编者按：全球人工智能竞赛的焦点正在发生转变。7 月 17 日，吉尔吉斯斯坦前总理、中国人民大学重阳金融研究院外籍高级研究员卓奥玛尔特 · 奥托尔巴耶夫（Djoomart Otorbaev）在中国日报发表英文评论文章指出，人工智能的经济逻辑正在改变，而战略优势可能会转向那些能够在大规模制造生态系统中部署人工智能的国家，而不仅仅是能制造最先进算法的国家。现将中文译文及英文原文发布如下：（中文译文约 2600 字，预计阅读时间 7 分钟）

多年来，全球关于人工智能的讨论主要围绕大型语言模型、算法和算力展开。竞争看似简单明了：谁能打造最强大的神经网络，谁就能主导数字经济。这种认知驱动着美国领先企业展开了一场激烈的技术竞赛，参与者包括 OpenAI、谷歌、微软、亚马逊和 Anthropic。在中国，深度求索（DeepSeek）、阿里巴巴、腾讯、华为和百度等公司则通过加大对大模型、芯片和云基础设施的投入予以回应。

但过去一年里，一场更深层的转型已日渐清晰。人工智能正稳步超越聊天机器人、搜索引擎和办公工具，进入制造、物流、制药、汽车工程、机器人和工业规划等领域。因此，全球人工智能竞赛的焦点发生了转变。竞争的下一个阶段，或许将不再那么依赖算法，而是更多取决于生产能力、工业组织能力，以及将人工智能嵌入大规模经济活动的能力。

这一转变改变了人工智能竞争的本质。在第一阶段，主要比拼的是前沿模型和算力霸权，各公司竞相争夺顶尖工程师、半导体、云基础设施和风险投资，目标是取得技术领先。而工业部署则是一项复杂得多的挑战，需要软件开发商、制造商、供应链运营商、大学、工业工程师、能源系统和政府间的协同配合。换句话说，人工智能的未来越来越取决于生态系统，而非孤立的技术突破。这也正是战略联盟在人工智能领域遍地开花的原因。

在美国，科技公司与工业企业之间的合作势头日益强劲。微软扩大了与跨国汽车制造商斯特兰蒂斯（Stellantis）的合作，在工程、网络安全和制造系统等领域开发人工智能应用。谷歌云深化了与德国制药公司默克（Merck）的合作，在整个药物研发管线中部署人工智能工具，同时还扩大了与玛氏（Mars）的合作，将人工智能代理整合到运营和工业流程中。与此同时，英伟达（Nvidia）已从单纯的芯片供应商，演进为全球人工智能部署的基础设施基石，与云服务提供商、制造商、机器人开发者及各国技术生态系统同步开展工作。美国企业日益意识到，决定长期领导地位的将是产业融合能力，而不仅仅是模型质量。

与此同时，大学也已成为新兴人工智能产业格局中的关键角色。OpenAI 的 " 下一代人工智能联盟 "（NextGenAI）汇聚了美国和英国的一流大学及研究机构，为研究人员和学生提供资助、算力和应用程序接口（API）访问权限。Anthropic 也与专注于计算机科学和人工智能工程的学术机构及教育组织建立了类似的伙伴关系。这些合作并非仅仅出于慈善目的，它们有助于培养未来人才、加速应用研究、创建行业标准，并让未来的工程师熟悉特定的人工智能生态系统。实际上，大学已经成为国家人工智能供应链的一部分。

中国正在走一条格外独特的道路，因为它拥有许多西方国家难以轻易复制的结构性优势：一个能将人工智能直接融入制造生产的完备工业生态系统。中国企业并不只专注于面向消费者的人工智能应用，而是将人工智能整合进工厂、机器人、物流系统、半导体、电信基础设施、电动汽车和工业自动化中。

华为就是一个鲜明的例子。该公司扩大了与制造商、地方政府、电信运营商及一流大学的合作，以加速人工智能在各工业领域的部署。与清华大学、北京大学等机构的合作，推动了芯片设计、分布式计算、工业人工智能系统和优化技术的进步。在中国看来，人工智能不是孤立的数字化产品，而是嵌入整个经济的基础设施。

深度求索（DeepSeek）展示了另一个重要趋势。其技术团队包含与中国顶尖大学相关的专家，尤其是在数学、工程和计算机科学领域。据报道，这种紧密结合使该公司能够优化训练效率、降低成本并保持有竞争力的性能，证明了学术专长能够直接增强产业竞争力。与此同时，阿里云和腾讯正加快在零售、物流、金融和智慧城市生态系统中部署人工智能。中国电动汽车制造商也在将人工智能融入生产线、电池系统、自动驾驶和供应链管理中。从机器人到先进制造业，人工智能开发商与工业企业之间的合作正在日益制度化。

这一演变至关重要，因为人工智能的经济逻辑正在改变。早期的人工智能竞赛青睐那些能打造最大通用模型的公司，但这些大模型的训练和维护成本非常高昂。企业发现，集成到工业工作流程中的、更小型且专业化的人工智能系统，要比通用聊天机器人创造更大的经济价值。因此，战略优势可能会转向那些能够在大规模制造生态系统中部署人工智能的国家，而不仅仅是能制造最先进算法的国家。

