UNCTAD Technology and Innovation Report 2025: Inclusive AI for Development

The $16.4 Trillion Frontier Technology Opportunity

The UNCTAD Technology and Innovation Report 2025 arrives at a pivotal moment in global development. Frontier technologies — spanning artificial intelligence, the Internet of Things, blockchain, 5G, robotics, renewable energy systems, nanotechnology, and gene editing — represented a $2.5 trillion market in 2023. UNCTAD projects this figure will surge sixfold to $16.4 trillion by 2033, driven by a compound annual growth rate of approximately 20%.

Within this landscape, artificial intelligence stands as the dominant force. The AI market alone is projected to reach $4.8 trillion by 2033, accounting for nearly one-third of the total frontier technology market. The generative AI segment, catalyzed by the launch of ChatGPT in late 2022, is forecast to expand from $137 billion in 2024 to $900 billion by 2030 — a staggering 37% compound annual growth rate that underscores the speed at which this technology is reshaping industries and economies worldwide.

Yet the report’s central thesis is not one of uncritical techno-optimism. As UNCTAD Secretary-General Rebeca Grynspan writes in the foreword, “History has shown that while technological progress drives economic growth, it does not on its own ensure equitable income distribution or promote inclusive human development.” The report systematically demonstrates that without deliberate policy intervention, ethical oversight, and international cooperation, AI will deepen — not bridge — the divides between wealthy and developing nations. For organizations seeking to understand how digital transformation intersects with global trade dynamics, the UNCTAD Digital Economy Report 2025 provides essential complementary analysis.

AI as a General-Purpose Technology Reshaping Every Sector

UNCTAD characterizes AI as a general-purpose technology — a designation reserved for innovations with three defining attributes: pervasiveness across sectors, dynamicity in continuous improvement, and innovational complementarities that enhance other technologies. This classification places AI alongside historic transformations like the steam engine, electricity, and the internet.

The report traces AI’s evolution through three distinct waves. The first wave (1950s–1960s) focused on rule-based systems following Alan Turing’s foundational concepts. The second wave (1990s onward) embraced statistical learning and pattern recognition, exemplified by milestones like ImageNet in 2007 and AlphaGo’s defeat of the world Go champion in 2016. The current third wave — contextual adaptation — has unleashed generative AI capable of producing text, images, and video with remarkable sophistication.

What makes this moment transformative is AI’s capacity to amplify every other frontier technology. When combined with IoT, AI creates intelligent autonomous systems for smart factories and precision agriculture. Paired with 5G, it enables real-time vehicle communication and intelligent transportation networks. Integrated with blockchain, AI strengthens cybersecurity, fraud detection, and supply chain optimization. In healthcare, AI image classifiers are transforming diagnostic capabilities, while in manufacturing, smart sensors powered by AI enable real-time control of energy and water usage.

The scale of investment reflects this transformative potential. AI-related investment is expected to double to $200 billion between 2022 and 2025 — roughly three times global spending on climate change adaptation. By 2030, AI investment could represent 2% of GDP in leading AI countries. The training costs for frontier models are escalating at 2.4 times per year since 2016, with GPT-4 utilizing an estimated 1.77 trillion parameters compared to GPT-3’s 175 billion.

The Widening AI Divide Between Nations

Perhaps the report’s most urgent finding is the documentation of a widening AI divide across four critical dimensions: infrastructure, service provision, investment, and knowledge creation.

In infrastructure, the United States commands approximately one-third of the world’s top 500 supercomputers and more than 50% of total computational performance. China holds 80 supercomputers in the top 500 but delivers less than one-tenth of the US’s computational capacity. Beyond these two giants, few developing countries possess significant computing infrastructure — with Brazil, India, and Russia as notable exceptions. Data centers remain overwhelmingly concentrated in developed economies.

The investment gap is equally stark. In 2023, the United States attracted $67 billion in private AI investment — 70% of the global total. China followed with $7.8 billion, while India, the highest-ranked developing country outside China, secured only $1.4 billion. The US produced roughly seven times more newly funded AI companies than China. Among the world’s top 100 corporate R&D investors, approximately 49% are headquartered in the United States and 13% in China. Outside of China, no developing country hosts a single top-100 R&D investor.

Knowledge creation follows the same pattern. China and the United States together produce approximately one-third of global AI peer-reviewed articles and two-thirds of all AI patents. Roughly 50% of the world’s top-tier AI researchers originate from China, 18% from the US, and 12% from Europe. The report notes a troubling slowdown in technology diffusion, which means the distance between frontier countries and latecomers is growing rather than shrinking. As global trade dynamics shift in response to technological competition, developing nations risk being locked out of the AI value chain entirely.

