Government AI Readiness Index 2025

The global AI leadership picture is increasingly bipolar, with the United States and China emerging as the two dominant forces shaping the technology’s trajectory. While the US retains the top position in our rankings—buoyed by its private sector ecosystem, research institutions, and compute infrastructure—China’s position in the Index likely understates its true capabilities. Measuring China’s AI readiness from outside presents significant methodological challenges: limited transparency around government initiatives, restricted access to deployment data, and a domestic technology ecosystem that operates largely independently of Western platforms and standards.

What we can observe, however, is increasingly compelling. China’s AI research output in 2024 matched the combined publications of the US, UK and EU, with 156 institutions each producing over 50 AI papers—a distributed research infrastructure unmatched elsewhere. Models like DeepSeek and Alibaba’s Qwen have demonstrated performance rivalling Western counterparts at a fraction of the development cost, challenging assumptions about the necessity of cutting-edge hardware.

Meanwhile, China is systematically building a sovereign AI stack: Huawei’s Ascend chip ecosystem and SMIC’s manufacturing advances are reducing reliance on Western semiconductors, backed by over US$140 billion in state investment through the National Integrated Circuit Industry Investment Fund. Control of critical supply chains—including 50% of global silicon carbide wafer production and dominance in rare earth processing—further insulates this infrastructure. More broadly, governance initiatives like the Shanghai Declaration on Global AI Governance signal Beijing’s intent to shape international AI norms, particularly among Global South nations.

China’s rise in our rankings reflects this growing evidence base, yet we assess that a complete picture would likely place it higher still. The gap between the US and China is almost certainly narrower than the rankings suggest—and in specific domains, China may already lead.

2025 saw governments, and countries, increasingly putting AI strategies and policies into action. With countries across the Index exploring AI within government, and across their societies and economies. From the release of Canada’s AI Strategy for the Federal Public Service (Canada 2025 Index ranking: 12th), to the launch of Nigeria’s AI Scaling Hub (70nd). The important role of AI as a tool within the public sector toolkit is also being recognised. With more and more countries now clearly setting out in public documents how they are – or intend to – use AI to improve the reach, efficiency, and effectiveness of government.

Governments are also looking outward, and identifying how they can also drive – and catalyse – their national AI innovation ecosystems. Countries are exploring the role of AI sandboxes and other innovative policy instruments, and also working more closely with the private sector – including Uzbekistan’s launch of the ‘President AI Award’, to support new innovators and entrepreneurs. Whilst many are also engaging internationally in order to strengthen their domestic talent and opportunity. For example, the International Center for AI Research and Ethics (ICAIRE), under the auspices of UNESCO.

In engaging further with their respective ecosystems, countries are also considering areas of particular AI opportunity – or even comparative advantage. Australia published its final report exploring safe and responsible AI in health care. Whilst Estonia launched its AI Leap initiative – weaving AI throughout the education system, and inspired by its 1990s ‘Tiger Leap’ programme that has positioned the country as a global digital leader. And Rwanda’s hosting of the Global AI Summit on Africa positions the country as a key AI actor on the continent.

Finally, as the countries continually ranking highly in the Index demonstrate, success in AI is not an overnight phenomenon. It requires deep, focused, and strategic engagement with shaping the foundations, components, and catalysts that can lead to AI having a positive and sustained impact across governments, societies, and economies. However, this success demands longer-term direction and investment. The former is increasingly evident through the development and implementation of AI strategies. But the latter – particularly the existence and extent – of dedicated AI funding remains a gap in many settings.

The Index has always recognised, and championed, the importance of cross-border collaboration on AI. As a general-purpose technology – one with the potential to fundamentally reshape our societies and economies – learning from the successes, and challenges, of all contexts and communities will be essential in ensuring that AI has the greatest positive impact. However, although international collaborations on AI governance and R&D continue, 2025 also saw increased attention centred around defining or driving national and regional AI sovereignty.

AI sovereignty is being explored through a number of separate approaches – although with considerable crossover in thinking, approach, and intended impact. Geopolitics, and jostling for supremacy in the AI ‘race’, are a driver in some countries. The July 2025 US Presidential Executive Order explicitly highlighted the importance of promoting the export of the American ‘AI technology stack’ as a mechanism to ‘secure [US] continued technological dominance’. Ranking 1st again in this year’s Index, the US continues to shape global AI discourse. Whilst the country’s long tail of hardware, software, diplomacy, and talent could see this shaping become more tangible.

AI sovereignty is also resonating in other national contexts. China (ranked 8th, up from 23rd in 2024) continues to strengthen its technology and ecosystem components. And, building on its earlier leadership in shaping a global approach to ‘Digital Public Infrastructure’, the IndiaAI Mission is a strategic play to shape an extensive domestic AI ecosystem. The India AI Summit, in February 2026, could see some of these efforts be recognised – and potentially incorporated – by other countries. And nuanced approaches could also be emerging. The Kingdom of Saudi Arabia – leveraging its sovereign wealth fund – launched ‘HUMAIN’, a commercial approach to shaping a ‘full-stack AI ecosystem’ for the country.

Finally, and perhaps in response to some of this direction, regions are exploring ways to strengthen their autonomy, agency, and self-reliance across the AI technology ‘stack’. August saw the launch of Latam-GPT – with an aim to develop a Large Language Model (LLM) entirely within the Latin America region. Whilst the European Union (EU) is also working on its own approach. Earlier in 2025, the EU launched the OpenEuroLLM initiative – a programme aiming to develop a family of open LLMs across the 24 official languages of the EU. The continent’s Apply AI Strategy will also be championing a ‘buy European’ approach, building on continued mentions of ‘technological sovereignty’ across policy documents and other bloc communications.

