Hi folks,
In today’s edition of the newsletter, I am covering two pieces from the latest Qiushi journal publication.
The first article is by Dr. Sun Chenghao, Senior Fellow and head of US-EU program at Center for International Security and Strategy, Tsinghua University.
The piece deals with the complexities of the US-EU relationship at present and in the future. Sun’s core argument is that the transatlantic relationship is undergoing severe strain, but it is unlikely to see a divorce. This is because it has evolved over the decades to be driven by “institutional mechanisms”; these provide “a platform for deliberation and consultation in decision-making, but also, through bureaucratic cooperation, civil-military coordination, military interoperability/coordination, and long-term cooperation, creates path dependence and raises the cost of ‘breaking up’ or ‘going separate ways’.”
That said, he believes that the “principal logic” of the transatlantic alliance is undergoing a transformation. This is reflected in differences over economic policy, identity and international responsibility. He contends that the US is shifting its approach from “leading the West” to “managing the alliance” to serve its changing interests. At the same time, the EU is putting strategic autonomy into practice. And the purpose of Europe’s strategic autonomy “is not to break away from the US, but rather to enhance its own capacity for action, ability to bear risks, and policy autonomy, while continuing to maintain the alliance relationship.” So, this will be the primary contradiction that will lie at the heart of the transatlantic alliance going forward.
The second article is by Li Yan, Director of the Institute of Sci-Tech and Cyber Security Studies at the China Institutes of Contemporary International Relations.
The piece basically pushes back at the criticism levelled against Chinese AI companies for engaging in distillation to steal American firms’ IP. Li argues that the US government and American companies are redefining what is a legitimate technical practice within a national security frame in order to stymie the rise of Chinese companies and monopolise AI tools. Li claims that this is being done because Chinese AI advancements have hit a “sensitive nerve in the United States”, which is already dealing with “strategic anxiety” around its relative decline. She then pivots to argue that what’s made this worse for the US is that the Chinese models “remain committed to an open-source and open approach”. Li frames this as China’s ‘universally beneficial and inclusive model (which) is fundamentally changing the global landscape of competition in AI.” Towards the end, Li explicitly calls to “establish a global governance framework (on AI) enjoying broad consensus”, while pitching China as the flag-bearer of the Global South.
Here are the two pieces:
In recent years, the transatlantic alliance relationship has suffered strong shocks. While strengthening ‘America First’, the current US administration has demanded that Europe shoulder more security responsibility, and has continued to pressure Europe in areas such as trade and technology; Europe, for its part, while maintaining alliance cooperation, has accelerated the advancement of strategic autonomy and the building of defense capabilities. The differences between the two sides over defense burden-sharing, economic and trade rules, technology governance, and the international order have grown increasingly prominent. Analyzing in depth the historical logic of the evolution of the transatlantic alliance relationship, the deep roots of the current fissures, its future prospects, and its impact on the international configuration holds practical significance for accurately grasping the profound changes unseen in a century and the development trend of world multipolarization. 近年来，跨大西洋同盟关系遭受强烈冲击。现任美国政府在强化“美国优先”的同时，要求欧洲承担更多安全责任，并在贸易、科技等领域持续对欧施压；欧洲则在维持同盟合作的同时，加快推进战略自主和防务能力建设。双方围绕防务分担、经贸规则、技术治理和国际秩序的分歧日益突出。深入分析跨大西洋同盟关系演变的历史逻辑、当前裂痕的深层根源、未来前景及其对国际格局的影响，对于准确把握百年变局和世界多极化发展趋势具有现实意义.
The transatlantic alliance is not simply an alliance formed for the sake of interests, but rather a long-term alliance jointly underpinned by three factors: the security structure, institutional support, and identification based on shared ideas or values. Since the end of the Second World War, the transatlantic alliance has undergone multiple rounds of tests, including Cold War confrontation, the dissolution of the Soviet Union and the restructuring of the international order, the impact of the war on terror, and great-power competition. 跨大西洋同盟并非简单的因利而合，而是由安全结构、制度维系和观念认同三重因素共同支撑的长期同盟关系。自第二次世界大战结束以来，跨大西洋同盟关系历经冷战对抗、苏联解体与秩序重构、反恐战争冲击以及大国博弈等多轮考验.
To meet the needs of responding to the Cold War, in 1949, 12 countries including the United States, Britain, and France signed the North Atlantic Treaty, establishing the North Atlantic Treaty Organization (NATO) with the so-called ‘collective defense’ as its core principle; its provisions stipulate that an armed attack on any one member state shall be regarded as an attack on the entire alliance. The establishment of NATO marked a high degree of interlocking between the US and Europe on security. In this period, the core logic of the alliance relationship was that of a ‘security community’, the essence of which lay in jointly responding to external threats. The United States maintained a large-scale military presence in Europe over the long term and played the role of NATO’s leader, while most European countries relied heavily on the US nuclear umbrella as well as its advanced military and intelligence capabilities. Although the United States and Europe differed on specific issues such as strategic autonomy, nuclear policy, and military deployments, these contradictions were consistently overshadowed by the more pressing security requirements. The facts of this period demonstrate that, so long as the external threat was sufficiently strong, internal contradictions within the US-European alliance occupied a secondary position. 为应对冷战的需要，1949年，美国、英国、法国等12国签署《北大西洋公约》，以所谓的“集体防御”为核心原则建立北大西洋公约组织（简称“北约”），其条款规定，对任何一个成员国的武装攻击都将被视为对整个同盟的攻击。北约的建立标志着美欧在安全上高度嵌合。这一时期，同盟关系的核心逻辑是“安全共同体”，其本质在于共同应对外部威胁。美国长期在欧洲保持大规模驻军，扮演北约领导者的角色，多数欧洲国家则高度依赖美国核保护伞及其先进军事和情报能力。尽管美欧在战略自主、核政策、军事部署等具体问题上存在分歧，但这些矛盾始终被掩盖于更优先的安全需求之下。这一时期的事实表明，只要外部威胁足够强烈，美欧同盟内部矛盾便处于次要地位。
After the end of the Cold War, the external threats faced by the US and Europe declined substantially; if one followed the assumptions of traditional alliance theory, the transatlantic alliance should have moved toward looseness or even dissolution. Yet the reality was precisely the opposite: NATO not only did not disappear, but achieved expansion and functional transformation. At the end of the Cold War, NATO had only 16 member states, whereas as of July 2026 the membership had expanded to 32, doubling in scale. NATO’s functions also gradually expanded from traditional collective defense to multiple domains such as regional intervention, counter-terrorism cooperation, and cybersecurity. The key lies in the institutionalized operation of the alliance, i.e., the transatlantic institutional system centered on NATO, which, through rules, procedures, and organizational structures, has mechanized/institutionalised US-European security cooperation, shifting the alliance from being ‘threat-driven’ to ‘mechanism-driven’. 为应对冷战的需要，1949年，美国、英国、法国等12国签署《北大西洋公约》，以所谓的“集体防御”为核心原则建立北大西洋公约组织（简称“北约”），其条款规定，对任何一个成员国的武装攻击都将被视为对整个同盟的攻击。北约的建立标志着美欧在安全上高度嵌合。这一时期，同盟关系的核心逻辑是“安全共同体”，其本质在于共同应对外部威胁。美国长期在欧洲保持大规模驻军，扮演北约领导者的角色，多数欧洲国家则高度依赖美国核保护伞及其先进军事和情报能力。尽管美欧在战略自主、核政策、军事部署等具体问题上存在分歧，但这些矛盾始终被掩盖于更优先的安全需求之下。这一时期的事实表明，只要外部威胁足够强烈，美欧同盟内部矛盾便处于次要地位.
