Diplomacy & Defense Think Tank News

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. […]

Over the previous two months, Turkish President Recep Tayyip Erdoğan hosted a NATO summit in Ankara and attended the Shanghai Cooperation Organisation summit in Bishkek. On his way back, Erdoğan presented the latter as part of Turkey’s “commitment to diversifying its economic, political, and strategic options”. This is not far removed from the usual analytical vocabulary. Ankara’s ability to straddle East and West is variously described as balancing, hedging or seeking strategic autonomy.

There are important differences between these concepts. Yet they often direct attention towards the same question: Where does Turkey stand in relation to the great powers? Turkish practices become evidence of the country’s position. The risk is that position then becomes a shortcut for assessing agency.

This can distort assessments in both directions. A focus on positioning can overestimate agency when the practice, purpose and effects of balancing are collapsed into one another. Maintaining relationships with competing powers may be intended to preserve flexibility and may, in turn, expand Ankara’s diplomatic repertoire. But such effects do not by themselves establish what Turkey can achieve through the options that these relationships create. At the same time, viewing Turkish foreign policy primarily through alignment can underestimate agency when practices that do not alter Turkey’s geopolitical position are overlooked, despite altering what Ankara can do.

The appeal of a balancing lens is understandable. It makes seemingly contradictory practices intelligible as part of a broader strategy. But explaining their strategic logic is not the same as assessing their effects. 

What becomes visible if we set aside balancing as the primary lens?

The test in the Black Sea 

Turkey’s response to Russia’s invasion of Ukraine has commonly been read as a balancing act. Ankara has not joined Western sanctions against Russia and has maintained political and economic ties with Moscow. At the same time, it has provided military support to Ukraine and applied the Montreux Convention to the passage of warships through the Turkish Straits, with consequences for Black Sea and non-Black Sea states alike. Read in this way, these policies appear to place Ankara somewhere between Russia and its Western allies. But a practice can matter without moving Turkey anywhere geopolitically.

More than four years later, Turkey continues to regulate naval access to the Straits. Together with Romania and Bulgaria, it operates the Black Sea Mine Countermeasures Task Group, whose mandate has expanded from mine clearance to the protection of critical undersea infrastructure following the Ankara NATO summit in July. Amid renewed attacks on commercial shipping, Turkish Foreign Minister Hakan Fidan has announced a proposal for the safe passage of grain, which Turkey is discussing with both Russia and Ukraine.

The basis of Turkish capacity is different in each case. Montreux gives Ankara authority over naval access within rules that also limit its discretion. Mine countermeasures are conducted through operational cooperation with Romania and Bulgaria. Relations with both Moscow and Kyiv create an opportunity for Ankara to broker an arrangement over commercial navigation, but not the conditions for agreement. Ankara’s agency is thus more specific, conditional and relational than the balancing vocabulary allows us to see.

What follows from where Turkey stands?

Turkey may well balance. The problem begins when balancing becomes the primary lens through which everything else is seen. It places great-power relations at the centre of the analysis, even where the capacity in question is produced elsewhere. Shifting the focus away from position raises a different set of questions: What can Ankara do? What enables it to do so? And which other actors does it depend on?

A practice that barely registers on an East–West axis may nevertheless alter what Turkey is able to do. For policy analysis, the relevant question is therefore not only where Ankara stands but also what it can affect, through which relationships and under what conditions.

The old establishment elites have opposed Namal’s arrest. They want to preserve a system where one of their own is not jailed for corruption.

• 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.

Un an après le lancement de l’opération Lance du Sud (Operation Southern Spear), la question de son efficacité se pose alors que les États-Unis veulent désormais l’étendre à tout le continent américain. Cette campagne militaire de surveillance, mise en place par l’administration de Trump, a pour but officiel de « détecter, perturber et dégrader les réseaux criminels transnationaux et les réseaux maritimes illicites ». Or, les accusations d’exécutions extrajudiciaires portées par de nombreuses ONG ainsi que les chiffres officiels qui annoncent 227 morts interroge sur son réel impact.