The international AI landscape is not solely about geopolitics, international rivalry, and discussions of zero-sum AI ‘progress’. AI is also emerging as a key pillar of regional discussions and direction. Perhaps in recognition of the cost and complexity of trying to build entire domestic approaches – particularly as AI technology, and its potential, continues to rapidly evolve. Cross-border AI efforts are also moving beyond abstract agreements and into action. With a cross-border toolkit of collaborative assets taking shape. From policy, to joint R&D efforts, and shaping practical regional norms and governance approaches.

AI was a central component of agendas at regional platforms throughout 2025. The African Union declared AI to be a ‘strategic priority’ for a dynamic continent with the world’s youngest population – whilst the EU saw tension during the Paris AI Summit, with the US and UK refusing to sign the event’s final declaration, but also progress as the bloc’s JUPITER exascale supercomputer came online. AI was also a central component of discussions at the COP30 climate conference in Brazil, and at the launch of the China-led Global Governance Initiative.

Bilateral AI collaborations are also emerging. The first-ever UK and US technology agreement saw the two nations committing to AI R&D collaborations in drug discovery, energy, and exploration. Malaysia and China agreed to technical exchange and mutual strengthening of AI infrastructure. Whilst Brazil explored both regional agreements – including with Ecuador, to increase AI skills development – and global partnerships, signing a Memorandum of Understanding with China to further shape AI collaboration between the two countries. This list is by no means exhaustive, but a strong signifier that national agreements continue at pace – but both interacting with and diverging from regional and global pacts.

2026 could prove to be another year of evolving connections, collaborations, and calculations. With a potential additional layer of geopolitical dynamics taking shape as the US hosts the G20, France – potentially advancing its Paris AI Summit efforts – takes on host duties of the G7, and India perhaps bringing a new perspective to the global AI landscape at the India AI Impact Summit. Whilst the Philippines, rising seven places in this year’s Index, will assume the Chair of the Association of Southeast Asian Nations (ASEAN) regional bloc, and could be an important voice in defining an ‘ASEAN way’ for the AI era.

Although regional and global collaborations continue at pace, practical aspects of global AI regulation are taking longer to come to fruition. Components of the EU AI Act entered into force throughout 2025, but the end of the year saw continued discussion and pushback concerning the extent, practicalities, and relevance of some of these provisions for national and regional AI development and direction. As a global bellwether for AI governance, 2026 will be an important year for the EU to drive a common direction across its member countries – and to send an important signal to its global partners, followers, and watchers.

Other global efforts are still emerging. At the 2025 United Nations General Assembly, the UN launched two new AI governance bodies: the Global Dialogue on AI Governance and the Independent International Scientific Panel on AI. The former will bring together experts and stakeholders from across all sectors, countries, and communities to share and explore governance approaches. Translating these national and broader learnings into wider practice will be key, including providing a voice to the many countries often not at the table when AI rules and norms are being determined.

And new AI governance divergence may also be taking shape. The US boycotted the 2025 G20 Summit in South Africa, an event which saw the establishment of the G20 Taskforce on Artificial Intelligence, Data Governance and Innovation for Sustainable Development. With the G20 represented by members of varying levels of AI readiness, the community remains an important actor in advancing global AI governance. The 2026 Summit, hosted by the US, could see a new narrative in this joint  AI action and direction – or reaffirm the geopolitical rivalries we saw continue throughout 2025.

Finally, although global regulatory and governance efforts are in flux, many countries are focusing on what they have within their control and are shaping domestic governance approaches. The Brazilian Senate has been closely following the EU AI Act, whilst many other countries around the world are shaping AI or broader sandboxes in order to define their governance approaches. Smaller gatherings could also be important in shaping a common direction. The 12-country Commonwealth Artificial Intelligence Consortium ‘minilateral’ could be an important platform to shape new AI thinking and approaches. Several of its members made good progress in this year’s rankings – including Tanzania (moving into the global top 100), Pakistan, and Jamaica.

2025 saw even more attention paid to the global trials, tribulations, and travails of the AI private sector. From new company acquisitions, to seven-figure signing salaries, and discussions of stock market floating. But, in amongst this talk of billion dollar valuations and million-dollar salaries, the essential role of governments in this AI era risks being diminished. As the Index shows, governments around the world are actively engaging with the role and potential of AI tools and technologies.

Governments are setting the rules of the game, and are also important – and increasingly successful – players in the AI landscape. From the work of Brazil in launching Latin America’s first AI sandbox, to Saudi Arabia’s public-private ‘HUMAIN’ full-stack AI collaboration, governments are pushing boundaries and encouraging the private sector to raise the bar. And governments are also essential in shaping and driving new AI opportunities. Fostering the development of national AI ecosystems.

The public sector is also a key catalyst for private sector success. And many governments are shaping long-term and formidable toolkits to strengthen domestic AI direction. From China’s launch of its ‘K Visa’ to attract AI and tech talent, to the significant digital, data, and AI infrastructure investments being made by India – and continued AI infrastructure investments and explorations in the European Union. Whilst cross-border collaborations could drive new AI opportunities that augment and amplify the international investments of the private sector.

AI developments and successes over the past decades have emerged from a Silicon Valley community that was heavily supported and nurtured by the US government. Identifying the role, importance, and value-add of governments in this next chapter of the AI story could be an important component in maximising the positive power and potential of these tools.