cannon, 2022; Soyaltin-Colella and Demiryol, 2023). AI could deepen this leverage because it ties hardware to software, data and continuing technical support. But the opportunity is conditional: it depends on reliable integration, affordability, battlefield performance and continued access to the advanced technologies on which these systems rely. AI does not replace Turkey’s defence-industrial model; it makes that model both more valuable and more vulnerable.

The New Dependency Paradox

Turkey’s growing use of AI in defence creates a problem that is easy to miss in discussions about localisation. Because, while more domestic production may reduce dependence on imported weapons, it may also increase dependence on the infrastructure needed to run advanced systems. AI requires much more than software. It needs advanced chips, computing power, cloud infrastructure, data centres, reliable electricity, cooling, skilled people and secure networks. Weymouth (2025) argues that control over these concentrated supply chains is becoming a source of geopolitical power. For Turkey, the question is therefore not only how much AI it can put into defence systems, but whether it can retain access to the infrastructure those systems need during periods of geopolitical tension.

Energy is central to this problem because AI is highly electricity-intensive. Indeed, the International Energy Agency estimates that global data-centre electricity use could rise from about 485 TWh in 2025 to roughly 950 TWh by 2030, with AI-focused facilities growing even faster (IEA, 2026). While Turkey does not need to copy the huge computing centres being built in the United States or the Gulf, the price, availability and reliability of electricity will nonetheless increasingly shape its technological capacity. Consequently, the advent of AI inextricably links energy security with technology policy. Expanding domestic computing requires generation capacity, stable grids, storage, cooling and some protection from imported-energy shocks. Turkey’s investment in renewables and nuclear power may therefore have a direct technological value. The relationship also works both ways: AI can help with forecasting, grid management and infrastructure efficiency (IEA, 2025).

A further issue is that Turkey’s energy position is itself uneven. The country has expanded its renewable capacity and is adding nuclear power, but it still depends heavily on imported energy. 

A further issue is that Turkey’s energy position is itself uneven. The country has expanded its renewable capacity and is adding nuclear power, but it still depends heavily on imported energy. This matters because AI infrastructure needs a stable and predictable supply of electricity. A sudden rise in energy prices, pressure on gas imports or problems with the grid would not only affect households and industry; it could also limit the expansion of data centres and high-performance computing. In that sense, energy dependence can become a technological dependence as well. The more Turkey wants to build domestic AI capacity, the more important it becomes to reduce exposure to external energy shocks and strengthen the resilience of the electricity system.

There is also a regional dimension. Gulf states have an obvious advantage in their ability to combine cheap energy, capital and large-scale investment in data centres and computing infrastructure. Turkey is unlikely to be able to compete with them on scale alone. Its advantage may instead come from combining a relatively diversified energy mix with industrial capacity, proximity to European markets and existing links to NATO and EU technological networks. This could make Turkey more attractive as a regional location for selected AI- and defence-related infrastructure, particularly if it can improve grid reliability and regulatory predictability. Energy policy therefore becomes part of a wider strategic question: not simply how much electricity Turkey can produce, but whether it can provide the stable, secure and internationally connected infrastructure needed for an AI-intensive economy.

This creates a second layer of dependence. Reducing reliance on imported defence platforms can simply move the vulnerability further upstream, to chips and computing capacity in particular. Advanced semiconductors are particularly sensitive because production is concentrated in a small number of countries and depends on highly specialised equipment and supply chains. Rone (2025) shows how semiconductor policy in Europe has become closely linked to national security. Export controls make the issue even more important. Weymouth (2025) argues that controls on chips and digital infrastructure are helping to reorganise technological interdependence into competing blocs. For Turkey, which has tried to maintain room for manoeuvre between Western institutions and a wider set of partners, this could make technological hedging more difficult. In other words, greater technological sophistication may actually make strategic autonomy harder to sustain.

