Governments are walking a fine line between the promotion of the AI ecosystem and its regulation. As the capabilities of AI advance faster than any previous general-purpose technology, governments face an intricate set of policy decisions, spanning across several policy areas and the entire AI value chain. This briefing details how governments are aiming to promote and regulate AI, and how competition policy is relevant to both. 

AI regulation is an iceberg, comprising both AI-specific policies and cross-cutting policies that affect the AI ecosystem. Debates on AI regulation focus mostly on AI-specific frameworks, such as the European Union’s AI Act. The Digital Policy Alert has documented 1243 AI-specific policy and enforcement developments, most of which occurred after the launch of ChatGPT. Recent examples include the adoption of the United States’ AI Action Plan (see below) and the BRICS declaration on AI governance. Beyond AI-specific developments, the Digital Policy Alert database contains an even larger number of cross-cutting developments (3251). These developments target the entire digital economy, not only to AI, but are equally important for the AI ecosystem. Data protection and intellectual property rules, for instance, apply to all sectors of the digital economy but considerably influence the training and output of AI systems. These cross-cutting policies represent the large, hidden, part of the AI regulation iceberg. 

Competition policy showcases the dimensions of the AI regulation iceberg. The Digital Policy Alert has documented 978 competition policy and enforcement developments across the globe. Notably, only 32 of these developments are AI-specific. They comprise mostly enforcement cases against AI firms, including the European Commission’s concluded investigation into the proposed acquisition of Run:ai Labs by Nvidia and the United Kingdom’s Competition and Market Authority’s concluded investigation into the partnership between Microsoft and OpenAI. In addition, singular international declarations, such as the G7 Digital Competition Communiqué on AI address competition with an AI-specific perspective. The majority of competition policies that affect AI are thus not AI-specific, but cross-cutting: 266 cross-cutting competition developments lie “under the surface”. 

Cross-cutting competition policies affect the AI ecosystem through different obligations and new enforcement powers. The Digital Policy Alert dataset reveals that cross-cutting competition developments are two-fold. A majority of the developments impose obligations, specifically regarding unilateral conduct (85), merger control (74), and anti-competitive agreements (25). A minority of the developments (71) focuses on empowering competition authorities to address the challenges of digital market competition, mainly through new enforcement tools. Both types of policies impact the AI ecosystem despite not being tailored to it. Rules restrict how AI firms can use their market power and transact with other firms in the AI market. Novel powers enable competition authorities to effectively reign in anti-competitive behaviour in the AI market.

AI promotion comprises an even broader variety of policy areas, since it spans across the AI value chain. While AI regulation focuses on the “application layer”, AI promotion also covers the hardware underlying AI applications. This includes energy, infrastructure such as data centres, and semiconductors. Governments aiming to promote AI can thus make use of various trade and industrial policy instruments, to both boost and protect the local AI ecosystem. Subsidies for (clean) energy, for instance, aim to boost the AI ecosystem. Export restrictions on semiconductors, on the other hand, aim to protect the local AI ecosystem from foreign competition. Both types of AI promotion are rising, as documented by the Global Trade Alert, signaling a pivot from open markets to strategic interventionism. 

Some governments are doubling down on AI promotion and slowing down on AI regulation. The United States exemplifies this shift. The new administration promptly rescinded the Biden-era Executive Order on Safe, Secure, and Trustworthy Development and Use of AI, which aimed to set guardrails for AI. In July 2025, the government adopted the AI Action Plan with three pillars: accelerating AI innovation, building American AI infrastructure, and leading international AI diplomacy and security. Of the three Executive Orders that implement the Plan, two focus on AI promotion. Specifically, their goals are accelerating federal permitting of data centre infrastructure and promoting the export of the American AI technology stack. In line with the goal of exporting AI technology, the new administration had previously rolled back certain export controls on semiconductors. Specifically, the AI Diffusion Rule and a licensing regime for the export of certain chips, such as Nvidia’s H20 and AMD’s MI308, were removed. Export controls on the most advanced semiconductors, however, remain in place. As the global AI race intensifies, other governments may well follow the United States’ example and increase AI promotion efforts – raising AI competition concerns of another kind.

India has not enacted a dedicated AI law, but applies both AI-specific and cross-cutting rules to AI. In 2024, the Economic Advisory Council to the Prime Minister proposed a complex adaptive systems approach to regulating AI, advocating for effective regulation through five principles. More recently, the Ministry of Electronics and Information Technology (MeitY) held a consultation on AI governance guidelines, while the Securities and Exchange Board consulted on guidelines for responsible AI use in security markets. Cross-cutting frameworks that significantly affect the AI ecosystem include the Digital Personal Data Protection Act and the Information Technology Act/Rules. MeitY previously issued advisories for intermediaries under the Information Technology rules, specifically on AI due diligence, preventing misinformation, and deepfakes.

In terms of AI promotion, the 2018 National Strategy on AI advanced an inclusive "AI for all" vision, focusing on overcoming barriers related to accessibility, cost, and the availability of skilled talent. In 2021, the IndiaAI Mission was launched with a budget of INR 10,371.92 crore, to foster the AI ecosystem by improving access to compute, data, and capabilities, among others. To implement the Mission, MeitY recently launched AIKosha, which provides an integrated development environment with over 300 datasets, 80 AI models, AI sandbox capabilities, and tools. Additionally, IndiaAI Compute offers relevant compute, network, storage, platform, and cloud services, initially providing 10,000 GPUs, with 8,693 more planned. Furthermore, the AI Competency Framework provides training and skill development for public sector officials in AI-related fields. More broadly, India’s Digital Public Infrastructure aims to provide a foundation for a thriving AI ecosystem. 

India also strives for international cooperation on AI. India participated in the recent BRICS declaration on AI governance and the declarations issued during the Bletchley AI Safety Summit and the Paris AI Action Summit. India also cooperates on AI bilaterally, including in its partnership with Indonesia, its Trade and Technology Council with the European Union, and its free trade agreement with the United Kingdom, which includes a commitment to cooperate on emerging technologies, including AI.