The newest Anthropic Economic Index data shows the top 20 countries account for 48% of all Claude usage, up from 45% six months earlier. AI adoption is concentrating, not spreading.
The top 20 countries by per-capita Claude usage now account for 48% of all conversations, up from 45% just six months earlier. AI adoption is not spreading evenly across the globe. It is concentrating. The new Anthropic Economic Index data shows a deepening divide between the countries leaning hardest into AI tools and the ones still on the runway, and the gap is widening, not closing.
The shift matters because almost every public narrative about AI assumes broad, fast diffusion. The data say otherwise. Inside the high-adoption cluster, usage is compounding rapidly. Outside it, growth is much slower. Anyone planning a career in the AI economy needs to understand which side of that divide their labor market sits on, because the structural pressure on roles looks different in each.
What the AUI actually measures
The Anthropic AI Usage Index (AUI) is a per-capita concentration ratio. For each country it divides that country's share of global Claude conversations by its share of global working-age population (ages 15 to 64). A score of 1.0 means usage matches population. A score of 4.0 means the country uses Claude four times as much as population alone would predict. The full leaderboard for every country and US state is published on the AI Adoption by Country page, updated through Anthropic's February 2026 snapshot.
The metric is useful precisely because it does not reward population size. China, India, and the United States dominate any league table by raw total usage simply because they are big. The AUI strips that out and reveals where AI is actually integrated into everyday knowledge work, regardless of how big the country is.
Who leads the leaderboard
In February 2026 the top ten by AUI are: Monaco (9.6x), Cayman Islands (4.9x), Andorra (4.8x), Singapore (4.6x), Liechtenstein (4.5x), Israel (4.4x), Luxembourg (4.2x), Switzerland (4.1x), the United States (3.8x), and Canada (3.7x). Several micro-states show outsized scores because a small population of high-tech residents drives a large share of per-capita usage. Among populous countries, Singapore, the United States, Canada, and Australia carry the most signal.
Israel was the global leader in August 2025 at 7.0x. It is now sixth at 4.4x. That drop reflects the rest of the leading cluster catching up rather than Israel using less AI in absolute terms. Israel's raw conversation volume actually grew, but its share of global usage fell as Monaco, Cayman Islands, and Andorra accelerated past it.
Why concentration is widening, not narrowing
Three structural factors explain the deepening concentration. First, the high-adoption countries already have the conditions AI adoption depends on: high-bandwidth English-language internet access, large stocks of knowledge workers, capital available to license enterprise tools, and developer talent that ships AI features into production. Those conditions compound. Each quarter of usage produces more in-country expertise, more local case studies, and more demand for the next generation of tools.
Second, emerging markets are hitting real infrastructure constraints. Adoption requires more than smartphones. It requires the ability to use English-language interfaces fluently, the ability to pay for premium tiers, and the workplace conditions in which AI tooling is welcome rather than restricted. Most of the global population is still building one or more of those preconditions. Until the constraints lift, the divide widens.
Third, the Anthropic Economic Index is measuring Claude specifically. Claude has strong adoption in high-income tech economies and weaker presence in many emerging markets. Other AI assistants are likely to show different geographic patterns. But the structural shape (a top cluster pulling away from a long tail) is consistent across most published AI usage datasets.
What concentration means for careers
If you work in one of the top-20 countries, you are inside the labor market where AI compression hits earliest and hardest. Your role's exposure to automation is realised on a shorter timeline than the equivalent role in a low-AUI market. That is both risk and opportunity. The roles that adapt first capture the wage premiums that come from being early to the augmentation layer. The roles that do not are the ones whose compensation gets reset downward first.
If you work in a low-AUI country, you have more runway. Adoption will reach your market; it will just take longer than the headlines suggest. The right move is not to wait. It is to use the runway to build the same authority-layer position that high-AUI countries are now selecting for. By the time the compression wave hits, the position should already exist.
Either way the structural reading is the same. AI does not replace roles uniformly; it reweights the task composition of each role. The routine cognitive layer compresses. The authority, judgement, and integration layer is where wages settle. The AI Career Index scores both layers for over a thousand specific roles in the AI Career Risk Rankings, and the acceleration plan maps the six-month moves into the durable layer.
How to read the next snapshot
Anthropic publishes the Economic Index every two to three months. Watch three things in the next release. First, whether the top-20 share keeps climbing past 48%. A second consecutive rise would confirm concentration is a structural feature, not a one-snapshot artefact. Second, whether Israel and Singapore continue to slide, which would suggest the early leaders are losing relative position even as they grow in absolute terms. Third, whether any large emerging-market economy (India, Indonesia, Brazil) shows real AUI gains, which would be the first sign that the bottom of the distribution is starting to move.
Right now the data point in one direction: the AI economy is being built first in a relatively small set of high-income countries, and the gap to the rest is widening. That is the structural backdrop every career plan needs to start from. The full leaderboard, with trend arrows running back to August 2025, is on the AI Adoption by Country page. The reading takes about ten minutes; the implications are worth more than that.
Run your structural assessment
The AI Career Index is free to take. Get your structural snapshot in 10 minutes, including your durability score, archetype, and projection.
Take the Free Assessment