Quel bilan tirer de cette première année ? Au-delà des chiffres officiels, quelles conséquences réelles pour le narcotrafic et pour la région des Caraïbes ? Et alors que le dispositif s’étend désormais à toute l’Amérique latine, assiste-t-on à l’émergence d’une nouvelle doctrine militaire américaine dans la région ?

Nouvelle chronique de l’Amérique latine avec Christophe Ventura, directeur de recherche à l’IRIS et responsable du Programme Amérique latine/Caraïbe.

L’article Opération Lance du Sud dans la Caraïbe : où en sont les promesses de Trump un an après ? est apparu en premier sur IRIS.

With around half a year remaining before the next presidential election, France is in danger of becoming a stress test for the eurozone: a political crisis, a deteriorating fiscal position and growing market doubts about the country’s ability to finance its mounting debt burden are all converging. France faces a significant structural imbalance in its public finances. The budget deficit exceeds 5 per cent of GDP, while public expenditure stands at around 57 per cent of GDP. At the same time, public debt is rising steadily and could, without sufficient fiscal consolidation measures, exceed 130 per cent of GDP by 2030. The interest burden on public debt is rising rapidly and could rise to more than 120 billion euros, equivalent to 3.5 per cent of GDP by 2030. The central challenge, however, may not lie in the fiscal arithmetic of the budget, but in the political capacity to address it. On the one hand, the costs of social welfare expenditure are no longer sustainable; on the other, the majority of the population continues to demand more state intervention. The situation is compounded by the lack of a parliamentary majority and limited prospects of achieving one in the near future. The global environment is also weighing heavily: Tensions involving Iran are driving up fuel prices and inflation, while US fiscal concerns are putting upward pressure on bond yields.

An election campaign lacking a commitment to fiscal consolidation

With just over six months to go before the first round of the presidential election, the French economic debate is dominated by promises of additional spending, including commitments to lower the retirement age, as well as by diversionary issues such as constitutional referendums and by nationalist and protectionist rhetoric often accompanied by strongly Eurosceptic undertones. Jean-Luc Mélenchon’s call for a partial write-off of French debt held by the Banque de France, a member of the Eurosystem, is adding to market concerns about France. Marine Le Pen’s proposals, too – such as maintaining the low retirement age or escalating conflicts with the EU (over Schengen, migration policy and the EU budget) – could increase pressure on French government bonds and further weaken the economy through capital outflows. The high and rising share of foreign investors in French public debt, which stands at 56 per cent, makes the country even more vulnerable to shifts in market sentiment.

This raises the crucial question: What happens if market pressure on France persists and there is no parliamentary majority in favour of fiscal consolidation? The eurozone has no legal or political mechanisms in place in the event that the political system of a key member state becomes paralysed. The only credible safety net remains the European Central Bank’s (ECB) bond-purchasing programme, the Transmission Protection Instrument (TPI). The problem, however, is that the Eurosystem cannot finance France’s deficit or purchase French government bonds at will, solely in response to rising yields. Should broader intervention become necessary in the absence of a credible fiscal consolidation path from Paris, the ECB Governing Council would face a legally and politically difficult decision. The ECB would then find itself caught in a conflict of objectives between its mandate to protect the stability of the eurozone and its duty to prevent the single monetary policy from replacing the French government’s fiscal policy.

The real test for the eurozone

The previous eurozone crisis was triggered by smaller economies on its periphery, yet it nearly caused the collapse of the monetary union. A growing loss of market confidence in France would be a far more severe test for the eurozone. Bond purchases can mitigate acute liquidity and refinancing risks, but they merely shift the problem into a chronic structural fiscal crisis, which, if it escalates further, could eventually overwhelm the capabilities of monetary policy. The real test for the eurozone will be whether it can maintain stability in the bond market in Europe’s second-largest economy without relieving it of responsibility for its own fiscal policy. Should this balancing act fail, the Eurosystem could, in the short term, use its instruments to limit acute market distortions in the bond markets. However, such measures cannot permanently cushion the consequences of an unsustainable fiscal trajectory. The longer the fiscal problems remain unresolved, the more they will exacerbate political tensions over the limits of European solidarity and the future of the monetary union, thereby undermining confidence in the European institutions and the cohesion of the EU.