Three Leverage Points: Infrastructure, Data, and Skills

UNCTAD organizes its analysis around three mutually reinforcing leverage points that determine a country’s AI readiness: infrastructure, data, and skills. These create what the report calls “transformational cascades” — positive feedback loops where progress in one area accelerates gains in the others.

Infrastructure encompasses computing power, server capabilities, storage, network connectivity, and security systems. Without robust digital infrastructure — including broadband access, cloud computing platforms, and data center capacity — countries cannot effectively deploy or develop AI systems. The report emphasizes that the rising costs of frontier AI training create increasingly high barriers to entry, concentrating capability among the wealthiest actors.

Data serves as the primary input for training, validating, and testing AI algorithms. The world is approaching 200 zettabytes of data generation by 2025, yet the quality, diversity, and accessibility of this data varies enormously across countries. Developing nations often lack the data governance frameworks, digital ID systems, and structured datasets necessary to train AI models relevant to their specific needs. Data and intellectual property policies in developed countries can further hinder technology development elsewhere.

Skills range from basic data literacy across populations to advanced expertise in developing and applying AI models. The report identifies a severe talent concentration problem: leading AI companies and research institutions in developed countries attract top researchers from developing nations, creating a brain drain that compounds the digital divide. Building domestic STEM education pipelines and AI literacy programs is therefore essential for any country seeking to participate meaningfully in the AI economy.

Countries like Brazil, China, India, and the Philippines are identified as developing nations that outperform their peers in technology readiness, offering models for how strategic focus on these three leverage points can accelerate progress despite resource constraints.

AI for Productivity and Worker Empowerment

The report delivers a nuanced analysis of AI’s impact on employment, finding that AI could affect approximately 40% of global employment. Unlike previous automation waves that targeted routine, low-skill functions, AI performs cognitive tasks — making its labor market disruption qualitatively different and potentially more far-reaching.

UNCTAD identifies a critical tension: corporate incentives tend to direct AI toward labor substitution rather than labor augmentation. If this pattern dominates, AI could erode the comparative advantage of low labor costs that has driven development in many emerging economies, “threatening much of the gains they have made in recent decades.” The report finds that technological advancements are already shifting value toward capital at the expense of labor.

However, the report also presents compelling evidence for a more optimistic trajectory. AI has the potential to augment worker capabilities, improve productivity, and potentially reverse the capital-shift trend — but only if countries adopt deliberate worker-centric approaches. This means considering workers throughout the entire AI life cycle, from development to deployment to monitoring, and ensuring that AI tools enhance rather than replace human roles.

Sector-specific applications in developing countries demonstrate this potential. In agriculture, AI-powered systems assist with pest detection, yield prediction, and precision irrigation — augmenting rather than replacing farmers. In healthcare, AI image classifiers extend diagnostic reach to underserved communities. In banking, AI predicts loan default rates, expanding financial inclusion. The Bain Technology Report 2025 explores how leading organizations are navigating this balance between automation and augmentation in practice.

National Policy Blueprints for Inclusive AI

Chapter IV of the report provides detailed policy guidance for countries at different stages of AI readiness. The overarching recommendation is that AI policies should be embedded within broader industrial and innovation strategies using a “whole-of-government approach” rather than being treated as standalone technology initiatives.

The report examines three distinct national approaches. China has adopted a state-led strategy with ambitious targets and centralized coordination. The European Union has taken a regulatory-forward approach, exemplified by the EU AI Act, which classifies AI systems by risk level and imposes corresponding obligations. The United States has pursued a market-driven model with targeted interventions.

For developing countries specifically, UNCTAD distills four critical takeaways from successful case studies. First, adapt AI solutions to local digital infrastructure limitations rather than attempting to replicate developed-country approaches. Second, utilize new and creative sources of data — including mobile phone data, satellite imagery, and community-generated datasets. Third, make AI easy to use by lowering skill barriers with simple interfaces. Fourth, build strategic partnerships across sectors and borders to access essential AI resources, research capacity, and expertise.

National AI strategies are the most common policy instrument globally, but the report finds a strong correlation between GDP per capita and AI governance preparedness. Most AI policies have been produced by developed countries, highlighting a governance gap that mirrors the technology gap itself.

Global AI Governance: Bridging the Fragmentation Gap

The report’s analysis of international AI governance reveals a deeply fragmented landscape. Seven major international governance initiatives exist, but they lack harmonization and are “largely driven by G7 members.” This creates a fundamental problem: the countries most affected by AI’s disruptions have the least voice in shaping the rules that govern it.