A stable institutional mechanism not only provides a platform for deliberation and consultation in decision-making, but also, through bureaucratic cooperation, civil-military coordination, military interoperability/coordination, and long-term cooperation, creates path dependence and raises the cost of ‘breaking up’ or ‘going separate ways’. Even though in recent years the US has become less willing to invest in European security and Europe has continually emphasized strategic autonomy, the institutional continuity of NATO still keeps both sides within the alliance framework. As the Russia-Ukraine conflict has dragged on, differences have emerged between the US and Europe over the sharing of defense costs, but European countries, driven by security needs, have still sharply increased defense spending and set up more defense-cooperation mechanisms. This further shows that the mechanism itself has become an important foundation for sustaining the transatlantic alliance relationship. 冷战结束后，美欧面临的外部威胁大幅下降，若遵循传统同盟理论假设，跨大西洋同盟本应走向松散甚至解体。然而现实情况恰恰相反，北约不仅没有消失，反而实现扩展与功能转型。冷战结束时，北约成员国仅16个，而截至2026年7月，成员国已扩至32个，规模翻了一倍。北约职能也从传统集体防御逐渐扩展至地区干预、反恐合作、网络安全等多个领域。其中关键在于同盟的制度化运作，即以北约为核心的跨大西洋制度体系，通过规则、程序、组织结构，将美欧安全合作机制化，使同盟从“威胁驱动”转向“机制驱动”。稳定的机制不仅提供了决策商讨平台，还通过官僚协作、军政配合、军事协同与长期合作，形成路径依赖，提高“分家”成本。即便近年来美国对欧洲安全投入意愿下降、欧洲不断强调战略自主，但北约的制度延续性仍将双方框定于同盟架构内。随着俄乌冲突延宕，美欧在防务成本分摊问题上出现分歧，但欧洲各国仍迫于安全需要大幅提升防务开支，并增设更多防务合作机制。这进一步说明，机制本身已成为维系跨大西洋同盟关系的重要基础.
The transatlantic alliance is not merely about security cooperation, but also constitutes a political and shared identity built upon so-called value concepts such as ‘freedom and democracy’. Unlike the security structure and the institutional mechanisms, the shared sense of identity does not play a role only in a particular historical stage, but is a cultural bond and spiritual foundation running through the entire process of the transatlantic alliance’s development. During the Cold War, the US and Europe gradually coalesced around a shared political and ideological value identity, including the liberal-democratic system, the market-economy model, and a ‘rules-based international order’. From the 1990s to the early 21st century, the US and Europe generally believed that their own institutional model possessed attractiveness, and their strategic focus gradually shifted from maintaining the existing security order to shaping and expanding the international order they dominated. Whether it was NATO’s eastward expansion, the EU’s eastward expansion, or external actions carried out in the name of ‘democracy promotion’, all carried behind them a distinct coloring of the export of values. Even after major differences emerged over issues such as the Iraq War, the US and Europe still maintained a relatively high degree of consistency in their basic cognition regarding the so-called ‘liberal international order’, and they had a common discourse and value foundation on issues such as global governance, the reform of international organizations, and the formulation of international rules, turning the transatlantic alliance from a traditional military alliance into a strategic relationship possessing a distinct political character and bloc character. 跨大西洋同盟不仅是安全合作，更是一种建立在所谓“自由、民主”等价值理念上的政治与意识形态认同。与安全结构和制度机制不同，观念认同并非只在某一历史阶段发挥作用，而是贯穿跨大西洋同盟发展全过程的文化纽带和精神基础。冷战时期，美欧逐步凝聚起政治与意识形态价值认同，包括自由民主制度、市场经济模式、“基于规则的国际秩序”等。20世纪90年代至21世纪初，美欧普遍认为自身制度模式具有吸引力，其战略重心也从维护既有安全秩序逐步转向塑造和扩展其所主导的国际秩序。无论是北约东扩、欧盟东扩，还是以“民主推广”为名义开展的对外行动，其背后都带有鲜明的价值理念输出色彩。即便在伊拉克战争等议题上出现重大分歧后，美欧围绕所谓“自由主义国际秩序”的基本认知仍保持较高一致性，在全球治理、国际组织改革、国际规则制定等问题上有着共同话语和价值基础，使跨大西洋同盟从传统军事联盟发展成为具有鲜明政治属性和阵营色彩的战略关系.
In recent years, the economic and trade frictions, disputes over defense burden-sharing, and differences in strategic perceptions within the transatlantic alliance have, on the surface, been specific policy contradictions; in reality, however, they reflect profound changes taking place in the economic foundations, political ideas, and value consensus that have long underpinned cooperation between the two sides. 近年来，跨大西洋同盟关系中的经济贸易摩擦、防务分担争议、战略认知分歧，表面上是具体政策矛盾，实则反映出支撑双方长期合作的经济基础、政治理念、价值共识正在发生深刻变化.