This is why Turkey’s NATO membership and its links to European technology networks can also be assets. NATO’s revised AI strategy places interoperability, good-quality data, testing, responsible use and an AI-ready workforce at the centre of allied planning. It also recognises the energy demands of AI and the need to protect systems from adversarial interference (NATO, 2024). Participation can give Turkey access to standards, testing networks, DIANA centres, industry links and shared practices that would be expensive to build alone. The paradox is that deeper participation in allied technology networks may increase Turkey’s practical autonomy. In this sense, autonomy can come through institutional embeddedness rather than separation.

That embeddedness also creates obligations, however. NATO interoperability is no longer only about whether systems can communicate; it increasingly involves shared standards for testing, data quality, explainability, accountability and human control. NATO’s Principles of Responsible Use stress lawfulness, accountability, traceability, reliability, governability and bias mitigation (NATO, 2024). EU regulation adds a further layer by linking access to markets and technological networks to risk management and technical standards (Kilian, Jäck and Ebel, 2025). Turkey could therefore face difficulties if military AI develops more rapidly than its systems for verification, accountability and regulation. Governance is not simply a normative issue here; it can directly affect interoperability, procurement and access to trusted technology networks.

Turkey has long struggled to retain its most highly skilled professionals. Research on brain drain shows how institutional and labour-market conditions shape the decision of skilled workers to leave or return.

Human capital is another possible bottleneck. Defence AI requires people who combine machine-learning skills with engineering, cybersecurity and an understanding of military systems. Turkey has long struggled to retain its most highly skilled professionals. Research on brain drain shows how institutional and labour-market conditions shape the decision of skilled workers to leave or return (Elveren and Toksöz, 2019). This matters even more in AI, where specialists can find highly paid jobs in the United States, Europe and increasingly the Gulf. Investment in hardware will not be enough if Turkey cannot retain the people needed to develop, maintain and improve these systems.

AI also creates new security risks that cannot be measured simply by looking at industrial capacity. Machine-learning systems can be manipulated through bad data or adversarial inputs. More networked military platforms create more cyber vulnerabilities. Faster decision-support systems may also reduce the time available for human judgement in a crisis. NATO therefore treats adversarial AI, system protection and testing as core aspects of military AI readiness (NATO, 2024). There is also a domestic issue: technologies developed for defence and security can expand state surveillance, making questions of accountability and proportionality increasingly important.

Turkey’s AI position is therefore shaped by a dependency paradox: the more successfully Ankara adds AI to domestic defence and security systems, the more important resources it does not fully control become: energy, chips, computing capacity, software, skilled labour and international standards. This does not make technological autonomy impossible, but it changes what autonomy should mean. The realistic goal is resilience, in the form more diverse supply chains, reliable energy, secure computing, access to allied innovation networks and credible governance. Turkey’s AI power will depend as much on how it manages these dependencies as on the technologies it develops.

The Political Limits of Turkey’s AI Ambitions

Turkey already has several of the ingredients it needs to become an AI-enabled middle power: a large defence industry, experience with unmanned systems, manufacturing capacity, NATO membership and a growing network of defence partners. But technology alone will not decide whether these advantages produce lasting geopolitical influence. The biggest obstacles may lie in the political, economic and institutional environment surrounding innovation.

The first problem is human capital. Advanced AI depends on strong universities, scientific freedom, international research links and the ability to attract and keep skilled people. Turkey’s authoritarian turn makes this more difficult. Pressure on universities, weaker institutional autonomy and doubts about meritocracy can encourage researchers, engineers and other highly educated professionals to leave (Aydin, 2021; Elveren and Toksöz, 2019). Recent research also finds that brain drain has had a measurable negative effect on technological progress in Turkey, where it has weakened both the domestic talent pool and the R&D capacity (Mike et al., 2025). This creates a basic contradiction: a state seeking technological sovereignty may be losing one of the resources it needs most.

Though Turkey’s defence industry shows that concentrated state support can produce strong results in selected areas, AI requires a much wider and more expensive ecosystem. With macroeconomic instability and competing public spending priorities, sustaining investment across the whole chain will be difficult. 

A second problem is economic capacity. Competitive AI requires long-term spending on computing infrastructure, data centres, energy, research and specialised staff. Though Turkey’s defence industry shows that concentrated state support can produce strong results in selected areas, AI requires a much wider and more expensive ecosystem. With macroeconomic instability and competing public spending priorities, sustaining investment across the whole chain will be difficult. Turkey may therefore develop pockets of advanced capability, especially in defence, without building the wider infrastructure needed to support them over time.

