AI’s gains are concentrating power

Artificial intelligence is often presented as a broad-based driver of productivity growth. Recent data suggests a more uneven outcome.

Research by consulting firm PwC finds that 74 percent of AI-driven financial gains are being captured by just 20 percent of firms. At the same time, its 2026 Global CEO Survey reports that 56 percent of companies have yet to record measurable returns from AI, while only a small minority report both revenue growth and cost reductions. The result is a widening gap between companies experimenting with the technology and those able to convert it into measurable gains.

A system that rewards scale

This divergence reflects how AI is being deployed within firms.

A 2025 global survey by McKinsey & Company finds that companies capturing value from AI are those using AI across large parts of their business, supported by investments in data systems and human capital. These capabilities are resource-intensive and take time to build, placing firms with existing capital and infrastructure at a clear advantage.

The effects are increasingly visible at the sector level. The PwC analysis shows that industries with higher exposure to AI are already experiencing significantly faster productivity growth, with gains in revenue per employee outpacing less exposed sectors by a wide margin. McKinsey estimates that more than half of generative AI’s economic potential is concentrated in a narrow set of functions, including software engineering, marketing, and customer operations.

Taken together, these findings suggest that AI is not diffusing evenly across the economy but is instead reinforcing differences between firms and sectors based on their ability to deploy it at scale.

Concentration across the AI ecosystem

The pattern is reinforced by the economics of the technology itself.

A 2025 Organisation for Economic Co-operation and Development (OECD) report on AI infrastructure finds that key components of the underlying AI systems, including semiconductor manufacturing, cloud computing, and data center capacity, require exceptionally high levels of capital investment and are already concentrated among a small number of firms. These barriers limit new entry and strengthen the position of established firms particularly when combined with control over data and distribution.

Investment patterns follow the same logic. The OECD reports that AI-related firms captured more than 60 percent of global venture capital in 2025, with the largest shares directed toward infrastructure and platform providers. These dynamics shape not only which firms succeed, but where value is created and retained.

From firms to economies

The same pattern is emerging across countries. Analysis by the International Monetary Fund finds that roughly 40 percent of global employment is exposed to artificial intelligence, with advanced economies both more exposed and better positioned to benefit. Higher levels of digital infrastructure, deeper capital markets, and stronger human capital allow these economies to translate adoption into productivity gains more quickly.

For smaller or financially constrained economies, exposure does not carry the same upside. Limited investment capacity and weaker institutional infrastructure constrain the ability to deploy AI at scale. In these conditions, firms tend to rely on externally developed platforms rather than build domestic capabilities, with much of the resulting value captured outside the local economy.

Lebanon’s position

Lebanon enters this transition with barriers that affect its ability to participate in this evolving landscape.

The government’s strategy, Lebanon’s Digital and AI Transformation: A Strategic Roadmap, published by the Ministry of Technology and Artificial Intelligence in May 2025 and updated March 2026, identifies limited access to capital, unreliable electricity supply, and uneven broadband coverage as core barriers to digital development. The same document highlights fragmented public systems, paper-based administrative processes, and weak data integration across ministries as additional constraints on scaling digital services.

The scale of the response reflects the scale of the challenge. The strategy sets out targets to digitize 80 percent of government services and migrate public agencies to cloud-based systems by 2030, while attracting at least $500 million in AI investment and increasing ICT employment from roughly 1 to 2 percent of the workforce to about 5 percent. It also projects that expanded broadband, digital infrastructure, and AI adoption could contribute up to 10 percent of GDP by 2035, equivalent to an estimated $3 to $4 billion in additional output.

Early implementation efforts point to incremental progress. By 2026, eight ministries had been integrated into a centralized digital dashboard tracking more than 1,000 initiatives, while projects such as national digital infrastructure, digital identity systems, and unified payment platforms remain in development. The strategy therefore outlines both the scale of Lebanon’s ambitions and the gap it will need to close to capture value from the ongoing expansion of AI.

A narrowing window

Across firms, sectors, and countries, the same pattern is evident: AI is rewarding those able to invest, integrate, and scale. For economies like Lebanon, the challenge is not the absence of adoption but the difficulty of translating that adoption into sustained economic returns. As investment, infrastructure, and technical capacity become the defining inputs of the AI economy, the window for capturing value narrows for those entering later or from a weaker starting point.