Le journal israélien Haaretz a révélé, ce 8 septembre, que Benyamin Netanyahou avait connaissance de l’avènement d’une opération majeure du Hamas sur le territoire israélien par le leader émirati Mohammed Ben Zayed, dix jours avant le 7 octobre.
Des révélations réfutées, qui mettent à mal Netanyahou dans le cadre des prochaines élections législatives d’octobre.

Au lendemain des sanctions annoncées par la France, la Grande-Bretagne et le Canada contre les produits issus des colonies israéliennes en Cisjordanie, c’est un véritable revers diplomatique et politique pour Benyamin Netanyahou.

Assiste-t-on, à l’aube des élections israéliennes, à la fin de l’ère Netanyahou ?

Mon analyse dans cette vidéo

L’article Clap de fin pour Netanyahou ? est apparu en premier sur IRIS.

En quelques mois, les robots humanoïdes sont devenus l’une des vitrines de l’accélération technologique et de la course mondiale en matière d’innovation, sur fond de poussée de l’intelligence artificielle d’un côté et de rivalités sino-états-uniennes de l’autre.

Plusieurs éléments concourent à relier le développement des humanoïdes avec les enjeux de l’agriculture et de l’alimentation, de la recherche à la production, de la transformation à la consommation, en passant par la logistique et la distribution. Tour d’horizon synthétique et prospectif d’un espace de réflexions appelé à s’élargir[1] et qui interrogera fortement l’interaction entre les machines et les humains dans les mondes de demain, notamment agricoles et alimentaires.

Tenir compte du contexte global

Pour comprendre en quoi l’irruption des humanoïdes dans l’univers agricole et alimentaire n’est pas à sous-estimer, prenons le soin de rappeler certaines dynamiques actuellement à l’œuvre, en se limitant à cinq d’entre elles qui nous paraissent déterminantes :

D’abord, un constat. Nourrir la planète constitue le premier secteur d’emploi mondial. Près de 1,4 milliard de personnes travaillent dans les systèmes agricoles et alimentaires, soit environ quatre actifs sur dix. L’agriculture primaire en mobilise à elle seule près de 900 millions, auxquels s’ajoutent les métiers de la pêche et de l’aquaculture, de la transformation, de la logistique, de la distribution ou encore de la restauration. C’est massif en valeur absolue, même si la part de l’agriculture dans l’emploi mondial recule progressivement : supérieure à 40 % au début des années 1990, elle se situe désormais autour du quart. Imaginer l’arrivée de robots humanoïdes dans ces secteurs, c’est donc se pencher sur l’avenir du travail, des compétences et des revenus de centaines de millions d’êtres humains.

Ensuite, une tendance. La population mondiale vieillit et les secteurs agricoles et alimentaires, dans toutes leurs composantes, n’échappent pas à la tendance. Complétons par l’érosion de l’attractivité, depuis de nombreuses années, pour les métiers productifs agricoles et ceux de l’industrie. Nous avons encore beaucoup de monde dans l’agriculture et l’agroalimentaire, comme indiqué au point précédent, mais ces personnes prennent de l’âge, ne sont pas souvent valorisées sur le plan sociétal et des salaires alors qu’elles exercent des métiers indispensables et difficiles, et leurs compétences ou savoir-faire peinent de plus en plus à être transmis ou renouvelés. Parler des humanoïdes en agriculture et dans l’agro-industrie, c’est inévitablement évoquer le défi du rapport au travail, dans des secteurs exigeants et pour lesquels peu de jeunes gens se tournent désormais…

En outre, un rappel. La technologie, le numérique et l’intelligence artificielle sont depuis longtemps très présents dans les secteurs agricoles et alimentaires. Pour entrer en détail dans ce jeu relationnel et en saisir à la fois l’amplitude et l’intérêt, il suffit de se pencher sur quelques travaux globaux récents[2] ou même certaines publications pédagogiques à l’échelle de la France[3]. Se projeter avec les humanoïdes, c’est anticiper le passage d’une intelligence artificielle (IA) qui assiste l’agriculteur ou le salarié d’une usine à une IA qui potentiellement agira seule, de manière autonome et pourrait même décider des tâches à accomplir. C’est aussi une IA qui oriente et structure les comportements alimentaires, avec pourquoi pas des robots en cuisine qui combinent recommandations nutritionnelles et capacités à préparer la juste ration pour servir les personnes…