Foreign policy creates a third constraint. Turkey is geopolitically important because of its location, NATO membership and military capabilities. That makes Ankara difficult for both allies and rivals to ignore. But being indispensable is not the same as being trusted. Transactional diplomacy, changing alignments and repeated disputes with Western partners can create short-term bargaining power while also making Turkey look unpredictable. This matters in an AI environment that is increasingly built around trusted networks, sensitive data, interoperability and controlled access to advanced technology. Positional power is useful only if partners believe that cooperation with Ankara will remain sufficiently predictable.

Turkey’s AI future therefore depends on much more than successful defence platforms. The key question is whether technological ambition can be matched by institutional quality, economic resilience and international trust. 

Turkey’s AI future therefore depends on much more than successful defence platforms. The key question is whether technological ambition can be matched by institutional quality, economic resilience and international trust. AI could strengthen strategic autonomy and increase Turkey’s middle-power influence, but authoritarianism, brain drain, economic weakness and declining trust could also place a clear ceiling on those ambitions. The real competition is not only over technology; it is also over the political and institutional capacity to sustain it.

References

Atwood, B. (2025) ‘Artificial intelligence in Iran: National narratives and material realities’, Iranian Studies, 58(4), pp. 680–697.

Aydin, B. (2021) ‘Politically motivated precarization of academic and journalistic lives under authoritarian neoliberalism: The case of Turkey’, Globalizations, 19(5), pp. 677–695.

Baysal, B. (2025) ‘Evaluating the advances and challenges in Turkey’s defence industry: A comparative analysis’, Southeast European and Black Sea Studies, 25(1), pp. 31–52.

Boyle, M.J. and Wolfe, W. (2026) ‘Diplomatic drones? How states exploit drone exports for leverage’, European Journal of International Security, 11(2).

Brown, R.D. and Ozluk, D. (2026) ‘Regulatory autonomy and AI sovereignty: Türkiye at a crossroad’, Asian Journal of Law and Society, First View, pp. 1–27.

Ciftci, S. (2023) ‘Military might: A domestic economy explanation of Turkish foreign policy’, Turkish Studies, 24(5), pp. 764–787.

Elveren, A.Y. and Toksöz, G. (2019) ‘Hidden gender dimensions of the brain drain: The case of Turkey’, New Perspectives on Turkey, 60, pp. 33–59.

Gormus, E. (2025) ‘NATO’s artificial intelligence strategy and interoperability challenges: The case of Turkey’, Journal of Balkan and Near Eastern Studies, 27(4), pp. 513–529.

Hutto, J.W. and Rogers, J.P. (2026) ‘The drone revolution: Towards a synthesis in the drone debate’, European Journal of International Security, 11(2), pp. 145–165.

Hwang, W.-J. and Song, S.-H. (2022) ‘The extension of Turkish influence and the use of drones’, Comparative Strategy, 41(5), pp. 439–458.

IEA (2025) Energy and AI. Paris: International Energy Agency.

IEA (2026) Key Questions on Energy and AI. Paris: International Energy Agency.

Johnson, J. (2026) Artificial Intelligence and the Future of Warfare: The USA, China, and Strategic Stability. Manchester: Manchester University Press.

Kilian, R., Jäck, L. and Ebel, D. (2025) ‘European AI standards – Technical standardisation and implementation challenges under the EU AI Act’, European Journal of Risk Regulation.

Landwehr-Matlé, J., Oppermann, K. and Lambach, D. (2026) ‘Great power competition for global leadership in artificial intelligence (AI): Reconstructing AI narratives of the United States, China, and the European Union’, Review of International Studies, First View, pp. 1–19.

Mike, F., Akyıldız, İ.E., Doğanlar, M. and Kızılkaya, O. (2025) ‘The innovation exodus: How Türkiye’s brain drain challenges technological progress’, Economics of Innovation and New Technology.