UNCTAD proposes a comprehensive framework for inclusive global AI governance built on several pillars. Digital public infrastructure (DPI) for AI would provide shared computing resources, open datasets, and interoperable systems that reduce barriers for developing countries. Open innovation in AI would promote open-source models and collaborative research platforms. Capacity-building partnerships would facilitate knowledge transfer, research collaboration, and technical assistance between developed and developing nations.

A particularly important recommendation is the establishment of AI public disclosure mechanisms to ensure accountability. The report argues that the dominance of multinational tech giants creates power asymmetries that require transparency and multi-stakeholder governance — including governments, private sector actors, civil society, and consumers. Understanding how different governance standards compare is critical; the ICC AI Governance Standards 2025 provide a business-oriented perspective that complements UNCTAD’s development focus.

The report also examines different regulatory approaches including risk-based frameworks, sector-specific regulation, self-regulation with industry standards, and principles-based approaches. No single model emerges as universally superior — context, institutional capacity, and development priorities should guide each country’s regulatory choices. The United Nations system, including UNCTAD, ITU, ILO, UNESCO, and WIPO, plays a critical role in facilitating multilateral coordination and ensuring developing country representation in these discussions.

From Industry 4.0 to Industry 5.0: The Human-AI Collaboration Era

The report frames the current technological moment as a potential fifth industrial revolution — Industry 5.0. While Industry 4.0 emphasized data connectivity and cyber-physical systems, Industry 5.0 is distinguished by three features: human-machine collaboration through co-creation, sustainability as an industrial imperative, and personalization driven by AI analysis of individual preferences.

This framing matters for developing countries because it suggests that the goal is not to replicate the capital-intensive, labor-displacing automation of Industry 4.0 but to pursue a more human-centric model. AI can serve as a “great equalizer” if deployed to augment human capabilities rather than replace them — enabling smallholder farmers to access precision agriculture, rural health workers to leverage diagnostic AI, and small manufacturers to optimize production without massive capital investment.

The report maps technology specialization across countries using “revealed technology advantage” metrics based on patent data. South Korea leads in 5G technology, Germany in wind energy, Japan in electric vehicles, India in nanotechnology and IoT, and the United States in gene editing and robotics. China has built strengths in big data, solar PV, and 5G. These specialization patterns suggest that developing countries need not compete across all frontier technologies but can focus on areas aligned with their comparative advantages and development priorities.

The environmental dimension also receives attention. AI currently generates more greenhouse gas emissions than the global airline industry, and data centers consume approximately 1% of global electricity demand. However, the report notes that AI could enable a 4% reduction in global greenhouse gas emissions by 2030 through efficiency improvements, reinforcing the sustainability pillar of Industry 5.0. For deeper analysis of how technical AI safety considerations intersect with these deployment decisions, additional resources are available in Libertify’s interactive library.

Actionable Roadmap for Developing Economies

Drawing together the report’s analysis, a clear actionable roadmap emerges for developing economies seeking to harness AI inclusively. The starting point is an honest assessment of national readiness across the three leverage points — infrastructure, data, and skills — using UNCTAD’s preparedness framework to identify specific gaps and priorities.

On infrastructure, countries should invest in broadband expansion, explore shared cloud computing arrangements, and consider regional data center initiatives. Public-private partnerships can accelerate deployment, while strategic engagement with international development finance institutions can help fund the capital-intensive elements of digital infrastructure.

On data, the priority is developing robust data governance frameworks that balance openness with privacy, investing in the creation and curation of locally relevant datasets, and building institutional capacity for data management. Countries should explore creative data sources — satellite imagery for agriculture, mobile phone records for economic monitoring, and community-generated health data — to overcome the limitations of traditional statistical systems.

On skills, UNCTAD recommends a dual approach: broad-based digital literacy programs for the general population combined with targeted STEM education and AI training programs. Addressing brain drain through competitive career pathways and supportive research environments is equally important. Upskilling existing workers through continuous learning programs helps ensure that AI deployment creates opportunities rather than displacement.

At the international level, developing countries should actively engage in multilateral governance forums to ensure their perspectives shape emerging AI rules and standards. Participation in open-source AI initiatives, regional cooperation on shared infrastructure, and South-South knowledge exchanges offer practical pathways to accelerate progress without waiting for top-down solutions.

The UNCTAD report ultimately makes a compelling case that inclusive AI is not merely a moral imperative but an economic necessity. With frontier technologies set to reshape $16.4 trillion in economic activity by 2033, countries that fail to invest in AI readiness risk permanent marginalization. Conversely, those that act strategically — building infrastructure, governing data wisely, and investing in their people — can leverage AI as a powerful accelerator of sustainable development.