Par ailleurs, une demande. Les attentes envers l’agriculture et l’alimentation sont immenses à l’heure des grandes transitions écologiques à mener. Préservation et régénération des écosystèmes, optimisation et résilience des rendements, décarbonation des activités, économie circulaire, réduction des pertes et des gaspillages sont autant de cadres stratégiques avec lesquels les acteurs agricoles et alimentaires doivent composer. Introduire la perspective d’humanoïdes dans ce registre, c’est se demander s’ils seraient plus à même que les humains à conduire ces transitions, accepter les changements et savoir s’y adapter. Ils pourraient travailler lorsque les températures deviennent difficiles, modifier instantanément leurs pratiques, optimiser l’usage de l’eau et de l’énergie ou appliquer sans résistance des protocoles nouveaux. Dit autrement, faire parfois mieux, plus vite et plus systématiquement ce que les humains, attachés à leurs habitudes, leurs intérêts et leurs contraintes économiques ou sociales, peinent à accomplir.

Enfin, la compétition géopolitique constitue la cinquième dynamique qui conditionne notre réflexion générale sur le sujet. Derrière les humanoïdes se cachent des jeux de puissance entre la Chine et les États-Unis notamment, mais pas uniquement. Le Japon, confronté à des virages sociodémographiques et agricoles considérables, ou encore l’Arabie saoudite, soucieuse d’investir le futur avec de l’IA et d’avoir une armée de travailleurs à domicile, participent d’ores et déjà, et à leur manière, à cette grande aventure technologique en cours. Pékin semble augmenter la cadence pour industrialiser sa production d’humanoïdes[4]. Tesla effectue un mouvement important pour passer de l’automobile aux robots. L’IA physique, avec des humanoïdes déployés, sera à la fois un bouleversement dans le monde du travail mais également un terrain majeur de compétition entre puissances, États comme entreprises. D’ailleurs, Bill Gates vient de souligner combien l’essor fulgurant de l’IA bouleversera les sociétés en profondeur, au point de plaider pour que certains emplois demeurent réservés aux humains. Il liste néanmoins le potentiel incroyable de l’IA dans certains secteurs stratégiques, dont l’agriculture[5].

De la curiosité à l’anticipation

Démographie, emploi, climat, technologie et rapports de force suffisent pour cadrer l’essentiel des enjeux qui gravitent autour de l’arrivée d’humanoïdes dans nos existences, à court et à moyen terme. Il serait conseillé de s’y préparer. Pour l’agriculture et l’alimentation, cinq domaines balisent potentiellement le champ des réflexions : la recherche, la production, la transformation, la distribution et la consommation.

L’IA dope les capacités scientifiques. Pour la recherche agronomique et végétale, pour la performance des biotechnologies, pour la santé du Vivant et le bien-être animal, pour la fertilisation raisonnée des sols, les progrès pourraient s’avérer significatifs. Réduire le temps de l’expérimentation, accélérer l’analyse des résultats, favoriser le développement de solutions plus rapidement applicables sur le terrain : autant d’avancées concrètes d’ores et déjà constatées. L’arrivée des humanoïdes pourrait ouvrir une étape supplémentaire. Capables de manipuler, tester, observer, mesurer et répéter des expérimentations, éventuellement en continu, ils pourraient contribuer à rapprocher davantage l’IA du travail physique de laboratoire ou de terrain. Que deviendront les stratégies de recherche et développement (R&D) lorsque l’IA ne se contentera plus d’analyser les résultats des expériences, mais pourra aussi participer physiquement à leur réalisation ? Entreprises, instituts techniques et organismes scientifiques seraient alors concernés. Après l’automatisation du calcul scientifique pourrait ainsi venir celle d’une partie du geste scientifique. Si les humanoïdes permettent aux laboratoires de fonctionner plus longtemps, de multiplier les expérimentations et de raccourcir les cycles de R&D, alors la compétition géopolitique sur la robotique devient aussi une compétition sur la vitesse de production de l’innovation. Celui qui possède les meilleurs modèles d’IA et les meilleures capacités physiques d’expérimentation pourrait apprendre plus vite que ses concurrents.

