Part Three of the analysis “The Twilight of the Unipolar Order: Economy, Security, Rearmament, and Immigration in the European Crisis” by Massimiliano Carta, currently being published in five installments by the Machiavelli Foundation. The first two installments are The Crisis of the European Union’s Foundations: Theses and Assessment and The Myth of the Self-Regulating Market: Theory, Political Power, and the Architecture of the Euro [ed.].
European strategic autonomy is at stake on two fronts that public debate tends to keep separate. The first is external: the war in Ukraine, uncertainty regarding U.S. protection, and rivalry with China have reignited the debate over rearmament, its timeline, and its obstacles. The second lies within the Euro-Atlantic relationship: a dependence on the United States that extends not only to defense but also to technology, digital infrastructure, energy, and finance—and which raises the question of whether the convergence between Europe and the United States represents a partnership of equals or a form of limited sovereignty. This article brings these two fronts together because they share the same limitation: the gap between the political intention of autonomy and the concrete ability to achieve it.
Why Europe Is Rearming
Ukraine and the End of an Illusion
The immediate trigger for European rearmament was Russia’s 2022 invasion of Ukraine, which disproved the assumption—upon which Europe had built three decades of the “peace dividend”—that a large-scale conventional war on the continent was now unthinkable: According to the European Defense Agency, military spending by EU member states has grown by more than 60% in real terms between 2020 and 2025, reaching approximately 381 billion euros.
The concept of the “peace dividend,” however, requires some clarification. It posited the absence of wars between established European states—an assumption that remained largely valid, at least in the perception of Western capitals, until 2022, despite the signs from Georgia (2008) and Crimea (2014)—not the absence of war in Europe tout court, which was disproved by the wars that led to the breakup of Yugoslavia (1991–1999). These conflicts involved direct NATO intervention through aerial bombing campaigns (Deliberate Force in 1995, Allied Force in 1999) and resulted, in Bosnia and Herzegovina alone, in approximately 100,000 deaths and the Srebrenica genocide. An unresolved legacy remains: the use of depleted uranium munitions in Bosnia and Kosovo, whose link to cancer cases among deployed military personnel is still controversial. A 2001 European Parliament resolution noted that a direct link had not been proven; however, an Italian parliamentary commission of inquiry (2006) and research by the Istituto Superiore di Sanità found a statistically significant excess of Hodgkin’s lymphoma among Italian military personnel, though they were unable to isolate the role of depleted uranium from other toxic factors present in the same theaters of operation, such as asbestos and heavy metals. The post-1989 “pacification” was therefore partial and not without long-term costs: a fact that complicates—without refuting—the narrative of a continent that supposedly rediscovered war only in 2022.
Uncertainty Regarding the American Ally
A second factor, likely more decisive in recent times, is the uncertainty regarding the continued existence of the U.S. security umbrella, which emerged with particular clarity during the Greenland crisis. In January 2026, President Donald Trump threatened to impose tariffs on Denmark, Norway, Sweden, France, Germany, the United Kingdom, the Netherlands, and Finland, effective February 1, “until” an agreement was reached on the transfer of the island, without initially ruling out the use of force. The withdrawal of the threat on January 22, following a meeting with NATO Secretary General Mark Rutte in Davos, did not end the crisis: the Greenlandic government and Danish Prime Minister Mette Frederiksen reiterated that “sovereignty is non-negotiable”; Denmark sent reinforcements to the island; European governments considered countermeasures, including the EU’s anti-coercion instrument; and the issue faced bipartisan opposition even in the U.S. Congress. At the NATO summit in Ankara in July 2026, Trump reiterated his position, and as of that date, the crisis remained unresolved. The episode affected European perceptions more than any statement on American disengagement, because it showed one NATO member openly threatening the territorial integrity of another.
The most emblematic response came from Germany. On April 22, 2026, Germany presented its first independent postwar military strategy, “Verantwortung für Europa,” with the goal of increasing active-duty forces from 185,000 to 260,000 by the mid-2030s, supplemented by a reserve of at least 200,000: a historic shift for a country that for decades had based its political identity on a culture of military restraint.
Two Threats of a Different Nature
Russia: An Existential Threat or a Limited Risk?
Whether Russia poses an existential threat to the entire continent or a limited and largely reactive risk remains one of the most contested issues in the strategic debate. The skeptical position, rooted in realism (represented by John Mearsheimer, among others), points to certain facts: Russia’s GDP is comparable to that of medium-sized European economies such as Italy or Spain, and the Russian military has faced considerable difficulties, with heavy losses, over more than four years of war against a single medium-sized state. According to this interpretation, the invasion of Ukraine—however criminal it may be under international law—would be a response to NATO’s expansion, perceived as an existential threat, with limited objectives, rather than the first phase of a continental conquest; a direct attack on a member of the Alliance, triggering Article 5, would be an irrational gamble for Moscow.