At the economic level, economic globalization has not only changed the relationship between the West and the world, but has also weakened the US-European consensus on economic development. Free trade, multiculturalism, and Western-dominated international institutional framework formed after the Cold War were once important pillars through which the US and Europe jointly promoted the so-called ‘liberal international order’. The US and Europe generally believed that open markets, capital flows, personnel exchanges, and institutional expansion could not only raise economic efficiency, but also drive the global order to develop in a direction favourable to their long-term dominance. However, as economic globalization advanced in depth, problems such as the relocation of industries abroad, the loss of manufacturing jobs, the widening of income gaps, and imbalances in regional development kept accumulating, and some groups gradually became disaffected with economic globalization. Especially after the 2008 international financial crisis, Western society’s confidence in free markets and economic globalization declined markedly, and an increasing number of people began to question whether an open economic system could continue to bring universal prosperity. On the choice of path for responding to the shocks of economic globalization, US-European differences have become increasingly evident. The US turned relatively early toward a new economic strategy centered on industrial policy, trade protection, and supply-chain security, increasingly stressing reliance on the power of the sovereign state to strengthen national capacity and to reshape strategic planning and resource allocation, so as to revive the domestic economy and enhance strategic competitiveness. Whether it is America First, the reshoring of manufacturing, or the strengthening of industrial subsidies and economic security policies, they all reflect its adjustment or even revision of the neoliberal-dominated model of economic globalization. By comparison, although Europe likewise faces problems such as declining industrial competitiveness and weak economic growth, constrained by its supranational governance model and its member-state coordination mechanisms, on the whole, it still hopes to seek reform within the framework of open markets, free trade, and the multilateral economic system, and is more inclined to advance policy adjustments through rules-based frameworks and collective consultation. Differences between the US and Europe have, therefore, continued to increase in such areas as trade, industrial policy, economic security, and the restructuring of supply chains.
经济层面，经济全球化不仅改变了西方与世界的关系，也削弱了美欧的经济发展共识。冷战后形成的自由贸易、多元文化主义及西方主导的国际制度框架，曾是美欧共同推动所谓“自由主义国际秩序”的重要支柱。美欧普遍相信，开放市场、资本流动、人员往来和制度扩展不仅能够提高经济效率，也能够推动全球秩序朝着有利于其长期主导的方向发展。然而，随着经济全球化深入推进，产业外迁、制造业岗位流失、收入差距扩大、区域发展失衡等问题持续积累，部分群体逐渐沦为经济全球化的“失意者”。特别是2008年国际金融危机后，西方社会对自由市场和经济全球化的信心明显下降，越来越多民众开始质疑开放经济体系能否继续带来普遍繁荣。在应对经济全球化冲击的路径选择上，美欧分歧日益明显。美国较早转向以产业政策、贸易保护和供应链安全为核心的新经济战略，越来越强调依托主权国家力量强化国家能力、重塑战略规划和资源配置，以重振国内经济并提升战略竞争力。无论是“美国优先”、制造业回流，还是强化产业补贴和经济安全政策，都体现出其对新自由主义主导的经济全球化模式的调整甚至修正。相比来说，欧洲虽同样面临产业竞争力下降、经济增长乏力等问题，但受制于超国家治理模式和成员国协调机制，总体上仍希望在开放市场、自由贸易和多边经济体系框架内寻求改革，更倾向于通过规则框架和集体协商推进政策调整。美欧在贸易、产业政策、经济安全以及供应链重组等领域的分歧不断增多.
At the political level, questions of identity and ‘boundaries’ have become its core issue. During the Cold War, the Soviet threat to a large extent masked the differences between the US and Europe on immigration, identity, and social governance. After the end of the Cold War, as the common external threat weakened and economic globalization developed in depth, the questions of who belongs to the ‘West’ and how the boundaries of the community should be defined gradually became important issues of political debate within the US and Europe. The US increasingly stresses the priority of national borders, sovereign belonging, and citizenship, and tends to build the political community on the basis of the nation-state; Europe, by contrast, has long been committed to supranational integration, seeking to construct a broader European identity beyond national identity. In recent years, the migration crisis, the refugee problem, and the difficulties of social integration have further magnified this difference. As a result, the differences between the US and Europe are no longer merely a dispute over specific immigration policies, but also a dispute over identity revolving around ‘Who are the people of the West’ (西方人是谁) and ‘Where is the boundary of the West.’ (西方边界在哪里). 政治层面，身份和“边界”问题成为其核心议题。冷战时期，苏联威胁很大程度上掩盖了美欧在移民、身份认同、社会治理上的差异。冷战结束后，随着共同外部威胁弱化和经济全球化深入发展，谁属于“西方”、共同体边界应如何界定，逐渐成为美欧内部政治争论的重要议题。美国越来越强调国家边界、主权归属、公民身份的优先性，倾向于将政治共同体建立在民族国家基础之上；欧洲则长期致力于超国家整合，试图在国家认同之外构建更广泛的欧洲身份认同。近年来，移民危机、难民问题以及社会融合困境进一步放大这种分歧。由此，美欧之间的分歧已不仅是具体移民政策之争，而且是围绕“西方人是谁”、“西方边界在哪里”的认同之争.
At the level of international responsibility, differences are also emerging between the US and Europe. After the end of the Cold War, the United States and Europe generally regarded themselves as shapers of the global order and promoters of rules, viewing multilateralism, free trade, human rights narratives, and the export of institutions as important manifestations of their strengths. In recent years, however, the two sides have gradually diverged in their understanding of what kind of international responsibilities they should assume and how the international order should be shaped.