Milli İstihbarat Akademisi (2026) Dünya Siyasetinde Teknopolitik Düzen Tartışmaları ve Türkiye. Ankara: Milli İstihbarat Akademisi.

Mishra, A. (2026) ‘Positional power: Middle power strategies in an era of institutional rivalry’, International Affairs, 102(4), pp. 1287–1315.

NATO (2024) Summary of NATO’s Revised Artificial Intelligence (AI) Strategy. Brussels: North Atlantic Treaty Organization.

Omer, N. (2026) ‘The AI kill chain in the defense sector: Strategic, ethical, and governance implications in the Middle East’, in Hacker, P. (ed.) Oxford Intersections: AI in Society. Oxford: Oxford University Press.

Öniş, Z. and Kutlay, M. (2021) ‘Turkish foreign policy in a post-Western order: Strategic autonomy or new forms of dependence?’, International Affairs, 97(4), pp. 1085–1104.

Rone, J. (2025) ‘The economy–security nexus: Risk, strategic autonomy and the regulation of the semiconductor supply chain’, European Journal of Risk Regulation, 16(1), pp. 279–293.

Rossiter, A. and Cannon, B.J. (2022) ‘Turkey’s rise as a drone power: Trial by fire’, Defense & Security Analysis, 38(2), pp. 210–229.

Rózsa, E.N. (2025) ‘Great power competition in defence technology – What can the Asia-Pacific offer for the Gulf?’, Asian Journal of Middle Eastern and Islamic Studies, 19(3), pp. 285–302.

Soyaltin-Colella, D. and Demiryol, T. (2023) ‘Unusual middle power activism and regime survival: Turkey’s drone warfare and its regime-boosting effects’, Third World Quarterly, 44(4), pp. 724–743.

Weymouth, S. (2025) ‘Digital disintegration: Techno-blocs and strategic sovereignty in the AI era’, International Organization, 79(S1), pp. 57–S70.

Der Rat der Europäischen Zentralbank (EZB) hat heute beschlossen, den Leitzins auf nun 2,5 Prozent zu erhöhen. Dazu eine Einschätzung von Marcel Fratzscher, Präsident des Deutschen Instituts für Wirtschaftsforschung (DIW Berlin):

Die EZB hat mit der Zinserhöhung einen notwendigen Schritt getan, um die Inflationserwartungen zu stabilisieren und ihre Glaubwürdigkeit zu schützen. Allerdings dürfte die Zinserhöhung nichts Substanzielles an der derzeit hohen Inflation ändern – auch nicht über das kommende Jahr. Denn die Inflation ist fast ausschließlich durch den Energiepreisschock infolge des Krieges im Nahen Osten verursacht worden. Gegen solch einen externen Schock hat die EZB nicht wirklich etwas in der Hand.

Die Inflationserwartungen im Euroraum sind zwar noch gut verankert, aber die EZB verschafft sich mit dem heutigen Schritt eine bessere Absicherung gegen mögliche Zweitrundeneffekte durch Unternehmen und Gewerkschaften. Sie sendet ein Signal an alle wirtschaftlichen Akteure, dass sie ihr Ziel der Preisstabilität ernst nimmt und dafür auch gewillt ist, die Wirtschaft im Euroraum zu bremsen.

Die EZB hält sich völlig zu Recht alle Optionen für die Zukunft offen, denn die Unsicherheit ist enorm hoch. Eine erneute Eskalation des Konflikts im Nahen Osten könnte die Energiepreise und damit die Inflation noch deutlich weiter erhöhen.

Allerdings ist Vorsicht geboten, den Bogen nicht zu überspannen. Denn die langfristigen Zinsen sind vor allem wegen der Zweifel von Wirtschaft und Märkten an der Handlungsfähigkeit der Politik – nicht nur, aber vor allem in den USA – deutlich gestiegen. Das reduziert den Druck auf die EZB, die Zinsen noch deutlich weiter zu erhöhen.

