Weekly #269 From technology to the infrastructure stack
From technology to the infrastructure stack
For years, digital infrastructure was largely equated with connectivity: more fibre, faster broadband, and wider mobile coverage. But recent developments, including the ones from last week, show a broader shift in how governments think about infrastructure. It is no longer just about the networks that move data, but about the entire technological stack that enables the digital economy and society, from semiconductor supply chains and cloud platforms to satellite systems and quantum technologies.
Chips, cloud, connectivity, quantum and satellites have graduated out of the category of ‘technology’ and into the category of ‘infrastructure,’ which is a much older and more heavily governed category. It’s the same institutional maturing that electricity went through a century ago and highways went through half a century ago: useful and largely unregulated at first, then gradually recognised as too foundational to the rest of the economy to leave ungoverned.
Canada and Germany’s new partnership on semiconductor supply chains reflects this evolution. Rather than focusing narrowly on access to chips, the agreement spans research, industrial development, investment, and innovation, signalling that semiconductor ecosystems are now seen as strategic infrastructure in their own right.
Singapore is taking a similar view of cloud infrastructure. Its proposed Digital Infrastructure Bill would bring cloud providers and data centres under a dedicated regulatory framework, recognising that the resilience of cloud services has become a matter of national interest rather than a purely commercial concern.
Some are already looking even further ahead. Türkiye’s first national quantum roadmap marks an early investment in technologies that may underpin future computing and communications, demonstrating that infrastructure planning is beginning to extend well beyond today’s commercially mature technologies.
At the same time, several governments are now explicitly building compute infrastructure as a national asset class. South Korea has announced a comprehensive AI infrastructure strategy centred on expanding high-performance computing capacity, including the procurement of 18,000 GPUs through a public-private national AI computing centre and a new national supercomputer. The strategy also couples infrastructure with capability-building: flagship LLM development, research access to compute and datasets, and talent pipelines through international AI labs and industry-university graduate programmes. Crucially, it extends infrastructure logic into deployment, mandating the rollout of domestically developed AI systems across core public services, from healthcare and education to disaster management and public administration.
Why does it matter? Digital infrastructure has always been about more than physical assets. Today’s policy agenda just makes that increasingly explicit. As governments rethink the foundations of their digital economies and societies, infrastructure is evolving from a collection of individual projects into a strategic architecture.
The EU. The Council of the European Union has given its final approval to a regulation that simplifies parts of the EU’s AI framework as part of the broader ‘Omnibus VII’ package to reduce regulatory complexity. The updated rules revise the implementation timeline for high-risk AI systems, postponing full application until December 2027 for standalone systems and August 2028 for AI systems integrated into regulated products. The regulation also strengthens safeguards by explicitly prohibiting AI-generated non-consensual sexual content, including manipulated intimate imagery and AI-generated child sexual abuse material. Additional changes aim to reduce administrative burdens and improve legal clarity. Deadlines for establishing AI regulatory sandboxes have been extended to August 2027, transparency obligations for AI-generated content have been streamlined, and the regulation clarifies the division of responsibilities between EU and national authorities while reducing overlap with sector-specific legislation. The framework also clarifies the division of responsibilities between EU and national authorities and introduces mechanisms to avoid overlap with sector-specific legislation.
Malaysia. Malaysia has launched the Malaysia Digital Action Plan 2030 (MD2030), a national roadmap that places the Ministry of Digital at the centre of efforts to achieve the country’s ambition of becoming an AI-driven nation by 2030. The strategy aims to transform Malaysia from a consumer of technology into a producer of homegrown digital innovation through a coordinated, whole-of-government approach. The five-year plan sets national priorities across economic growth, digital public services, infrastructure, talent development, cybersecurity and AI innovation.
Japan. Japan’s industry ministry has approved ¥387.3 billion in funding for a domestic AI project to develop a multimodal foundation model for physical AI systems that control robots. The project aims to develop a foundation model that can be widely adopted by Japanese companies to support industrial automation and robotics. Officials see the initiative as a strategic effort to narrow the technology gap with the USA and China in next-generation AI.
The UK-Germany. The UK and Germany have agreed to strengthen cooperation on AI safety and security. The partners will deepen institutional cooperation by sharing best practices in AI evaluation, aligning research priorities and exchanging expertise. The collaboration will also examine the cybersecurity implications of advanced AI systems and contribute to the international evidence base on AI safety.