The current US administration no longer regards maintaining the West and global order as a ‘responsibility’, but rather stresses America First, and demands that Europe prove the value of the alliance at a higher cost. Europe, by contrast, still tends to regard multilateralism, the expansion of norms, and global governance as important components of its international identity, and remains wary of the US’s increasingly evident unilateralism. Discussions in Europe concerning American reliability, strategic autonomy, and European responsibility have continued to intensify, reflecting Europe’s concern over changes in the US role and its reconsideration of its own positioning. These differences in the understanding of international responsibility have caused policy differences between the two sides to continue widening on issues, including security, trade and economics, technology, climate change, and global governance. 国际责任层面，美欧也在产生分歧。冷战结束后，美欧普遍将自身视为全球秩序塑造者和规则推广者，把多边主义、自由贸易、人权叙事、制度输出作为自身优势的重要体现。近年来，双方对于应承担何种国际责任、如何塑造国际秩序的认识逐渐分化。现任美国政府不再把维护西方和全球秩序视为“责任”，而是强调“美国优先”，并要求欧洲以更高成本证明其同盟价值。欧洲则仍倾向于将多边主义、规范扩展、全球治理视为其国际身份的重要组成部分，并对美国日益明显的单边主义保持警惕。欧洲关于“美国可靠性”、“战略自主”、“欧洲责任”的讨论持续升温，反映出欧洲对美国角色变化的担忧和对自身定位的重新思考。这种国际责任理解上的差异，使双方在安全、经贸、科技、气候变化和全球治理等议题上的政策分歧不断扩大.
It should, however, be pointed out that the current changes in the transatlantic alliance do not mean that the common interests and foundations for cooperation between the US and Europe have disappeared altogether. Rather, the principal logic that has sustained the functioning of the alliance is undergoing adjustment. Although it is increasingly difficult for the two sides to maintain the highly consistent strategic consensus of the post-Cold War period, common needs in terms of security interests, institutional ties, and global interests still exist. Whether in responding to regional security risks, maintaining transatlantic economic ties, or coordinating on global issues, the two sides still need to cooperate with one another, and neither can easily bear the costs of a complete breakdown of the alliance. For precisely this reason, the transatlantic alliance will not disintegrate simply because differences have increased. Rather, it is more likely to enter a new phase of adjustment under conditions in which common interests continue to exist while common understandings gradually weaken. 应当指出的是，当前跨大西洋同盟关系发生的变化，并不意味着美欧共同利益和合作基础已全然消失，而是维系同盟运行的主要逻辑正发生调整。双方虽越来越难以维持冷战后高度一致的战略共识，但在安全利益、制度联系、全球利益上的共同需求仍然存在。无论是应对地区安全风险、维护跨大西洋经济联系，还是协调全球性议题，双方仍需彼此合作，也都难以承受同盟关系全面破裂的代价。正因如此，跨大西洋同盟不会因分歧增多而走向瓦解，而更可能在共同利益仍然存在、共同认知逐渐弱化的条件下，进入新的调整阶段.
From a more long-term perspective, the security cooperation, economic and trade ties, and institutional coordination within the transatlantic alliance relationship will continue to play an important role, but their internal cohesion will be more subject to the influence of practical needs and strategic considerations. 从更长远角度看，跨大西洋同盟关系中的安全合作、经贸联系和制度协调仍将继续发挥重要作用，但其内在凝聚力将更多受到现实需求和战略考量的影响.
From the US perspective, its attitude toward the transatlantic alliance is shifting from ‘leading the West’ to ‘managing the alliance’. This is because the strategic priorities of the United States, both domestically and internationally, are changing. Compared with the post-Cold War period, when it long assumed the role of the ‘guardian’ of the Western liberal international order, the US is increasingly emphasizing giving priority to addressing such issues as domestic economic competitiveness, industrial development, and social governance. It also has tended to devote its limited resources to directions of greater strategic priority, such as the Indo-Pacific region. Accordingly, the importance of the transatlantic alliance has not declined in substantive terms, but its functions are undergoing adjustment. In the future, it will increasingly serve the US’s need to maintain its competitive advantages externally and shape a favourable strategic environment. Against this background, the US will not readily abandon Europe as an important ally. Its support for Europe will continue, but it will place increasing emphasis on cost-benefit considerations. 从美国角度看，其对待跨大西洋同盟的态度正从“领导西方”转向“管理同盟”。这是由于美国内外战略重心正发生变化，相较于冷战后长期承担西方自由主义国际秩序“维护者”的角色，美国越来越强调优先解决国内经济竞争力、产业发展、社会治理等问题，并倾向于将有限资源投入“印太”地区等更具战略优先性的方向。因此，跨大西洋同盟的重要性并未实质性下降，但其功能正发生调整，未来将更多服务于美国维护对外竞争优势和塑造有利战略环境的需要。在这一背景下，美国不会轻易放弃欧洲这一重要盟友，其对欧洲的支持仍将持续，但将愈发强调“成本收益”.
From the European perspective, the transatlantic alliance relationship is undergoing an adjustment from dependence on the US toward enhancing strategic autonomy. In recent years, Europe has continued to advance defense cooperation, the integration of the military-industrial sector, economic security, and the building of critical-technology capabilities; strategic autonomy is gradually being transformed from a political proposition into actual policy. However, differences exist within Europe in terms of levels of development, security perceptions, and strategic culture. Moreover, the nuclear deterrent, intelligence system, and strategic power-projection capabilities provided by the US will remain difficult to replace in the short term. Therefore, the practical objective of European strategic autonomy is not to break away from the US, but rather to enhance its own capacity for action, ability to bear risks, and policy autonomy, while continuing to maintain the alliance relationship. At the same time, Europe is becoming increasingly aware that genuine strategic autonomy is reflected not only in the development of military capabilities, but also in industrial competitiveness, technological innovation capability, and the ability to shape global rules. How to strengthen its own strategic agency while maintaining cooperation with the US will become an important challenge that Europe faces in managing the transatlantic alliance in the future. 从欧洲角度看，跨大西洋同盟关系正经历从依赖美国向增强战略自主的调整。近年来，欧洲持续推动防务合作、军工整合、经济安全和关键技术能力建设，战略自主正逐步从政治主张转化为现实政策。但欧洲内部发展水平、安全认知、战略文化存在差异，且美国提供的核威慑、情报体系和战略投送能力短时期仍难以被替代。因此，欧洲战略自主的现实目标并非摆脱美国，而是在继续维持同盟关系的同时，提高自身行动能力、风险承受能力和政策自主性。同时，欧洲越来越意识到，真正的战略自主不仅体现在军事能力建设上，也体现在产业竞争力、科技创新能力、全球规则塑造能力等方面。如何在保持对美合作的同时增强自身战略能动性，将成为未来欧洲处理跨大西洋同盟关系时面临的重要挑战.
