Key findings
- 21 countries score 10.0 and 22 score 0.0 on risk velocity, an index of digital infrastructure (broadband and secure servers), not of AI adoption or job loss
- For 20 of the 21 top countries our data has no broadband figure, so the 10.0 rests on secure-server counts alone; only the Cayman Islands has both indicators
- Vietnam reaches 10.0 with a low AI exposure of 3.27/10, and South Africa with GDP per capita of $6,598: readiness and exposure are separate questions
- Germany (9.6), Canada (9.2), Japan (8.8) and Australia (8.5) fall short of 10.0 partly because they have a broadband figure just under the cap, a data-coverage effect rather than weaker infrastructure
- The index says nothing about how many workers use AI or when jobs will change; for that, use occupation exposure and measured AI use
Readiness and Exposure Are Two Different Questions
Most coverage of AI and work asks how many jobs are exposed. A second question is whether a country has the digital infrastructure that AI tools run on. WorldJobsData's risk velocity index answers only that second question. It combines two World Bank indicators, fixed broadband subscriptions per 100 people and secure internet servers per million people, into a score from 0 to 10.
Read it as readiness, not as a forecast. A high score means the basic infrastructure for deploying AI tools is widely available. It does not tell you how many workers use AI, which employers have adopted it, or when any job will change. Our methodology page says the same. The "1-3 years" to "12+ years" wording that earlier versions of this article used was a label for score ranges, and it was wrong to present it as a prediction.
The 21 Countries at the 10.0 Maximum
21 of the 198 countries with a score reach 10.0, covering about 436 million workers. They range from Switzerland ($114,769 GDP per capita) to Vietnam ($5,066). Four small territories (Isle of Man, Cayman Islands, US Virgin Islands, Seychelles) also reach the maximum and are left out of the table.
| Country | Risk velocity | Secure servers per 1M people | AI exposure | GDP/capita | Workers |
|---|---|---|---|---|---|
| United States | 10.0/10 | 196,554 | 5.06/10 | $90,027 | 144.0M |
| Russia | 10.0/10 | 26,130 | 4.62/10 | $17,547 | 73.5M |
| Vietnam | 10.0/10 | 5,734 | 3.27/10 | $5,066 | 53.7M |
| United Kingdom | 10.0/10 | 68,449 | 5.08/10 | $57,602 | 34.1M |
| Turkiye | 10.0/10 | 11,103 | 4.11/10 | $18,599 | 32.4M |
| Spain | 10.0/10 | 29,986 | 4.72/10 | $38,627 | 22.2M |
| Poland | 10.0/10 | 41,132 | 5.04/10 | $28,420 | 17.1M |
| Ukraine | 10.0/10 | 9,937 | 4.63/10 | $5,866 | 15.0M |
| South Africa | 10.0/10 | 16,658 | 4.84/10 | $6,598 | 12.7M |
| Romania | 10.0/10 | 27,230 | 4.26/10 | $22,538 | 7.7M |
| Sweden | 10.0/10 | 55,097 | 5.32/10 | $63,133 | 5.3M |
| Portugal | 10.0/10 | 28,710 | 4.88/10 | $32,082 | 5.2M |
| Switzerland | 10.0/10 | 167,358 | 5.35/10 | $114,769 | 4.6M |
| Serbia | 10.0/10 | 12,942 | 4.69/10 | $15,262 | 2.8M |
| Slovakia | 10.0/10 | 43,287 | 4.91/10 | $28,544 | 2.6M |
| Singapore | 10.0/10 | 209,665 | 5.35/10 | $98,814 | 2.3M |
| Slovenia | 10.0/10 | 68,783 | 4.86/10 | $37,376 | 990K |
Source: World Bank Open Data (secure servers per million people, 2024; GDP per capita), ILO ILOSTAT (employment), WorldJobsData (AI exposure, model-estimated). Risk velocity scores as of 2026.
One caveat matters when reading this table. For 20 of the 21 top countries our data has no broadband figure, so their score rests on secure-server counts alone, and every one of them is above the 5,000 per million cap. Only the Cayman Islands has both indicators. Across all 198 scored countries, 63 lack a broadband figure.
Why Germany, Canada, Japan and Australia Are Not at 10.0
These four countries have a broadband figure and a secure-server count far above the cap (Germany: 152,114 per million). Their score therefore depends on broadband alone, and each sits just under the 50 per 100 cap: Germany 45.6, Canada 42.5, Japan 38.6, Australia 36.5. That gives 9.6, 9.2, 8.8 and 8.5. Countries with no broadband figure skip that step and can reach 10.0 on servers alone, so part of the gap between 9.6 and 10.0 is a data-coverage effect, not a real infrastructure difference.
Vietnam and South Africa: Reaching the Maximum on Servers Alone
Vietnam scores 10.0 with 5,734 secure servers per million people, above the 5,000 cap, and no broadband figure in our data. Its average AI exposure is 3.27/10, the lowest of the 17 larger countries in the table, and ILO data put its informal employment at 67.0%. South Africa scores 10.0 with 16,658 servers per million, GDP per capita of $6,598 and informal employment of 34.8%.