这将有利于拥有深厚工业能力、完备供应链、强大工程传统，以及大学、产业和国家机构间紧密协调的经济体。人工智能已与能源基础设施、半导体生产、机器人、工业自动化和物流网络密不可分。在许多方面，新兴的人工智能竞争已开始类似于历史上工业强国间的竞争。全球人工智能竞赛的下一个阶段，可能不会主要在硅谷或由基准排名来决定，而是在工厂、港口、实验室、工业园区和供应链中展开。从这个意义上说，人工智能已不再只是一场软件革命，它已经成为一场新的工业革命。

AI rivalry shifts from smart models to smart factories

For years, the global discussion about artificial intelligence revolved around large language models, algorithms and computing power. The competition appeared straightforward: Whoever built the most powerful neural network would dominate the digital economy. That perception fueled a fierce technological race among leading US companies such as OpenAI, Google, Microsoft, Amazon and Anthropic. In China, companies such as Deep-Seek, Alibaba, Tencent, Huawei and Baidu responded by accelerating investment in large models, chips and cloud infrastructure.

But over the past year, a deeper transformation has become evident. AI is steadily moving beyond chatbots, search engines and office tools into manufacturing, logistics, pharmaceuticals, automotive engineering, robotics and industrial planning. Consequently, the focus of the global AI race has changed. The next phase of the competition may depend less on algorithms and more on productive capacity, industrial organization and the ability to embed AI into large-scale economic activity.

This shift has changed the very nature of AI rivalry. The first phase of the AI race primarily featured frontier models and computing supremacy. Companies competed for elite engineers, semiconductors, cloud infrastructure and venture capital. The goal was technological leadership.

But industrial deployment is a far more complex challenge. It requires coordination among software developers, manufacturers, supply chain operators, universities, industrial engineers, energy systems and governments. In other words, the future of AI increasingly depends on ecosystems rather than isolated technological breakthroughs.

This is precisely why strategic alliances are proliferating across the AI sector.

In the United States, collaboration between technology firms and industrial corporations has gathered momentum. Microsoft has expanded its partnership with the multinational automaker Stellantis to develop AI applications in engineering, cybersecurity and manufacturing systems. Google Cloud has deepened cooperation with the German pharmaceutical company Merck to deploy AI tools across the entire drug-discovery pipeline, while also broadening its collaboration with Mars to integrate AI agents into operational and industrial processes.

Meanwhile, Nvidia has evolved from just a chip supplier to the infrastructure backbone of global AI deployment, working simultaneously with cloud providers, manufacturers, robotics developers and national technology ecosystems. US companies are increasingly realizing that industrial integration — and not just model quality — will define long-term leadership.

At the same time, universities have become central actors in the emerging AI-industrial landscape. OpenAI's NextGenAI consortium brings together leading universities and research institutes in the US and Britain, offering grants, computing power and API access to researchers and students. Anthropic has established similar partnerships with academic institutions and educational organizations focused on computer science and AI engineering.

These collaborations are not just philanthropic initiatives. They help cultivate future talent, accelerate applied research, create industry standards, and familiarize future engineers with specific AI ecosystems. In effect, universities have become part of national AI supply chains.

China is pursuing a particularly distinctive path because it enjoys a structural advantage that many Western countries can't easily replicate: a comprehensive industrial ecosystem capable of integrating AI directly into manufacturing and production.

Chinese companies are not focusing solely on consumer-facing AI applications. Instead, they are integrating AI into factories, robotics, logistics systems, semiconductors, telecommunications infrastructure, electric vehicles and industrial automation.

Huawei is a clear example. The company has expanded cooperation with manufacturers, local governments, telecommunications providers and leading universities to accelerate AI deployment across industrial sectors. Partnerships with institutions such as Tsinghua University and Peking University have contributed to advances in chip design, distributed computing, industrial AI systems and optimization technologies.

Rather than treating AI as an isolated digital product, China regards it as infrastructure embedded across the economy.

DeepSeek illustrates another important trend. Its technical team includes specialists connected to China's leading universities, particularly in mathematics, engineering and computer science. This close integration has reportedly enabled the company to optimize training efficiency, reduce costs and maintain competitive performance, demonstrating how academic expertise can directly enhance industrial competitiveness.

Meanwhile, Alibaba Cloud and Tencent are accelerating AI deployment across retail, logistics, finance and smart-city ecosystems.

Chinese electric vehicle manufacturers are integrating AI into production lines, battery systems, autonomous driving and supply chain management.

In sectors ranging from robotics to advanced manufacturing, collaboration between AI developers and industrial enterprises is becoming increasingly institutionalized.

This evolution matters because the economics of AI are changing.

The early AI race rewarded companies capable of building the largest general-purpose models. But these large models are very expensive to train and maintain. Businesses are discovering that smaller, specialized AI systems integrated into industrial workflows generate greater economic value than universal chatbots.

As a result, the strategic advantage may shift toward countries that can deploy AI across large-scale manufacturing ecosystems rather than simply produce the most advanced algorithms.

This would favor economies with deep industrial capacity, integrated supply chains, strong engineering traditions and close coordination between universities, industry and state institutions. AI has become inseparable from energy infrastructure, semiconductor production, robotics, industrial automation and logistics networks.

In many ways, the emerging AI competition has begun to resemble the historical competition among industrial powers. The next phase of global AI rivalry may not be decided primarily in Silicon Valley or by benchmark rankings. Instead, it may unfold inside factories, ports, laboratories, industrial parks and supply chains. In that sense, AI is no longer just a software revolution. It has become a new industrial revolution.