Changes in the transatlantic alliance will further exert profound impacts on the international system and global governance. From the perspective of the international system, internal adjustments within the transatlantic alliance mean that the ability of the US and Europe to dominate the international agenda and shape international rules will decline relatively. In the future, international politics will increasingly be characterized by interaction and coordination among different forces, rather than by an order arranged under the dominance of a single power. The Global South and emerging major powers will continue to gain greater autonomy in international affairs, and their discursive power in such areas as reform of the global governance system, international financial governance, global climate governance, and digital governance will be further strengthened. From the perspective of global governance, the growing divergence in understanding and interests within the transatlantic alliance may weaken the ability of the relevant countries to coordinate on certain international issues, giving the global governance system more of the characteristics of being consultative, pluralistic, and decentralized. In the future, the formation of international rules will depend increasingly on coordination and cooperation among different countries and regions, rather than being driven primarily by a small number of Western countries. 跨大西洋同盟关系之变将进一步对国际体系和全球治理产生深远影响。从国际体系层面看，跨大西洋同盟内部调整意味着美欧主导国际议程、塑造国际规则的能力将相对下降。未来国际政治将越来越表现为不同力量之间的互动与协调，而非单一力量主导下的秩序安排。全球南方国家和新兴大国在国际事务中的自主性将持续上升，在全球治理体系改革、国际金融治理、全球气候治理、数字治理等领域的话语权将进一步增强。从全球治理层面看，跨大西洋同盟内部认知和利益分化可能削弱相关国家在部分国际议题上的协调能力，使全球治理体系更加呈现协商化、多元化、分散化特征。未来国际规则的形成将更多依赖不同国家和地区之间的协调与合作，而不再主要由少数西方国家主导推动.
The profound adjustment of the transatlantic alliance relationship also shows that the old security logic built on the foundation of exclusive alliances, camp divisions, and bloc politics is increasingly ill-suited to the trend of the times of world multipolarization, economic globalization, and the democratization of international relations. As the international balance of power undergoes profound adjustment and developing countries rise as a group, an increasing number of countries hope to participate in international affairs on an equal footing and jointly shape international rules, rather than passively accepting institutional arrangements dominated by a small number of Western countries. It is precisely against this backdrop that the Global South countries, represented by China, hold high the glorious banner of building a community with a shared future for mankind, conform to the trend of the times of the democratization of international relations, uphold true multilateralism, and advocate an equal and orderly world multipolarization and a universally beneficial and inclusive economic globalization, which will surely drive the international order to develop in a more just and reasonable direction. 跨大西洋同盟关系的深刻调整也表明，建立在排他性同盟、阵营划分和集团政治基础上的旧安全逻辑，越来越难以适应世界多极化、经济全球化和国际关系民主化的时代潮流。随着国际力量对比深刻调整和发展中国家群体性崛起，越来越多国家希望以平等身份参与国际事务，共同塑造国际规则，而不是被动接受少数西方国家主导的制度安排。正是在这一背景下，以中国为代表的全球南方国家高举构建人类命运共同体的光辉旗帜，顺应国际关系民主化的时代潮流，坚持真正的多边主义，倡导平等有序的世界多极化和普惠包容的经济全球化，必将推动国际秩序朝着更加公正合理的方向发展.
Recently, a rather absurd development has emerged in the field of AI: a foundational technique known as ‘distillation’, which has been widely used in the technology community for more than a decade, has suddenly been imbued with security and political overtones. Some American politicians have, with great seriousness, accused Chinese AI companies of ‘industrial-scale distilling’ of American models, essentially packaging a method originally intended to lower the threshold for computing power requirements and promote the popularization of technology and innovation as a byword for ‘intellectual property theft’. Congressional hearings, White House memoranda, and joint accusations by companies have followed one after another, creating quite a spectacle. But upon closer examination, the entire affair appears suspicious at every turn. Viewed through the lens of a longer historical perspective, this is nothing more than an act of hegemony cloaked in the garb of technology, a replay of an old script on a new stage. 近来，人工智能领域出现一个颇为荒诞的事件：一个在技术界通行了10余年的基础技术——“蒸馏”，突然被赋予了安全和政治色彩。美国一些政要煞有介事地指控中国人工智能企业“工业级蒸馏”美国模型，将一个原本旨在降低算力门槛、推动技术普及创新的操作方法包装成“知识产权窃取”的代名词。国会听证、白宫备忘录、企业联名控诉接踵而至，好不热闹。但此事细究起来，处处透着蹊跷。如果把这些疑问放在更长的历史镜头下审视，这不过是一场披着技术外衣的霸权行径，是旧剧本在新舞台上的重演.
To explain this farce clearly, it is first necessary to clarify what ‘distillation’ is. Distillation is originally a foundational technique in the field of AI. In 2015, Geoffrey Hinton, who is hailed as the ‘father of deep learning’, and others formally introduced this term for the technique. Its core logic is as follows: a large model with abundant computing power serves as the ‘master’, while a lightweight small model serves as the ‘apprentice’. The latter learns the former’s way of solving problems by observing its inputs and outputs, ultimately replicating its capabilities. The value of this technical approach lies in the fact that it enables a model that is smaller in size and more efficient to run to achieve performance close to that of cutting-edge large models while substantially reducing computing requirements. For this reason, distillation quickly became a widely recognized standard practice within the global AI community. American technology giants such as Google, OpenAI, and Amazon have not only made extensive use of this technique, but have even provided it externally as a standardized service. In other words, distillation is, in essence, a neutral technical method. It is one of the important means by which AI technology undergoes iterative improvement and innovative development, and an important pathway for lowering technological barriers and promoting the inclusive development of AI. 要讲清楚这场闹剧，还得先厘清何为“蒸馏”。“蒸馏”本是人工智能领域的一项基础技术。2015年，被誉为“深度学习之父”的杰弗里·辛顿等人明确提出这一技术命名，其核心逻辑是：让算力充沛的大模型做“师傅”，轻量级的小模型做“徒弟”，后者通过观察前者的输入输出，学习其解决问题的思路和方法，最终实现能力复刻。这种技术路径的价值在于，它能让一个体积更小、运行更高效的模型，在大幅降低算力需求的同时，获得接近尖端大模型的性能表现。正因如此，“蒸馏”迅速成为全球人工智能界公认的常规操作。谷歌、开放人工智能公司（OpenAI）、亚马逊等美国科技巨头，不仅广泛使用这一技术，甚至将其作为标准化服务对外提供。换言之，“蒸馏”本质上是一种中性的技术方法，是人工智能技术迭代与创新演进的重要方式之一，是降低技术门槛、推动人工智能普惠发展的重要路径.