South Korea-Japan. South Korea and Japan have agreed to expand defence cooperation, following talks between South Korean Defence Minister Ahn Gyu-Back and Japanese Defence Minister Shinjiro Koizumi in Seoul. The agreement included a commitment to strengthen ties in state-of-the-art science and technology, including AI.
UNDP. UN Development Programme (UNDP) and the UN Department of Political and Peacebuilding Affairs (DPPA) have published a technical guide on AI in electoral administration. The publication, From Promise to Practice: AI in Electoral Administration, doesn’t endorse the use of AI for electoral processes; It provides a framework to help electoral authorities assess whether, where and how AI can be adopted responsibly.
Pax Silica members meet for second annual summit
Last Friday (26 June), the Second Pax Silica Summit, hosted by the USA, concluded. The initiative brings together a growing coalition of 24 countries and economies committed to building secure, resilient, and innovation-friendly AI ecosystems.
At the centre of the summit was a Joint Statement on AI Opportunity, signed by nearly three dozen partners including the USA, the EU, India, Japan, the UK and a wide range of economies across Latin America, the Middle East, and Asia. The statement signals a shared preference for a pro-growth and innovation-driven approach to AI governance, emphasising support for startups, developers, and private-sector dynamism, while also strengthening trust and security across global AI supply chains.
The summit also moved beyond principles into implementation. A pilot project in Panama will test an AI-enabled logistics and customs platform designed to track and verify sensitive goods such as semiconductors, critical minerals, and AI infrastructure components. The aim is to speed up trade for trusted shipments while improving oversight of strategic technologies. If successful, the system could be expanded across other Pax Silica members.
In parallel, the USA and Stanford University launched the “Foundry School” initiative, an education and workforce programme focused on advanced manufacturing. It combines executive-level seminars with a transferable curriculum designed to help partner countries build domestic expertise in industrial systems that underpin technological competitiveness.
Why does it matter? Having expanded by 10 members at its second summit, Pax Silica points to the gradual formation of a technology-aligned geopolitical bloc, where influence is exercised less through traditional diplomacy and more through control over industrial ecosystems, trade infrastructure, and standards-setting coalitions.
Singapore updates its cybersecurity playbook
Singapore has unveiled two initiatives that strengthen its approach to governing digital risks across multiple fronts.
Singapore’s Online Safety Commission has begun operations, giving victims of online harms a dedicated channel to seek faster support and redress. In its first phase, the commission will support victims affected by five categories of online harm: online harassment, including online sexual harassment, doxxing, online stalking, intimate image abuse and image-based child abuse.
Victims of online harassment and online stalking are generally expected to report harmful content to the relevant platform first. If the platform fails to respond promptly or provides an inadequate response within 24 hours, the platform may be reported to the commission. More serious harms, including doxxing and image-based abuse, can be reported directly.
Where there is reason to suspect that online harm has occurred, the Commissioner may issue directions to the person who posted the content, the administrator of the online space or the platform hosting it. These directions may require access to harmful content to be disabled or an account to be restricted. Non-compliance is a criminal offence.
Singapore is also introducing court-based remedies through statutory torts. Victims may bring civil claims against communicators, administrators, or platforms that fail to meet the duties set out in the law. For intimate image abuse and image-based child abuse, courts must award at least $5,000 for each image or recording if the claim succeeds.
The commission will also work with community partners that can provide counselling and practical support to victims and families.
Separately, Singapore’s Cyber Security Agency (CSA) has outlined a broad package of initiatives to strengthen cybersecurity as AI reshapes the threat landscape. According to the Singapore Cyber Landscape 2025/2026 report, AI is enabling faster and more sophisticated cyberattacks, with agentic AI capable of automating parts of the cyber kill chain and accelerating vulnerability research, exploit development, and attack preparation. At the same time, AI is improving cyber defence by enhancing threat detection, incident response, and vulnerability management, making the secure deployment of AI systems an increasing priority.
To support this, CSA has published guidelines for securing AI systems, developed guidance on agentic AI security, and is working with international partners on AI security standards. The agency also warned that AI is making phishing and scams more convincing through voice cloning, deepfakes, and personalised attacks, prompting initiatives such as a national simulated scam exercise to improve public awareness.
