When the Chinese State Council issued its “AI Plus” guideline on August 26, 2025, it set a target that no other major economy has dared to publish: a 90% penetration rate for smart terminals and next-generation AI agents by 2030, and already 70% by 2027. The document presents AI as the next general-purpose technology, comparable to electricity, and identifies six sectors in which it must dominate: science and technology, industry, consumption, public welfare, governance and global cooperation.
By 2030, China expects AI to be the main driver of growth; by 2035, a “smart economy and society” sufficient to support what Xi Jinping calls socialist modernization. Analysis of Trivium China noted that no Chinese policy, including the 2025 Guideline, has defined a unit of measurement for AI’s contribution to GDP, but the policy signal is clear: AI is now on par with the 2015 “Internet Plus” initiative that gave birth to Meituan, DiDi and the cashless economy. Beijing is ready to legislate AI adoption, not just subsidize it.
The implications for Europe, which is still debating compliance, are unflattering. And China is not the only Asian country getting into AI.
South Korea has gone furthest in formalizing its approach. The orientation law for the development of artificial intelligence, known as the law Basic Law of AIcame into force on January 22, 2026, making South Korea the first country to adopt a comprehensive AI statute combining governance, industrial policy and risk management in a single law.
The 2026 national budget AI spending tripled to 10.1 trillion won (about $6.94 billion), and President Lee Jae-myung has pledged to make South Korea an “AI G3” power by 2030. Nvidia’s Jensen Huang announced in October 2025 that 260,000 Blackwell GPUs will be deployed on Samsung, SK, Hyundai, Naver and government infrastructures. South Korea chose an EU-recognized path (regulate first), but associated it with a capital deployment that the EU did not follow.
Japan presents the most interesting contrast. The use of generative AI among the Japanese public stood at 26.7% in 2024, compared to 68.8% in the United States and 81.2% in China, according to the Ministry of the Interior and Communications. White Paper 2025. The Japan Times reported in December 2025 that Tokyo’s proposed AI core program aims for a public use rate of 50%, or even 80%, as well as about 1 trillion yen (about $6.4 billion) in private R&D investment. NTT alone is committing $59 billion through 2027, and SoftBank has tied up with OpenAI. Stargate Project with more than $40 billion in commitments. Japan is behind and knows it – now Tokyo is making amends.
India plays a different role: computing as a public service. THE IndiaAI Missionapproved in March 2024 with an expenditure of more than 103.7 million rupees ($1.24 billion), has already been deployed on 38,000 GPUs (including 1,050 Google Trillium TPUs), well beyond the initial target of 10,000. Startups and academics can access H100-class compute at 65 rupees (about $0.72) per hour, the cheapest subsidized rate in the world.
Minister Ashwini Vaishnaw said in February 2026 that investments in AI in India already stand at $90 billion, and are expected to exceed $400 billion in the AI sector within two years. The Indian government predicts that AI will add $1.7 trillion to GDP by 2035.
Singapore is the case for precision. It is National AI 2.0 Strategy (launched in 2023, with an updated 10-priority update in May 2026) tripled the AI talent pool target from 4,500 to 15,000 and anchored more than 70 enterprise AI centers of excellence on the island. According to IMDA’s Singapore Digital Economy Report, AI adoption among small and medium enterprises (SMEs) increased from 4.2% in 2023 to 14.5% in 2024, and adoption by non-SMEs increased from 44% to 62.5% in the same year. The new National AI Impact Program aims to integrate 10,000 companies over three years.
Singapore is small enough to measure what it deploys, which is precisely why it outpaces larger economies in per capita adoption. The table below summarizes where the six countries stand in terms of targets, current adoption and capital commitments.
In the face of this, the EU’s digital decade seems lukewarm. The main target is for 75% of European businesses to use cloud, AI or big data by 2030. Eurostat data for 2025 puts cloud adoption at 39%, data analytics at 33.3% and AI at 13.5% in EU businesses. Basic digital skills stand at 55.6 percent among adults, compared to a target of 80 percent.
That of the Commission in June 2025 State of the digital decade The report concedes that at the current pace, the bloc will not achieve the 2030 goals until around 2040. An analysis commissioned by AWS by The public first found that the EU is on track to unlock just €1.3 trillion in projected digital value by 2030, leaving up to €1.5 trillion on the table. The EU regulates an economy that it has not yet built.
The EU has world-class regulation (the AI Act), serious chip ambitions (20% of global production value by 2030, up from 10.5% in 2024) and almost nothing on the demand side that compares to what Asian governments are imposing on their economies. China has publicly set a 90 percent penetration target and history suggests that Beijing will use its industrial policy to achieve it. South Korea legislated a national AI architecture and tripled its budget in a single year. India treats GPUs like the public highway. Singapore measures business adoption and acts on the result. Japan, which is embarrassingly behind in utilization rates, is spending tens of billions on this deficit.
Brussels is confusing the rules of the game with the rules of the game. If the digital decade ends in 2030 with the EU’s AI adoption rate at 30 to 40 percent while China sits at 90 percent, no well-drafted compliance measure will close that gap, and the productivity dividends will accrue elsewhere.