However, this very technology, a foundational technique that is technically neutral and widely used throughout the industry, has quietly been given a different face in the narrative promoted by certain political and business forces in the United States. The trigger for this farce can be traced back to 23 February 2026, when the American AI company Anthropic suddenly issued a statement accusing three Chinese AI companies—DeepSeek, Moonshot, and MiniMax —of carrying out so-called ‘industrial-scale distillation attacks’ against its AI models. It claimed that these companies had ‘systematically extracted the capabilities of American AI models, posing ‘a significant national security risk’ to the United States. What is particularly intriguing is that OpenAI, Anthropic, and Google—three companies engaged in intense competition with one another—nevertheless set aside their differences and, through coordinating their positions and sharing information, jointly established an inter-company anti-distillation coordination mechanism. 然而，就是这样一项技术中性、业界通用的基础技术，在美国某些政商势力的叙事中，却被悄然换了面目。这场闹剧的导火索，可以追溯到2026年2月23日，美国人工智能公司Anthropic突然发布通告，指控深度求索（DeepSeek）、月之暗面、稀宇科技（MiniMax）三家中国人工智能企业，对其人工智能模型实施了所谓“工业规模的蒸馏攻击”，称这些公司“系统性抽取美国人工智能模型能力”，对美国造成“重大国家安全风险”。令人玩味的是，开放人工智能公司、Anthropic、谷歌这三家彼此竞争激烈的企业，竟不约而同放下嫌隙，通过协调立场、共享信息，联手建立企业间反“蒸馏”协同机制。
The US government quickly followed suit. On 23 April, Michael Kratsios, Director of the White House Office of Science and Technology Policy, wrote that certain foreign entities were ‘deliberately distilling American frontier AI systems on an industrial scale’. At the same time, the US House Committee on Foreign Affairs was advancing a package of export-control bills, including the ‘Stop the Theft of American AI Models Act’. Meanwhile, the US Senate Committee on the Judiciary held a hearing to discuss the issue of ‘distillation’ within the narrative framework of ‘the theft of American intellectual property’. By this point, ‘distillation’ had been thoroughly politicized and securitized, transforming it from a neutral technical term. Evidently, the US government also felt that forcibly changing the face of an already widely accepted technical term did not hold water, so it made a great fuss over the wording, deliberately using exaggerated and aggressive terms such as ‘adversarial distillation’ and ‘industrial-scale distillation’ to forcibly distinguish it from normal ‘distillation’ technique. Relevant experts have pointed out that such terminology is clearly difficult to sustain as a technical distinction. More ironically, Anthropic was itself accused of downloading approximately seven million pirated books from a ‘shadow library’ to train its models, and paid at least US$1.5 billion in settlement. In response, Elon Musk, CEO of Tesla, mocked the situation, saying: ‘How can someone steal something that Anthropic stole from human programmers?’ With a single remark, he pierced the false façade of this campaign to shape public opinion. 美国政府很快跟进。4月23日，白宫科技政策办公室主任迈克尔·克拉齐奥斯发文称，某些外国实体“以蓄意、工业规模的方式蒸馏美国前沿人工智能系统”。与此同时，众议院外交事务委员会正在推进一揽子出口管制法案，其中就包括《阻止美国人工智能模型窃取法案》等；参议院司法委员会召开听证会，在“窃取美国知识产权”的叙事框架下讨论“蒸馏”议题。至此，“蒸馏”从一个中性的技术术语，彻底被政治化、安全化。显然，美国政府也觉得强行把一个已经被广泛接受的技术术语改头换面说不过去，就在表述上大做文章，刻意以“对抗性蒸馏”、“工业级蒸馏”等充满夸大和攻击性的词语，强行与正常“蒸馏”技术相区别。相关专家指出，这种表述显然在技术区分上难以成立。更具讽刺意味的是，Anthropic曾因涉嫌从“影子图书馆”下载约700万本盗版书籍来训练模型，支付了至少15亿美元和解金。对此，特斯拉首席执行官马斯克揶揄道：“怎么有人偷Anthropic从人类程序员那里偷来的东西？”一语戳破这场舆论造势的假面具.
Stripping away the layers of rhetorical packaging, the root of this ‘distillation’ farce lies in the fact that the rapid development of China’s AI has touched a sensitive nerve in the United States. In the midst of intense strategic anxiety, it has deliberately fabricated allegations of intellectual property infringement in order to preserve its own AI hegemony. At present, competition in AI is becoming increasingly intense, and the pace of model iteration is continuously accelerating. Chinese large language models have achieved significant breakthroughs in such areas as coding capability, mathematical reasoning, and multimodal integration, with some models already occupying a globally leading position. Since DeepSeek became a breakout success, Moonshot recently released its open-source Kimi K3 model, whose performance is already comparable to the advanced proprietary models of American companies, such as OpenAI and Anthropic. Against this backdrop, the anxiety of leading American AI companies is genuine. At present, the cost of a single training run for cutting-edge American LLM has already reached the order of hundreds of millions of US dollars, placing enormous pressure on research and development expenditures. What causes the United States even greater concern is that China’s large AI models remain committed to an open-source and open approach. Developers around the world are all able to download them free of charge, use them, and further develop them. This universally beneficial and inclusive model is fundamentally changing the global landscape of competition in AI. 剥开层层话语包装，这场“蒸馏”闹剧的根源是中国人工智能的快速发展触动了美国的敏感神经，其在极度战略焦虑下故意采取捏造侵权罪名的方式，以护持自身人工智能霸权。当前，人工智能竞争日趋激烈，模型迭代加速，中国大模型在代码能力、数学推理、多模态融合等领域取得显著突破，部分模型能力已处于全球领先地位。自DeepSeek火爆出圈之后，近期月之暗面发布的Kimi K3开源模型，效果已可媲美美国开放人工智能和Anthropic等闭源公司的先进模型。对此，美国头部人工智能企业的恐慌是真实的。目前，美国前沿大模型单次训练投入已达数亿美元量级，研发投入成本压力巨大。更令美国担忧的是，中国人工智能大模型坚持开源开放，全球开发者均可免费下载、使用和再开发，这种普惠模式正在从根本上改变全球人工智能竞争格局.