En matière de production, le secteur agricole affiche déjà de fortes connexions avec les machines et les robots. Ne revenons pas sur la mécanisation (dont le tracteur est le symbole) qui sur le temps long aura été l’un des moteurs des révolutions agricoles passées, et qui reste encore attendue dans de nombreux pays, où la précarité technique prive de nombreux agriculteurs de meilleurs rendements et de solutions à même de réduire la pénibilité du travail. D’ailleurs, à ce sujet, soyons attentifs au développement des exosquelettes, qui permettent de soulager des douleurs, de faciliter des tâches et de prolonger l’effort dans l’agriculture. Il s’agit d’innovations regardées de près dans des sociétés où le vieillissement des producteurs s’accroît. Au Japon, leur moyenne d’âge frôle désormais les 70 ans ! Pas étonnant que le conglomérat Panasonic, via sa filiale Atoun, s’était activé sur ce registre dès la décennie 2010, comme le font désormais des universités et des start-ups nipponnes. Si le Japon semble vouloir habiller l’homme en robot avant d’envisager de mettre le robot à sa place, ce modèle questionne l’efficacité d’un tel dispositif et son coût. Car in fine, pour effectuer des activités difficiles, potentiellement dangereuses et surtout répétitives, le robot l’emportera. Au Japon, comme ailleurs. À noter aussi que si les grandes cultures (céréales, oléagineux, etc.) sont déjà très équipées en machines et robots, d’autres filières misent sur ces dynamiques pour aller de l’avant. Ainsi notamment du secteur des fruits et légumes, où beaucoup de choses restent à faire. La fragilité de ces productions, en plein champ ou sous serres, explique en partie cette situation. Mais la concentration des récoltes sur de courtes périodes, où des mois de culture se jouent, exige de pouvoir compter sur une main d’œuvre disponible, compétente et capable d’efforts fournis sur quelques jours. Nul doute que des innovations robotiques y seront déployées, surtout si le recours à des travailleurs étrangers s’estompe et qu’il faille maintenir des volumes sur les étals de consommateurs avides de végétal. Nul doute que l’entreprise Unitree Robotics sera aux aguets à ce sujet, à moins qu’elle ne le soit déjà.

Au niveau de la transformation, la robotisation progressera certainement plus nettement dans les dédales des usines agroalimentaires ou agro-industrielles. Reste à déterminer quelle place spécifique y prendront les humanoïdes. Leur avantage ne résidera probablement pas dans la répétition d’un geste standardisé, où le robot industriel classique excelle déjà, mais dans leur polyvalence et leur capacité à évoluer dans des espaces conçus pour les humains. Mais à la différence de ces derniers, ils auront sans doute l’avantage de ne jamais devoir s’arrêter, sauf pour changer de batteries ou les recharger…Vaste question donc quant à l’automatisation accrue demain dans ces usines, quand bien même les femmes et les hommes n’y disparaitront pas, mais seront davantage dans la supervision, le contrôle et la gestion. Nous aurons une coexistence, robots et humains, où chacun devra trouver sa place, son rôle et sa valeur ajoutée. C’est là aussi le défi : savoir identifier les moments, les lieux et les actions où la décision humaine restera primordiale, ou décisive, alors que l’IA physique et des humanoïdes seront à la manœuvre. Mais comme l’ont bien expliqué deux spécialistes de ces enjeux[6], l’essentiel du débat devrait surtout se polariser sur la faisabilité d’une part, et l’acceptabilité d’autre part. Dans le cas d’industries agroalimentaires, où les produits transformés proviennent du Vivant et dans lesquelles les problématiques de sécurité et traçabilité sanitaire sont immenses, serons-nous tous à l’aise si ce qui deviendra nourritures ne passe plus du tout dans les mains des êtres humains ? Certains y verront peut-être une rassurance, là où d’autres seront à l’inverse contrariés. Comme dans la santé et la médecine, les puissances de calcul, l’IA et les robots améliorent des diagnostics, des cadences et des rapidités d’exécution dans l’univers de l’agroalimentaire. Mais l’imbrication avec l’espèce humaine s’avère nécessaire, encore, sauf à concevoir que tout est remplaçable progressivement, y compris la responsabilité. Or dans le champ alimentaire, le consommateur est-il prêt à ne plus pouvoir connaître les origines, les processus et les interlocuteurs si les machines brouillent les cartes, rendent peu de comptes et ne répondent pas aux services après-ventes ?