The opposing view, prevalent among Eastern European governments and in Atlantic intelligence circles, does not foresee a large-scale invasion, but rather a more targeted risk: a test of Article 5 against a peripheral and vulnerable target, such as the Baltic states, which have substantial Russian-speaking minorities. It cites the sequence of events in Georgia (2008), Crimea (2014), and Ukraine (2022), which has already repeatedly disproved the assumption that “Russia will not attack.” The disagreement, therefore, does not concern a total invasion of Europe—which almost no analyst foresees—but rather the nature—defensive or revisionist—of Russian ambitions, upon which the proportionality of rearmament depends.
China: A Systemic Rival, Not a Direct Military Threat
China poses a different kind of problem. Since 2019, the European Union has defined China—using a phrase that has become standard—as a partner, economic competitor, and systemic rival; at the 2024 Washington summit, NATO identified it as a “decisive enabler” of the Russian war, due to the support provided through exports of dual-use goods. Not everyone agrees with this tougher stance: Hungary, while signing the final declaration, ruled out turning the Alliance into an anti-China bloc, and within the EU, systemic rivalry is not unanimously accepted as the dominant framework for relations with Beijing. In fact, China has never condemned the invasion of Ukraine nor called it an act of aggression, and its trade with Russia has grown significantly since 2022, partially offsetting the impact of Western sanctions.
For Europe, however, the Chinese threat is not one of territorial invasion, but of structural vulnerability: dependence on rare earths, critical components, and semiconductors; the risk of a systemic shock in the event of a crisis in the Taiwan Strait; and, since 2026, the so-called “China Shock 2.0,” which, by hitting the heart of Germany’s and Europe’s advanced manufacturing sector, makes the threat less abstract and more immediately political. For this reason, rearmament aimed at countering China primarily concerns industrial resilience and the security of critical infrastructure, rather than conventional capabilities. The situation is different for the United States, for which Taiwan would also pose a direct military risk. Its commitment is not a mutual defense treaty comparable to Article 5: since 1979, the Taiwan Relations Act has established a policy of “strategic ambiguity,” which commits Washington to viewing a Chinese attack as a “matter of grave concern” and to providing Taipei with means of self-defense, without guaranteeing direct intervention—a choice that remains discretionary and is the subject of lively debate, given that the military balance in the Strait is increasingly favorable to Beijing. In any case, a conflict would expose U.S. bases in the Western Pacific—Guam, about 2,700 kilometers from Taiwan, and the Japanese installations in Okinawa—which are already within range of Chinese missiles, while a conquest of the island would offer Beijing a forward base from which to project power further afield; the U.S. mainland would remain at risk only in a scenario—unlikely but with strong deterrent value—of nuclear escalation.
Timelines and Obstacles to Rearmament
The formal targets converge on 2030. The Readiness 2030 plan (formerly ReArm Europe) aims to mobilize up to 800 billion euros in additional defense investment, while NATO has set a target for 2025 of 5% of GDP by 2035, of which 3.5% is for defense in the strict sense: a level that, according to the European Defense Agency, would require EU member states alone to spend over 630 billion euros per year. However, commitment remains uneven: by the end of July 2025, only ten of the twenty-seven member states had announced the explicit goal of reaching at least 3% of GDP by 2030.
The real bottleneck, according to the most recent analyses, is not financial but industrial: the ability to transform resources into delivered weapons. In Germany, defense orders nearly doubled between 2025 and early 2026, while production grew only marginally; with average delivery times estimated at three years, orders would need to be placed within the next two years to meet the 2030 deadline. This is clearly illustrated by Poland’s turn to South Korean industry—approximately 180 K2 tanks, over 200 K9 self-propelled howitzers, and 48 FA-50 aircraft, totaling more than $12 billion—which has made South Korea Warsaw’s second-largest supplier after the United States: not due to a lack of European technology, but because of insufficient European production capacity at the required pace. The SAFE fund itself, created to strengthen European industrial sovereignty, is opening up to non-European manufacturers just over a year after its creation.
Added to these obstacles is a constraint rarely discussed: the demographic and fiscal sustainability of rearmament. Eurostat projects a 13% decline in the working-age population (ages 15–64) between 2022 and 2060, with a fertility rate of approximately 1.4 children per woman (1.38 in 2023) compared to a replacement level of 2.1. A rearmament effort that requires a skilled industrial workforce, military personnel, and growing fiscal resources thus comes up against a shrinking demographic base—the very same factor that makes funding pensions increasingly burdensome and fuels Europe’s demand for foreign labor. In light of these dynamics, the realistic timeframe indicated by independent analysts—including the Hudson Institute—for credible European rearmament falls between 2026 and 2032, while full strategic autonomy—an industrial capacity independent of U.S. support in terms of scale and speed—would require at least a decade, and in any case, the political will to uphold the commitments made.