Therefore, it is by no means accidental that American political and business forces have chosen this particular juncture to hype up the issue of ‘distillation’. First, they seek to stigmatize the development of Chinese AI by attributing the technological advances of Chinese models to ‘theft’ rather than indigenous innovation, thereby weakening the global influence of China’s AI technology. Second, they seek to conduct public opinion mobilization and lay policy groundwork for even stricter technology controls in the future. Following restrictions on chips, they intend to direct regulatory tools more precisely toward the level of models and algorithms. Third, by constructing a narrative of ‘security threats’, they seek to forge consensus among domestic political parties, thereby providing legitimacy for technological containment of China. As a spokesperson for China’s Ministry of Commerce pointed out, the US side disregards such facts as that the capabilities of some Chinese models have already reached a world-leading level, and labels Chinese companies with the accusation that they rely on ‘distillation to steal American intellectual property’. Such accusations lack a factual basis, lack legal justification, and apply double standards in practice. They constitute a typical act of AI hegemonism. In fact, the research paper on DeepSeek-R1 published in September 2025 in the journal Nature explained precisely how reinforcement learning was used to improve the model’s reasoning capabilities, and did not adopt the ‘distillation’ approach. 因此，美国政商势力选择在这个节点炒作“蒸馏”议题绝非偶然。其一，污名化中国人工智能发展，将中国模型的技术进步归因于“窃取”而非自主创新，以削弱中国人工智能技术的全球影响力；其二，为后续更严厉的技术管制做舆论动员和政策铺垫，在芯片封锁之后，再将管制工具精准指向模型和算法层面；其三，通过制造“安全威胁”叙事，试图捏合其国内各党派共识，为对华技术遏制提供合法性。正如中国商务部发言人指出的，美方罔顾中国部分模型能力已处于世界领先地位等事实，给中国企业扣上所谓依靠“蒸馏窃取美国知识产权”等帽子，缺乏事实依据，没有法理支撑，在实践上搞双重标准，属于典型的人工智能霸权主义行径。事实上，在2025年9月《自然》杂志刊发的关于DeepSeek—R1相关研究论文，阐明的正是如何依靠强化学习提升模型推理能力，恰恰没有采用“蒸馏”路线.
If this ‘distillation’ controversy is viewed within a historical frame of reference, one discovers a clear pattern. The methods by which the US maintains its technological hegemony have never changed. In 2013, the Edward Snowden disclosures concerning the PRISM revealed that the National Security Agency had for years monitored global communications networks and carried out large-scale cyber intrusions against numerous countries. The breadth and depth of these activities were shocking. Yet, during that very same period, the US government repeatedly beat the drum of the so-called ‘China cyber threat’ narrative and played up stories of ‘Chinese hacker attacks’. Even though the US itself was the world’s largest ‘hacker empire’, it explained its own cyber operations as being ‘to safeguard national security interests’ and ‘consistent with international practice’. At the same time, it turned around and accused China of launching ‘cyber attacks’ in order to ‘steal intellectual property and commercial secrets’. More than 10 years have passed, and the same script is once again being played out in the field of AI. The methods are exactly the same, and the underlying logic is entirely unchanged. The United States has consistently repeated the same playbook at critical junctures in technological competition. First, it establishes hegemony through technological superiority; then, when that superiority is challenged, in order to suppress those catching up and its competitors, it coins a sensational new term, portrays itself as the ‘victim’, smears its opponents as ‘thieves’ and ‘copycats’, and uses this as the justification for imposing sanctions and carrying out suppression. It is simply old wine in a new bottle, utterly lacking in originality. 如果把这场“蒸馏”争议放在历史坐标中观察，便会发现一个明显的规律：美国维护自身技术霸权的套路，从来没有变过。2013年，斯诺登披露的“棱镜计划”显示，美国国家安全局长期监听全球通信网络，对多国实施大规模网络入侵，范围之广、程度之深，令人震惊。然而，就在同一时期，美国政府不断鼓噪所谓“中国网络威胁论”，渲染“中国黑客攻击”。明明美国自己是世界上最大的“黑客帝国”，却将其网络攻击行为解释为“维护国家安全利益”、“符合国际惯例”，反过来污称中国为了“窃取知识产权和商业机密”发动“网络攻击”。10多年过去，同样的剧本在人工智能领域再次上演，手法一模一样，逻辑一脉相承。美国始终在重要技术竞争节点上重复同样的操作：先以技术优势建立霸权，再在优势受到挑战时，为打压追赶者和竞争者，制造一个耸人听闻的新词，把自己包装成“受害者”，将对手抹黑为“小偷”、“抄袭者”，以此为由进行制裁和打压。可谓新壶装旧酒，毫无新意.