Autant de défis que les acteurs de la logistique et de la distribution alimentaire auront aussi à préparer, vu la vitesse avec laquelle pour eux aussi les dynamiques d’IA embarquée se déploient. Forçons la fiction : des humanoïdes demain partout en magasin ? Après tout, les caisses sont devenues de plus en plus automatiques, scannant seules et efficacement les produits du client, tout heureux parfois de ne pas devoir échanger un mot avec l’interface machine qui froidement identifie les achats, présente l’addition et propose plusieurs modes de paiement. Nous avons encore néanmoins des personnes dans les rayons…même si en Chine ou ailleurs, des géants du numérique ont déjà mis en place des supermarchés sans présence humaine. Est-ce le souhait de tous les consommateurs ? Jusqu’où cette tendance peut-elle aller ? Allons-nous voir des humanoïdes s’installer en caisse, pour redonner un peu d’âme à ce passage obligé ? Ou allons-nous surtout les voir aller faire les courses, pour celles et ceux qui n’aiment pas cela, n’ont pas le temps et cherchent à diversifier leurs canaux d’approvisionnement à la maison ? Les livraisons à domicile explosent depuis quelques années, notamment pour les jeunes générations, mais aussi pour des personnes ou des couples qui veulent optimiser leur quotidien sans devoir cuisiner ou sans devoir aller acheter à manger. Osons même renverser le raisonnement : remplacer l’humain par la machine est-il toujours synonyme de déshumanisation ? Les conditions de travail parfois très précaires des livreurs des plateformes numériques invitent à nuancer cette évidence[7]. Un robot effectuant une tâche pénible, dangereuse ou socialement dévalorisée ne détruit pas nécessairement de l’humanité. Tout dépend de ce que la société fait des femmes et des hommes ainsi libérés de cette tâche. En tout cas, il sera assurément intéressant de voir comment les entreprises de la distribution comme celles de la restauration se comporteront avec la révolution humanoïde. Imaginons la plus grande enseigne mondiale de repas rapide, d’origine états-unienne, présente sur tous les continents, employant 2,5 millions de personnes actuellement, dont 75 000 en France, faire le choix de tout numérique, du tout robotique et d’un Ronald humanoïdisé à l’accueil du restaurant ou à la fenêtre du passage en drive…

Quant à la maison…faut-il encore des espaces pour cuisiner ou manger diront certains ? Si tout est livré et si plus personne ne partage de repas en même temps, dans une pièce dédiée, les architectes ont de belles conceptions inédites à envisager. Avant cela, constatons avec lucidité que les appareils électroménagers ont envahi nos cuisines bien avant l’invention du smartphone, mais que des robots sans cesse plus performants y prennent une place grandissante depuis quelques années ! Ne citons que le Thermomix pour ne pas y aller par quatre chemins, dont chaque nouvelle édition est augmentée de fonctions numériques. Qui sait, les prochaines versions proposeront peut-être un Thermomix à deux pattes, deux bras, à la fois capables de cuisiner, servir, ranger, nettoyer et commander en ligne tous les produits alimentaires qui commencent à manquer au domicile du propriétaire. À moins que cela soit cela le projet de Tesla avec Optimus, son humanoïde censé apporter une aide inédite à la maison. Que dira-t-il de nos sociétés, de nos relations sociales qui souvent s’animent autour d’un repas et qu’enregistra-t-il de nos vies les plus intimes, si tant est qu’il en existe encore. D’ailleurs, publier la photographie de son assiette constitue déjà l’une des images les plus abondantes sur Instagram. Alimentation, santé et assistance vont alors converger. Le robot connaîtrait nos goûts, nos allergies, nos objectifs nutritionnels et éventuellement certaines données médicales. Il pourrait adapter les portions, réduire le sel, surveiller le sucre, limiter le gaspillage ou ajuster un menu à l’activité physique de chacun. Nous passerions alors de l’alimentation personnalisée à l’alimentation personnellement administrée. Sinon des frais supplémentaires pour se soigner ou s’assurer, faute de comportements appropriés !