A Dependence Deeper Than the Military
Subordination or Convergence?
Rearmament measures Europe’s military dependence, but it does not exhaust it. The more strictly political question is whether the convergence between European and U.S. interests is the result of a genuine affinity or the expression of a structurally asymmetrical relationship, which it is legitimate to define, without excessive euphemisms, as subordination.
Several arguments support this second thesis. Institutionally, NATO was created to ensure a permanent U.S. military presence on the continent,
with an integrated command led by the United States. Historically, the European Union’s enlargements toward Central and Eastern Europe have always followed—never preceded—those of NATO. And decisions made within the Atlantic alliance can have consequences that Europe bears alone: the 2011 intervention in Libya, which toppled Gaddafi and led to the collapse of the Libyan state, opened the Central Mediterranean migration route, with the burden falling primarily on border countries, starting with Italy.
Four Structural Dependencies
Unlike military dependence—which arose from a postwar political choice and is, in principle, reversible—technological, infrastructural, energy, and financial dependence stems from market dynamics—economies of scale, network effects, and first-mover advantage—and is therefore more insidious: there is no treaty from which to withdraw.
On the technological front, Amazon Web Services, Microsoft Azure, and Google Cloud control approximately 70% of the European cloud market, and in June 2026, the European Commission decided to subject them to the requirements of the Digital Markets Act. The issue is also jurisdictional: the U.S. CLOUD Act allows American authorities to request access to data held by U.S. companies wherever it is stored, which conflicts with the GDPR, meaning that data hosted in Frankfurt or Paris remains exposed to non-European authorities. Countermeasures remain incomplete: the Gaia-X project, launched in 2019, has not bridged the scale gap, and the AWS European Sovereign Cloud, launched in January 2026 with 7.8 billion euros in investments in Brandenburg and European staff, remains an Amazon product—one of operational sovereignty, not ownership. The gap is even more pronounced in artificial intelligence: according to a report by MIT researcher Michiel Bakker (June 2026), Europe accounts for about 5% of the world’s computing capacity dedicated to AI, compared to 80% in the United States; Mistral, Europe’s leading AI lab, has revenue and capital amounting to only a small fraction of those of OpenAI and also relies on Microsoft Azure’s infrastructure, with which it formed a partnership in 2024; and the public InvestAI initiative, with its 200 billion euros, still falls short of the amount that major U.S. tech companies alone spend annually on AI infrastructure—in the range of several hundred billion dollars.
Even further upstream lies the internet infrastructure. Approximately 95–99% of intercontinental traffic travels via undersea cables, of which the four American giants—Alphabet, Meta, Amazon, and Microsoft—now use between 66% and 70% of the capacity, according to estimates by TeleGeography and the Atlantic Council, compared to less than 10% in 2012; consortia of national telecommunications operators are now joined by private cables built exclusively for platform traffic. Even the logical coordination of the network is concentrated: of the twelve entities that manage the thirteen root DNS servers, eight are U.S.-based—including NASA, the Department of Defense, and the U.S. Army—three are non-U.S. (Netnod, RIPE NCC, WIDE Project), and one is ICANN, whose oversight shifted to a multistakeholder model in 2016. Here, unlike with the cloud, there are no European alternatives of comparable scale, nor is there an alternative to the root servers that would not risk fragmenting the network.
On the energy front, the attempt to reduce one dependency has created another. According to the Institute for Energy Economics and Financial Analysis (May 2026), in 2026 the United States will supply about two-thirds of the liquefied natural gas imported by the European Union, becoming its leading gas supplier ahead of Norway, and that share could rise to as much as 80% by 2028–2030; U.S. LNG is, on average, the most expensive for European buyers, while Russia remains the second-largest supplier of LNG, having set a quarterly import record in early 2026. Commission Executive Vice President Teresa Ribera warned that the EU risks replacing one dependency with another, rather than truly diversifying.
On the financial front, BlackRock, Vanguard, and State Street collectively manage approximately $26,000 billion and, according to Bruegel, hold 16.4% of the British stock market and 19% of the Irish stock market, with a presence in Europe that is growing faster than that of European asset managers. The issue is not just one of ownership but of governance: European fund managers vote in favor of shareholder resolutions on social and environmental issues about 80% of the time, while their U.S. counterparts—led by BlackRock—saw their support drop from 49% to 17% between 2021 and 2024. In other words, along with capital, corporate governance standards are also being imported.