Beyond AI, the report highlights a sharp rise in infected infrastructure, continued ransomware threats—particularly for SMEs—and growing risks from insecure consumer IoT devices. Singapore is responding by tightening cybersecurity requirements for residential routers and critical infrastructure operators, supporting SMEs through advisory and recovery programmes, advancing its National Quantum-Safe initiative, and expanding cybersecurity capabilities through exercises, partnerships, and new certification requirements. Together, the measures reflect Singapore’s effort to prepare for a threat environment which is increasingly shaped by AI, critical infrastructure risks, and future quantum technologies.
Why does it matter? As AI accelerates both online harms and cyber threats, policymakers are complementing regulation with new institutions, operational capabilities, technical standards, and support mechanisms.
CJEU court upholds €4.7 billion fine against Google in Android case
The European Commission imposed the record fine in 2018 after finding that Google had abused its dominant position by requiring smartphone manufacturers to pre-install Google apps and imposing contractual restrictions that limited competition. The ECJ dismissed Google’s appeal, confirming the Commission’s findings.
A lower EU court reduced the penalty from €4.34 billion to €4.1 billion in 2022 while largely upholding the Commission’s decision. Google has argued that Android increases consumer choice and supports developers, and said it modified its contractual arrangements following the original ruling to comply with EU competition rules.
The judgement comes as the EU continues to intensify scrutiny of major technology companies through both competition law and the Digital Markets Act, alongside investigations into digital advertising and other platform practices.
Why does it matter? The ruling reinforces the EU’s long-standing approach to digital competition, confirming that dominant technology companies can face substantial penalties when their market position is used to restrict consumer choice or limit competition. It also strengthens the Commission’s record in pursuing landmark antitrust cases against major digital platforms.
EU introduces €3 duty on low-value e-commerce imports
The European Commission introduced a temporary €3 customs duty on low-value goods imported into the EU from outside the bloc, from Wednesday, 1 July 2026. The duty applies to a wide range of commonly purchased goods, including clothing, toys, and electronics, covering items worth up to €150.
The duty is charged per customs tariff classification rather than by quantity. For example, purchasing five T-shirts attracts a single €3 charge because they share the same tariff code, whereas buying three T-shirts and a watch incurs two €3 charges because they fall under different classifications. Sellers or importers will declare and pay the duty through the customs process.
Why does it matter? The measure is intended to create fairer competition for the EU businesses, improve consumer protection by strengthening oversight of imported goods, reduce customs fraud linked to undervaluation and false declarations, and address the environmental impact of growing volumes of low-value shipments. It is also widely seen as a measure for discouraging shopping via low-cost e-commerce platforms.
Mauritius unveils fintech strategy to boost digital finance growth
The roadmap focuses on six areas: regulation and innovation, digital infrastructure and cybersecurity, skills development, market growth, international cooperation and consumer protection.
Implementation will run until 2030 and will be overseen through a dedicated governance framework. Planned targets include shorter licensing approval times, expanded digital onboarding, stronger digital infrastructure and training more than 5,000 people each year in specialised fintech skills.
The government says the strategy is intended to position Mauritius as Africa’s trusted fintech hub while supporting sustainable growth and the wider digital transformation of financial services.
Why does it matter? Mauritius’ strategy reflects a wider African policy trend: governments are trying to move fintech from fragmented innovation into structured national development plans.
The 62nd session of the HRC continues
The 62nd session of the UN Human Rights Council (HRC) is ongoing until 7 July. Last Friday, Peggy Hicks, Director, Thematic Engagement, Special Procedures and Right to Development Division of the Office of the United Nations High Commissioner for Human Rights, presented seven reports to the Human Rights Council. Two of them are related to digital technology.
The Report of the UN High Commissioner for Human Rights on Using digital tools for stakeholder engagement: aligning corporate practice with the Guiding Principles on Business and Human Rights (A/HRC/62/25) examines how companies are increasingly using digital tools to engage with workers, communities, and other affected stakeholders in the context of the UN Guiding Principles on Business and Human Rights. It finds that while these tools can improve reach and responsiveness, they can also deepen exclusion, weaken trust, and raise privacy concerns when poorly designed or used without safeguards. Importantly, digital engagement is not automatically meaningful—its value depends on whether it is appropriate, accessible, and rights-respecting for the people involved.