Ouvertures en conscience

Jusqu’à présent, nous avons surtout adapté l’environnement aux machines : chaînes de production, entrepôts automatisés, robots de traite, tracteurs autonomes ou bras robotisés. Avec l’humanoïde, la logique peut s’inverser. Nous cherchons désormais à fabriquer une machine capable de s’adapter à un monde conçu pour le corps humain. Peu de secteurs permettent d’observer cette révolution aussi bien que l’agriculture et l’alimentation. Pour nous nourrir, il faut semer, observer, récolter, traire, soigner, pêcher, porter, trier, découper, transformer, transporter, stocker, cuisiner et servir. Bref, mobiliser presque toute la palette des gestes humains.

La robotisation agricole est déjà une réalité. C’est également vrai dans l’élevage. Ces technologies répondent à plusieurs défis simultanés : productivité, manque de main-d’œuvre, pénibilité, réduction des intrants ou adaptation au changement climatique. À ce titre, et si les robots s’adaptaient plus vite que nous ? Les systèmes agricoles et alimentaires doivent simultanément réduire leurs émissions, économiser l’eau, protéger les sols, limiter certains intrants, préserver la biodiversité et maintenir une production suffisante. Or le paradoxe des transitions est connu : nous savons souvent ce qu’il faudrait modifier, mais les choses changent… lentement. La machine, le robot et l’humanoïde ne craindront pas de modifier des habitudes. Cela ne signifie évidemment pas qu’ils résoudront la crise climatique, d’autant que leur fabrication et leur utilisation seront énergivores… Mais ils pourraient devenir de puissants accélérateurs d’adaptation. La question écologique rejoindrait alors celle de la délégation : sommes-nous prêts à laisser les machines effectuer plus rapidement certaines transformations que nous n’arrivons pas collectivement à accomplir ?

Alors oui, le futur de l’agriculture robotisée ne sera pas nécessairement humanoïde. Pourquoi donner deux jambes à une machine lorsqu’un système roulant est plus stable ? Pourquoi reproduire une main humaine lorsqu’un outil spécialisé est plus efficace ? Dans une parcelle céréalière de plusieurs centaines d’hectares, le tracteur autonome conservera probablement longtemps l’avantage. Mais dans une serre, un élevage, une cave, un laboratoire, un atelier ou une exploitation horticole conçus autour des déplacements humains, la polyvalence de l’humanoïde pourrait faire sens, plus rapidement sans doute que ce que l’on pense. Et pourquoi s’arrêter à la terre ? Pêche et aquaculture connaissent elles aussi une automatisation croissante : surveillance des poissons, alimentation automatisée, inspection des cages, robots sous-marins, tri, manutention ou transformation. Dans une ferme aquacole ou un port, les environnements difficiles et parfois dangereux constituent un puissant moteur de robotisation. L’expérience en mer, sur un navire, peut s’avérer en revanche moins favorable, sauf si les humanoïdes font preuve d’un pied marin inattendu…

Après l’agriculture de précision pourrait ainsi venir l’agriculture de délégation. Dans l’industrie agroalimentaire, l’IA incarnée devrait également s’exprimer. Les robots servent déjà des plats et des boissons dans certains restaurants et l’humanoïde se prépare à entrer dans le domicile. Préparons tout cela, si nous voulons trouver le bon équilibre, car il ne s’agit pas de refuser l’évolution technique, mais bien d’anticiper au mieux ces ruptures potentielles et de définir les contours de la responsabilité future. C’est d’ailleurs le message principal de l’Encyclique du Pape Léon publiée cette année[8], à lire sans hésiter même si l’on n’est pas croyant. Le Pape y rappelle que les systèmes d’IA, aussi performants soient-ils, ne possèdent pas de conscience morale et ne peuvent « assumer le poids des conséquences » (paragraphe 99). Les humanoïdes ne vont pas envahir demain matin nos champs, nos usines, nos magasins et nos cuisines. Mais ils nous obligent dès aujourd’hui à poser une question essentielle : non pas seulement ce que nous accepterons de leur faire faire, mais ce que nous accepterons de leur laisser décider. Car produire, transformer, transporter ou préparer notre alimentation engage toujours une responsabilité. À mesure que les machines deviendront plus autonomes, il faudra donc déterminer où s’arrête la délégation et où doit demeurer la décision humaine.