However, this time the United States has run into stiff resistance. Unlike the era of the PRISM scandal over a decade ago, the international community today—and the global tech sector, tempered by the information age, in particular—possesses considerable discernment; America’s old playbook is, therefore, destined to fail. The most direct counterevidence comes from within the United States itself. In July 2026, when the US government threatened to sanction Chinese AI companies and restrict American developers’ use of Chinese open-source AI models, Silicon Valley’s reaction was unexpected. Nearly 200 Silicon Valley start-ups jointly sent a letter to the US government explicitly opposing the cutting off of access to Chinese open-source AI models. The letter stated: ‘American leadership requires two things: world-leading American open-weight models, and ensuring that American developers can continue to use existing open models from around the world’. Jensen Huang, founder of Nvidia, together with more than 20 important American companies and organizations, including Microsoft, Dell Technologies, and Hugging Face, issued a joint letter publicly supporting open-weight AI models. Jensen Huang said that ‘Chinese models are very good. Excellent open-source models ought to be used.’ He added that restricting open-source models on the grounds of national security ‘may instead make the United States more vulnerable.’ These voices of opposition from Silicon Valley convey a clear message: within the deeply interconnected technological ecosystem of the era of economic globalization, attempts to sever the flow of knowledge and monopolize technological advantages through political means constitute hegemonic behaviour that not only will not work externally, but will also encounter opposition domestically. Once a technology has become a driving force behind global innovation, any attempt to privatize it or weaponize it will ultimately backfire on oneself. 然而，这一次美国是踢到铁板了。与10多年前“棱镜门”时期不同，今天的国际社会，尤其是经历过信息网络时代洗礼的全球科技界，已经具备了相当的辨别力，美国的老套路注定难以奏效。最直接的反证来自美国内部。2026年7月，当美国政府威胁要制裁中国人工智能企业、限制美国开发者使用中国开源模型时，硅谷的反应出人意料。近200家硅谷初创企业联名致信美国政府，明确反对切断对中国开源人工智能模型的访问。信中写道：“美国的领导地位需要两样东西：世界领先的美国开放权重模型，以及让美国开发者能够持续使用全球范围内现有的开放模型。”英伟达创始人黄仁勋与微软、戴尔、抱抱脸（Hugging Face）等20多家美国重要企业与机构发布联名信，公开支持开放权重人工智能模型。黄仁勋表示，“中国模型非常出色，优秀的开源模型理应得到应用”，以国家安全为由限制开源模型，“可能反而使美国更加脆弱”。这些来自硅谷的反对声音，传递出一个清晰的信号：在经济全球化时代深度交织的技术生态中，试图通过政治手段切断知识流动、垄断技术优势的霸权行径，不仅对外行不通，对内也会遭到反对。当一项技术已经成为推动全球创新的驱动要素，任何试图将其私域化、武器化的行径，最终都将反噬自身.
On 17 July, in his keynote address at the opening ceremony of the 2026 World Artificial Intelligence Conference and High-level Meeting on Global AI Governance, General Secretary Xi Jinping stated clearly that ‘we should jointly oppose the practice of expanding the concept of national security in the field of AI and placing one’s own country’s security above that of other countries’. General Secretary Xi’s important judgment strikes at the crux of the AI hegemonic practices of a small number of countries. He put forward four proposals:
adhere to openness and win-win cooperation, and drive innovative development;
strengthen risk awareness and ensure security and controllability;
encourage inclusiveness and mutual accommodation, and promote mutual learning among civilizations;
advocate solidarity and working together in harmony, and improve global governance.
These provide important guidance for overcoming AI hegemony and jointly building a fair and reasonable global AI governance system. AI is a new engine for global economic growth and an accelerator for the transformation between old and new drivers of growth. All countries of the world should seize this rare historical opportunity, encourage open-source development and openness, cooperation and sharing, comprehensively promote AI-driven scientific and technological innovation, industrial development, and application scenarios, and enable AI to empower thousands of industries, so that it truly becomes a powerful source of positive energy for advancing human civilization. It is necessary to practice true multilateralism, effectively give full play to the important role of multilateral platforms such as the United Nations and the World AI Cooperation Organization, bring in the broad participation of developing countries in global AI governance, strengthen the alignment and coordination of AI development strategies, governance rules, and technical standards, break the pattern in which technical standards are monopolized by a small number of countries or enterprises, abandon exclusive and bloc-based governance models, and at an early date, establish a global governance framework enjoying broad consensus. In light of the serious lag in implementing the United Nations General Assembly Sustainable Development Goals, it is necessary to steadfastly and persistently advance the implementation of the UN General Assembly resolution on ‘Enhancing International Cooperation on AI Capacity-building’, help Global South countries strengthen their technological capacity-building, support all countries in tapping their own potential and distinctive advantages, and strengthen the international forces for open and jointly governed cooperation. 7月17日，习近平总书记在2026世界人工智能大会暨人工智能全球治理高级别会议开幕式上发表的主旨讲话中明确指出，“应当共同反对在人工智能领域泛化国家安全概念、把本国安全凌驾于他国安全之上的做法”。习近平总书记的重要论断切中少数国家人工智能霸权行径的要害，提出“坚持开放共赢，驱动创新发展”、“强化风险意识，确保安全可控”、“鼓励包容并蓄，促进文明互鉴”、“倡导和衷共济，完善全球治理”4点意见，为破解人工智能霸权、携手构建公正合理的全球人工智能治理体系提供了重要遵循。人工智能是世界经济增长的新引擎和新旧动能转换的加速器，世界各国都应抓住难得的历史性机遇，鼓励开源开放、合作共享，全面促进人工智能科技创新、产业发展、场景应用，让人工智能赋能千行百业，使其真正成为推动人类文明进步的强大正能量。要践行真正的多边主义，切实发挥联合国、世界人工智能合作组织等多边平台的重要作用，吸纳广大发展中国家参与全球人工智能治理，加强人工智能发展战略、治理规则、技术标准的对接协调，打破由少数国家或企业垄断技术标准的格局，摒弃排他性、阵营化治理模式，早日形成具有广泛共识的全球治理框架。针对联合国可持续发展目标执行严重滞后的实际，持之以恒推进落实“加强人工智能能力建设国际合作”联大决议，帮助全球南方国家加强技术能力建设，支持各国发掘自身潜力和独特优势，壮大开放共治的国际力量。
As the saying goes, ‘a single string cannot produce a tune; a lone tree does not make a forest’. The development of AI should not be a solo performed by one country, but rather a symphony of global cooperation. All countries should adhere to the principles of putting people first, AI for good, fairness and universal benefit, and collaborative co-governance. They should discuss and resolve the major issues and pressing challenges jointly faced by humanity in the intelligent age, join hands in opposing AI hegemonic practices, and ensure that AI promotes common development and truly benefits all of humanity. 单弦不成音，独木不成林。人工智能发展不应该是某个国家的独奏，而应当是全球合作的交响。各国应坚持以人为本、智能向善、公平普惠、协同共治原则，讨论和解决智能时代人类共同面对的重大问题、紧迫挑战，携手反对人工智能霸权行径，确保人工智能促进共同发展、真正造福全人类.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.