Spyros Blavoukos, Head, EU Institutions & Policies Programme, ELIAMEP; Head of the ‘Ariane Condellis’ European Programme;  Professor, Athens University of Economics & Business;  Apostolos Samaras, Research Fellow, European Programme ‘Ariane Condellis’ and Panos Politis Lamprou, Research Fellow, Defence Hub, ELIAMEP examine the recent debate over the future of the EU’s Common Foreign and Security Policy (CFSP), focusing on two developments that have brought questions about the Union’s foreign-policy architecture back to the forefront: proposals to reform the European External Action Service (EEAS) and renewed discussions on the use of qualified majority voting in CFSP decision-making. It places these developments in their broader institutional and historical context, exploring what is genuinely new in the current debate and what they could mean for the way the EU organises and conducts its external action.

Read the ELIAMEP Explainer here.

Which development will be at the centre of the European agenda in the coming months? The war in Ukraine, geopolitical shifts in the Eastern Mediterranean and the wider region, as well as the remarkable rise of the far right across Europe, are contributing to an increasingly complex environment. Which of these developments will have the greatest geopolitical significance and have the strongest impact on Europe’s future? ELIAMEP Analysts share their assessments. (the post in available only in Greek)

 

Climate-induced planned relocation is becoming an unavoidable policy issue for some highly exposed communities. Although relocation should remain a last resort after other realistic options for adaptation in place have been assessed, ensuring that the money is available when needed to support relocation is critical. The challenge is that funding is insufficient, drawn from diverse sources and often arrives too late and in fragmented forms: for example, one project for housing, another for infrastructure, and little predictable support for consultation, land negotiations, safeguards, livelihoods, host communities or long-term maintenance. The financing problem is therefore not only the scale of the costs, but the difficulty of organising money over time and across institutions and safeguards. This can leave governments reacting after crises rather than planning before risks become unmanageable.
Sovereign trust funds offer one practical way to address this gap. They do not create finance by themselves, but they can provide a country-owned platform for receiving, sequencing and reporting domestic revenue, bilateral support, multilateral development bank (MDB) finance, climate funds, disaster risk finance, and loss and damage resources. They present an alternative to predominantly loan- and grant-based financing for planned relocation. Fiji’s Climate Relocation of
Communities Trust Fund shows both the promise and limits of this approach: It provides a legal and institutional basis for relocation finance within government and links international funding to procedures, but it still requires capitalisation, administrative capacity and long-term technical support. The model presents an opportunity to support sovereign, rights-based financing systems rather than relying solely on donor-funded, often piecemeal relocation projects. It can also make funding available at the appropriate time and without undue time pressure.
Key policy messages:
• Finance planned relocation before a crisis, especially assessment, consent, land, safeguards and project preparation.
• Treat relocation as a long-term investment in the collective good, rather than solely as a construction or emergency response measure.
• Support sovereign trust funds where they are legally mandated at the national level, budget-linked, transparent and capitalised.
• Use finance to uphold quality and rights, including community-led processes, support for host communities and support for long-term livelihoods.

Climate-induced planned relocation is becoming an unavoidable policy issue for some highly exposed communities. Although relocation should remain a last resort after other realistic options for adaptation in place have been assessed, ensuring that the money is available when needed to support relocation is critical. The challenge is that funding is insufficient, drawn from diverse sources and often arrives too late and in fragmented forms: for example, one project for housing, another for infrastructure, and little predictable support for consultation, land negotiations, safeguards, livelihoods, host communities or long-term maintenance. The financing problem is therefore not only the scale of the costs, but the difficulty of organising money over time and across institutions and safeguards. This can leave governments reacting after crises rather than planning before risks become unmanageable.
Sovereign trust funds offer one practical way to address this gap. They do not create finance by themselves, but they can provide a country-owned platform for receiving, sequencing and reporting domestic revenue, bilateral support, multilateral development bank (MDB) finance, climate funds, disaster risk finance, and loss and damage resources. They present an alternative to predominantly loan- and grant-based financing for planned relocation. Fiji’s Climate Relocation of
Communities Trust Fund shows both the promise and limits of this approach: It provides a legal and institutional basis for relocation finance within government and links international funding to procedures, but it still requires capitalisation, administrative capacity and long-term technical support. The model presents an opportunity to support sovereign, rights-based financing systems rather than relying solely on donor-funded, often piecemeal relocation projects. It can also make funding available at the appropriate time and without undue time pressure.
Key policy messages:
• Finance planned relocation before a crisis, especially assessment, consent, land, safeguards and project preparation.
• Treat relocation as a long-term investment in the collective good, rather than solely as a construction or emergency response measure.
• Support sovereign trust funds where they are legally mandated at the national level, budget-linked, transparent and capitalised.
• Use finance to uphold quality and rights, including community-led processes, support for host communities and support for long-term livelihoods.