The report stresses that digital tools should complement, not replace, direct engagement with stakeholders such as workers, unions, and communities. It provides guidance for companies, technology providers, regulators, and civil society on assessing whether such tools are suitable, safe, and genuinely contribute to meaningful participation. Ultimately, it argues that digital engagement is only as rights-respecting as the design choices and governance frameworks behind it.
The Study on States’ obligations under international human rights law, relevant norms and commitments, and the human rights responsibilities of business enterprises in line with the Guiding Principles on Business and Human Rights, across the lifecycle of new and emerging digital technologies (A/HRC/62/33) frames human rights due diligence as a continuous process to identifying, preventing, mitigating, and remedying harms, triggered by actual violations as well as foreseeable risks. States are expected both to avoid rights violations in their own use of AI and to regulate private actors through robust legal frameworks, oversight, and, where necessary, restrictions on high-risk applications.
UN scientific panel publishes first global AI assessment
The UN Independent International Scientific Panel on Artificial Intelligence has published its first preliminary report, providing an evidence-based assessment of AI’s opportunities, risks, and societal impacts.
The report describes AI as a general-purpose technology capable of boosting productivity, innovation, and access to services. However, it stresses that these benefits are unevenly distributed, with significant gaps in access to computing power, talent, infrastructure, and data. Developing countries risk becoming consumers rather than producers of AI unless they invest in local ecosystems, digital skills, and research capacity, including multilingual systems and locally relevant datasets.
Alongside opportunities, the report identifies growing risks, including AI-enabled cyberattacks, fraud, disinformation, election interference, and market concentration. It also highlights the lack of technical capacity in many governments to independently assess frontier AI systems. Environmental impacts are also flagged, particularly rising energy, water, and hardware demands linked to large-scale AI deployment.
The Global Dialogue on AI Governance (AI Dialogue) will hold its inaugural meeting on 6-7 July 2026. The programme of the Global Dialogue will include high-level governmental plenary segments, multistakeholder exchanges, and the presentation of the report of the multidisciplinary Independent International Scientific Panel on Artificial Intelligence, thematic discussions across four clusters, and side events.
The 2026 edition of WSIS Forum will convene representatives from governments, international organisations, civil society, the private sector, academia, and technical communities to exchange knowledge, highlight innovative solutions, and engage in multistakeholder dialogue to advance the implementation of WSIS Action Lines and contribute to broader global development goals.
The AI for Good Global Summit 2026 starts on Tuesday, 7 July, in Geneva. The summit will focus on exploring and advancing the practical application of AI to address global challenges and support the SDGs. The programme includes keynote speeches, panels, exhibitions, workshops, and related initiatives to identify scalable AI applications that contribute to sustainable development and societal benefit. The event also offers opportunities for networking, knowledge exchange, and stakeholder engagement across the global AI ecosystem.
The High-level Political Forum on Sustainable Development 2026 (HLPF 2026) will convene from 7 to 16 July. Discussions will place digital technologies, including AI, at the centre of SDG 9 and broader development strategies, highlighting their role in transforming industry, innovation systems, and global value chains. At the same time, they stress persistent digital divides, with uneven access to connectivity, computing capacity, and digital infrastructure—particularly in LDCs, LLDCs and SIDS
The International Advisory Body for Submarine Cable Resilience will hold its final meeting on 10 July (Friday). During the meeting, IAB Members will review and formally approve the final publication incorporating the reports and recommendations of the three Working Groups.
In our June 2026 issue, we examine Geneva’s role as a central hub for global AI governance, where diplomacy, technical standards, and human-centred approaches increasingly converge. We also look at how Anthropic’s brief suspension and reinstatement of Fable 5 reflects shifting US export controls and increasingly reactive AI regulation. We explore Africa’s efforts to strengthen its voice in global digital diplomacy amid coordination gaps, alongside the European Commission’s Technological Sovereignty Package aimed at boosting capabilities in AI, semiconductors, and cloud infrastructure. We also introduce a new structured snapshot of AI developments across policy, industry, and governance.
When AI learns from humanity, it inherits more than knowledge. It also absorbs the biases, assumptions, and inequalities embedded in society.
Follow the road to the 2027 Geneva AI Summit, a global journey shaping AI’s future. Bookmark the page and register for updates to stay informed and involved every step of the way to Geneva 2027.
