Uneven AI Adoption: Insights from Anthropic’s September 2025 Economic Index

In mid-September 2025, Anthropic released a report under its Economic Index banner titled “Uneven geographic and enterprise AI adoption.” The document synthesises survey data, usage metrics and macroeconomic indicators to paint a nuanced picture of how AI is actually being used in the economy—not just in headline-grabbing pilots, but in day-to-day workflows across countries and industries.

The central finding is simple but important: AI adoption is highly uneven. A relatively small cluster of advanced economies and large firms account for a disproportionate share of investment, experimentation and productivity gains. Within countries, the gap between AI leaders and laggards is often wider than the gap between sectors. A handful of “frontier adopters” integrate models into core processes—supply-chain optimisation, product design, customer operations—while many others confine AI to isolated chatbots or proof-of-concept projects.

Anthropic’s report highlights several drivers of this unevenness. Access to talent and infrastructure is one. Organisations with in-house data teams, MLOps capabilities and robust cloud partnerships are far more likely to move beyond pilots. Regulatory clarity and digital maturity also matter: countries with clear data-protection regimes, digital ID systems and e-government initiatives tend to see more confident AI deployment by the private sector. Conversely, regions with fragmented connectivity, limited digital skills or uncertain regulatory environments lag behind, even when high-level interest in AI is strong.

The report also points to cultural and organisational factors. Firms that treat AI as a strategic, C-suite issue—linked to measurable business outcomes—tend to progress faster than those that frame it purely as an IT experiment or a marketing gimmick. Strong cross-functional collaboration between business, technology, risk and operations teams is a common trait among successful adopters. In contrast, siloed organisations often struggle to move from promising prototypes to systems that actually change how work gets done.

From a macroeconomic perspective, uneven adoption poses both opportunities and risks. On the upside, there is significant room for catch-up growth. If lagging firms and regions can close even part of the gap, the aggregate productivity gains could be substantial. On the downside, there is a risk of growing divergence: leading firms may compound their advantages as they reinvest AI-driven profits into more automation and innovation, while laggards fall further behind. This dynamic could exacerbate inequality across regions, sectors and income groups.

For policymakers, Anthropic’s findings suggest that debates about AI should move beyond frontier capabilities and focus more on diffusion. That means investing in digital infrastructure, skills, open tools and shared resources that make it easier for smaller organisations to adopt AI responsibly. It also means designing regulations that protect fundamental rights without inadvertently raising barriers to entry for smaller players.

From Synergy AI Tech Solutions’ vantage point, the Economic Index report is a useful tool for clients to benchmark themselves. Are you in the leading, middle or trailing cohort for your sector and region? Do you have a clear AI roadmap tied to your strategy, or are you still in demo mode? Are your investments balanced between experimentation, productionisation and governance? Honest answers to these questions can help prioritise the next 12–24 months of work.

Practically, closing the adoption gap often starts with a portfolio approach: identify a small number of high-impact, low-regret use cases (for example, internal knowledge search, code assistance, document automation), implement them well with strong governance, and use the resulting experience and credibility to fuel more ambitious initiatives. Anthropic’s September 2025 analysis may describe a world of uneven adoption—but it also implicitly shows the path that successful organisations are taking to move up the curve.