Climate-induced planned relocation is becoming an unavoidable policy issue for some highly exposed communities. Although relocation should remain a last resort after other realistic options for adaptation in place have been assessed, ensuring that the money is available when needed to support relocation is critical. The challenge is that funding is insufficient, drawn from diverse sources and often arrives too late and in fragmented forms: for example, one project for housing, another for infrastructure, and little predictable support for consultation, land negotiations, safeguards, livelihoods, host communities or long-term maintenance. The financing problem is therefore not only the scale of the costs, but the difficulty of organising money over time and across institutions and safeguards. This can leave governments reacting after crises rather than planning before risks become unmanageable.
Sovereign trust funds offer one practical way to address this gap. They do not create finance by themselves, but they can provide a country-owned platform for receiving, sequencing and reporting domestic revenue, bilateral support, multilateral development bank (MDB) finance, climate funds, disaster risk finance, and loss and damage resources. They present an alternative to predominantly loan- and grant-based financing for planned relocation. Fiji’s Climate Relocation of
Communities Trust Fund shows both the promise and limits of this approach: It provides a legal and institutional basis for relocation finance within government and links international funding to procedures, but it still requires capitalisation, administrative capacity and long-term technical support. The model presents an opportunity to support sovereign, rights-based financing systems rather than relying solely on donor-funded, often piecemeal relocation projects. It can also make funding available at the appropriate time and without undue time pressure.
Key policy messages:
• Finance planned relocation before a crisis, especially assessment, consent, land, safeguards and project preparation.
• Treat relocation as a long-term investment in the collective good, rather than solely as a construction or emergency response measure.
• Support sovereign trust funds where they are legally mandated at the national level, budget-linked, transparent and capitalised.
• Use finance to uphold quality and rights, including community-led processes, support for host communities and support for long-term livelihoods.

The AU–EU partnership should embrace the plurality within both unions, make interests transparent, and enable pragmatic cooperation between coalitions of willing countries, helping to bring proclaimed ambitions closer to actual practice, explain Benedikt Erforth and Lena Gutheil.

The AU–EU partnership should embrace the plurality within both unions, make interests transparent, and enable pragmatic cooperation between coalitions of willing countries, helping to bring proclaimed ambitions closer to actual practice, explain Benedikt Erforth and Lena Gutheil.

The AU–EU partnership should embrace the plurality within both unions, make interests transparent, and enable pragmatic cooperation between coalitions of willing countries, helping to bring proclaimed ambitions closer to actual practice, explain Benedikt Erforth and Lena Gutheil.

This paper examines how leadership gender configurations shape digitalisation and innovation in Egyptian manufacturing. Using the 2020/21 Egyptian Industrial Firm Behavior Survey (EIFBS) covering 2,338 firms, we construct a four-category gender variable (female owners and male managers (FOMM), male owners and female managers (MOFM), female owners and female managers (FOFM) and all-male baseline) and analyse how these owner–manager gender mixes relate to the adoption of digital technologies (DT) and to innovation outputs, and how these relationships vary by firm size and DT use. Our analysis suggests that firms with male owners and female managers (MOFM) are most likely to adopt DT across specifications. The cross-sectional data suggests that mixed-gender firms are associated with a higher probability of spending on R&D, but not with higher innovation outputs in firms with female owners (i.e. FOMM and FOFM) – pointing toward an innovation conversion gap in those firms. Heterogeneity results show that the MOFM adoption premium of DT and a negative association between FOMM and innovation are strongest in small firms. DT use moderates gender gaps: female-owned firms not using DT are significantly less likely than all-male firms to generate innovation outputs, but this penalty disappears when female-owned firms use DT. A combined size–sector analysis suggests that the divergence between MOFM and FOMM/FOFM is driven mainly by small food manufacturers, with average marginal effects elsewhere broadly comparable. The results highlight leadership composition as a correlate of technology adoption and the role of DT in converting innovation inputs into outputs.