Sur le terrain géopolitique, la course est lancée entre une Chine qui se veut championne mondiale de la robotique industrielle et entend distancer les États-Unis sur la gamme d’humanoïdes performants et peu coûteux. Quid de l’Europe, qui voudra assurément moraliser ces sujets tout en explorant les avantages de ces innovations technologiques pour les secteurs qu’elle jugera stratégiques. Sans doute sera-t-elle tentée de vouloir mettre en œuvre une robotique humaniste et tracer des frontières hermétiques pour limiter le champ d’action des humanoïdes. Une question de souveraineté se posera également. Qui maîtrisera les données captées par ces machines, les modèles d’IA qui les pilotent, leurs composants, leurs mises à jour et éventuellement leur contrôle à distance ? La souveraineté alimentaire pourrait donc demain intégrer la maîtrise des machines intelligentes qui concourent à nous nourrir. Quant à l’Afrique, allons-nous robotiser un continent jeune, où l’agriculture reste le premier employeur et où des millions d’emplois seront à créer dans les prochaines années au regard de la dynamique socio-démographique ? Les robots et les humanoïdes n’y seront pas une réponse à la raréfaction de la main-d’œuvre. S’ils se déploient là-bas, préparons-nous en Europe à devenir plus humanistes.

[1] L’édition 2028 du Déméter, ouvrage annuel de prospective sur les questions agricoles et alimentaires, réalisé et édité conjointement par le Club DEMETER et IRIS Éditions, portera précisément sur la dialectique humanités / humanoïdes. Cet ouvrage paraîtra en février 2028.

[2] FAO, The State of Food and Agriculture, Leveraging automation in agriculture for transforming agrifood systems, SOFA Report, 2022 ; Drew Ashton, AI-Powered Agriculture : the next green revolution, eBookklt, 2024 ; Alka Arora, Rajender Parsad, Rajni Jain, Sudeep Marwaha, Harvesting Intelligence: The AI Revolution in Agriculture. From Fields to Algorithms, Cultivating Future Harvests, Springer, 2026.

[3] Manon Longvixay, François Brun, Christian Germain, Nouha Jeema, Guy Waksman, Intelligence Artificielle Générative en agriculture Enjeux, principes et applications, Rapport conjoint de l’Académie d’Agriculture de France, ACTA et Chaire AgroTic, 2026 ; Karine Cailleaux-Breton, David Joulin et Hervé Pillaud, L’IA des champs. Les graines d’un futur possible, La France Agricole Editions, 2026.

[4] The Economist, « China is training up thousands of humanoid robots », 23 August 2026.

[5]  Bill Gates, « The turbulent AI era is here. The choices we make now are critical », Blog, Gates Notes, 26 August 2026.

[6] Eric Hazan et Olivier Sibony, Faut-il encore décider ? La décision humaine à l’ère de l’intelligence artificielle, Flammarion, 2026.

[7] Marwân-Al-Qays Bousmah, Kevin Poperl, Anne Gosselin, Katy Fenech, Ana Rivadeneyra, et al., Santé et conditions de vie et de travail des livreurs des plateformes numériques en France : rapport de l’enquête Santé-Course, Ined, CoopCycle-Association, Maison des Coursiers, Maison des Livreurs, Médecins du Monde, Ville de Paris, 2026.

[8] Pape Léon XIV, Magnifica Humanitas, Lettre encyclique sur la protection de la personne humaine à l’ère de l’intelligence artificielle (Édition française, Artège, 2026).

L’article Humanoïdes : révolutions agricoles et alimentaires ? est apparu en premier sur IRIS.