European regulatory capacity remains a reality, however—the GDPR has imposed European standards even on American giants, the so-called “Brussels effect”—and the internet infrastructure, whose protocols are developed by an open, non-governmental body such as the Internet Engineering Task Force, cannot be reduced to national ownership. It remains to be seen, however, whether regulation can translate into substantive autonomy without public industrial investment comparable in scale to that of the United States: this is the same gap between intention and capacity that has already emerged with regard to rearmament.
Convergence Among Peers and the Strongest Argument
The opposing argument notes that the Union enjoys full legal autonomy, an internal market larger than that of the United States, and a foreign policy that is nominally autonomous, however fragmented; that convergence is based on genuine affinities—democratic systems, integrated market economies, similar legal systems—; and that military dependence was a rational choice, which made it possible to allocate resources to the welfare state: after 1990, Europe could have established an autonomous defense system at any time, and did not do so for reasons of domestic politics, not because of an American veto.
The strongest argument in favor of the thesis of subordination, however, emerges on a purely political level. On issues such as Iran, China, and the Middle East, European positions tend to align with Washington even when European economic interests are objectively divergent—consider German exports to Iran or French and Italian investments in the Middle East. This discrepancy between economic interest and political alignment is difficult to explain solely by shared values and suggests de facto limited sovereignty, rather than full parity among allies—an interpretation that the four dependencies just outlined make even more concrete.
Why the Responses Remain Incomplete
In the face of such well-documented dependencies, European responses remain incomplete, inadequate, or uncoordinated, and part of the explanation lies in institutional factors. The Common Foreign and Security Policy requires unanimity among the twenty-seven member states: an instrument of cohesion that has, in practice, become a tool for vetoing decisions. Between 2011 and 2025, depending on the source, there were between 45 and 48 vetoes, with a marked increase in recent years, mostly from Hungary (between 19 and 21) and Poland (7). Not all of these reflect substantive disagreements: In 2020, Cyprus blocked sanctions against Belarus to exert pressure in a bilateral dispute with Turkey, and Slovakia and Malta withdrew their veto on the 18th package of sanctions against Russia in July 2025 only after securing concessions regarding gas supplies and maritime regulation.
In 2026, according to Human Rights Watch, the government of Viktor Orbán—the most frequent user of the veto—was defeated in the elections, and Commission President Ursula von der Leyen, with the support of High Representative Kaja Kallas, revived qualified majority voting in foreign policy. The proposal, however, remains trapped in a paradox: to be approved, it in turn requires unanimity—including from states such as Poland, which view the veto as an indispensable guarantee of their influence.
Europe, of course, is not standing idly by: on every front, there are at least embryonic initiatives—the SAFE Fund and Readiness 2030 for defense, the Digital Markets Act and InvestAI for technology, REPowerEU for energy, and the Savings and Investment Union for finance. None, however, has public investment on a scale comparable to that of its competitors, nor a decision-making mechanism capable of translating intent into timely action. Mario Draghi, author of the 2024 landmark report on European competitiveness, highlighted this at the informal summit in Alden-Biesen in February 2026, denouncing a “deterioration of the economic landscape” and the urgency of addressing the fragmentation of stock markets, energy costs, and the need for a targeted “European preference” in certain sectors. On the decision-making front, Draghi referred to the enhanced cooperation mechanisms provided for in the treaties—introduced by the Treaty of Amsterdam (1997) and never before used on a large scale for industrial policy—which allow a group of states to proceed without waiting for unanimity among the 27 member states: a way to circumvent the veto rather than abolish it, which is more realistic in the short term but depends on the formation of a critical mass of countries willing to move forward.
Conclusions
Rearmament and dependence are two sides of the same problem. External threats are real but asymmetric—the Russian threat is military and disputed in its scope, while the Chinese threat is systemic and industrial—and the resulting rearmament is constrained more by production capacity and demographics than by financial resources. Dependence on the United States, in turn, extends beyond defense to encompass technology, networks, energy, and finance, to the point of influencing European political alignment even against its own economic interests.
In both cases, the diagnosis is clear and the tools exist on paper; what is lacking is the ability to translate them into timely action on an adequate scale, hindered by a decision-making mechanism that allows individual states to block joint action. European strategic autonomy is, more than a matter of budgets, first and foremost a matter of institutions.
Bibliographic Note
The data and references cited in this article are drawn, among other sources, from the European Defense Agency, the Hudson Institute, the European Parliament, the Italian National Institute of Health, Eurostat, the European Commission, Bruegel, TeleGeography, the Atlantic Council Digital Forensic Research Lab, ICANN, IEEFA (Institute for Energy Economics and Financial Analysis), MIT (Michiel Bakker), Human Rights Watch, official statements by NATO leaders, and Mario Draghi’s remarks at the informal EU summit in Alden-Biesen (February 2026), as well as theoretical contributions by John Mearsheimer. Data updated as of July 2026.