This paper examines how leadership gender configurations shape digitalisation and innovation in Egyptian manufacturing. Using the 2020/21 Egyptian Industrial Firm Behavior Survey (EIFBS) covering 2,338 firms, we construct a four-category gender variable (female owners and male managers (FOMM), male owners and female managers (MOFM), female owners and female managers (FOFM) and all-male baseline) and analyse how these owner–manager gender mixes relate to the adoption of digital technologies (DT) and to innovation outputs, and how these relationships vary by firm size and DT use. Our analysis suggests that firms with male owners and female managers (MOFM) are most likely to adopt DT across specifications. The cross-sectional data suggests that mixed-gender firms are associated with a higher probability of spending on R&D, but not with higher innovation outputs in firms with female owners (i.e. FOMM and FOFM) – pointing toward an innovation conversion gap in those firms. Heterogeneity results show that the MOFM adoption premium of DT and a negative association between FOMM and innovation are strongest in small firms. DT use moderates gender gaps: female-owned firms not using DT are significantly less likely than all-male firms to generate innovation outputs, but this penalty disappears when female-owned firms use DT. A combined size–sector analysis suggests that the divergence between MOFM and FOMM/FOFM is driven mainly by small food manufacturers, with average marginal effects elsewhere broadly comparable. The results highlight leadership composition as a correlate of technology adoption and the role of DT in converting innovation inputs into outputs.

This paper examines how leadership gender configurations shape digitalisation and innovation in Egyptian manufacturing. Using the 2020/21 Egyptian Industrial Firm Behavior Survey (EIFBS) covering 2,338 firms, we construct a four-category gender variable (female owners and male managers (FOMM), male owners and female managers (MOFM), female owners and female managers (FOFM) and all-male baseline) and analyse how these owner–manager gender mixes relate to the adoption of digital technologies (DT) and to innovation outputs, and how these relationships vary by firm size and DT use. Our analysis suggests that firms with male owners and female managers (MOFM) are most likely to adopt DT across specifications. The cross-sectional data suggests that mixed-gender firms are associated with a higher probability of spending on R&D, but not with higher innovation outputs in firms with female owners (i.e. FOMM and FOFM) – pointing toward an innovation conversion gap in those firms. Heterogeneity results show that the MOFM adoption premium of DT and a negative association between FOMM and innovation are strongest in small firms. DT use moderates gender gaps: female-owned firms not using DT are significantly less likely than all-male firms to generate innovation outputs, but this penalty disappears when female-owned firms use DT. A combined size–sector analysis suggests that the divergence between MOFM and FOMM/FOFM is driven mainly by small food manufacturers, with average marginal effects elsewhere broadly comparable. The results highlight leadership composition as a correlate of technology adoption and the role of DT in converting innovation inputs into outputs.

This paper investigates the determinants and dynamics of labour demand and specifically informal labour in Egypt’s manufacturing sector, using nationally representative firm-level data from the 2020/21 Egyptian Industrial Firm Behavior Survey. Applying ordinary least squares and fractional logit models, we analyse total employment, the share of informal labour, and its average annual change over the firm life cycle. Three key findings emerge. First, employment is positively associated with capital, exporting, technology adoption, innovation, industrial zones, worker training, and managerial education, and negatively associated with sole proprietorships, wages, and total factor productivity. Second, informal employment is more common among private sector firms, sole proprietorships, and firms using more part-time workers, and less prevalent among firms adopting technology or led by more educated managers. Third, changes in informality over time are modest: most formal firms exhibit no change in the share of informal workers. Notably, formal firms that did not initially employ informal labour tend to increase their informal share, while firms that formalised continue to rely heavily on informal employment. Together, these findings underscore the persistence of informality and limited transitions towards full formalisation within Egypt’s formal manufacturing sector.