Earlier versions of this article explained Vietnam's score by factory automation and robotics. The index has no sector or robotics data, so that explanation was not supported by the data and has been removed. What the score does show is that secure-server counts, a hosting measure, are high in both countries. It does not show that workers in either country are using AI tools.
The 22 Countries at 0.0 and Close to It
22 countries score 0.0, and Tanzania and Ethiopia score 0.1. Where we have both indicators the pattern is the same: almost no fixed broadband (DR Congo: 0.03 per 100 people) and very few secure servers (10 per million).
| Country | Risk velocity | Broadband per 100 people | Secure servers per 1M people | AI exposure | Workers | Informal employment |
|---|---|---|---|---|---|---|
| Afghanistan | 0.0/10 | 0.08 | 41 | 3.12/10 | 7.4M | 86.1% |
| DR Congo | 0.0/10 | 0.03 | 10 | 3.35/10 | 30.6M | 97.3% |
| Madagascar | 0.0/10 | 0.12 | 25 | 2.91/10 | 11.9M | 96.1% |
| Malawi | 0.0/10 | 0.09 | 32 | 3.38/10 | 5.0M | 92.7% |
| Mozambique | 0.0/10 | 0.14 | 37 | 3.15/10 | 12.2M | 95.5% |
| Niger | 0.0/10 | 0.11 | 5 | 3.12/10 | 9.3M | 98.5% |
| Haiti | 0.0/10 | 0.31 | 8 | 3.29/10 | 1.9M | 91.6% |
| Burundi | 0.0/10 | 0.02 | 11 | 3.05/10 | 5.0M | 98.0% |
| Tanzania | 0.1/10 | n/a | 52 | 2.85/10 | 31.0M | 90.8% |
| Ethiopia | 0.1/10 | 0.59 | 24 | 2.94/10 | 36.6M | 77.7% |
Source: World Bank Open Data (broadband per 100 people and secure servers per million, 2022-2024), ILO ILOSTAT (employment and informal employment), WorldJobsData (AI exposure, model-estimated). Tanzania and Ethiopia are shown at 0.1.
These economies also have very high informal employment (78% to 98% in the countries above that have an ILO figure). That is our reading of why AI tools have fewer formal workplaces to enter, not something the index measures.
What This Does and Does Not Mean
A low score is not safety. It means the infrastructure is thin today; infrastructure changes, and the index is a snapshot of 2022 to 2024 data. A high score is not a warning that jobs will go in a set number of years. To judge risk for a specific job, start with the exposure of that occupation using the job risk calculator, and for how much AI is actually being used, see AI exposure vs AI use by country.
Limits of This Index
- Secure-server counts reflect where websites are hosted. Hosting hubs sit far above the cap (Singapore 209,665 per million, Netherlands 198,682), so the index can rate a hub higher than its workers' real tool use warrants.
- 63 of 198 scores lack a broadband figure and rely on servers alone.
- Data years vary from 2022 to 2024, and the index does not include regulation, skills, cost or employer adoption.
See readiness and exposure for your country
Interactive data for all 206 countries: workforce structure, exposure scores and the risk velocity index, with a note on what each means.
Explore the Data →Was this analysis useful?
Let us know what you think - your reaction helps us understand what to cover next.
Thanks for your reaction!
Get new country analyses, data updates, and AI labour market insights. No spam - one email when something worth reading drops.
Methodology
Risk velocity is built by WorldJobsData from two World Bank indicators: fixed broadband subscriptions per 100 people (IT.NET.BBND.P2, capped at 50) and secure internet servers per million people (IT.NET.SECR.P6, capped at 5,000), combined as a geometric mean and scaled from 0 to 10. Where only one indicator exists for a country, the score uses that one alone. Data years are 2022 to 2024. The index does not use occupations, measured AI adoption or any forecast; the "1-3 years" to "12+ years" wording on some pages is a label for score ranges, not a prediction. AI exposure scores are model-estimated by an AI language model and are not measurements.
AI exposure, robotics, offshoring and work-from-home scores are model-estimated: an AI language model (Claude Opus 4.7) scored each ISCO-08 occupation group on 28 May 2026, guided by published research. They are informed judgements, not measurements, and not forecasts of job losses. "Risk velocity" and its time windows come from a digital-infrastructure index (broadband and secure servers), not from occupations or measured AI adoption. Read the full methodology and how the scores were checked.
Frequently asked questions
Which countries are most digitally ready for AI rollout?
Which countries score lowest on AI readiness?
Does risk velocity predict when AI will disrupt jobs?
Where does the risk velocity data come from?
Data sources
- World Bank Open Data - Fixed broadband subscriptions per 100 people (IT.NET.BBND.P2) and secure internet servers per million people (IT.NET.SECR.P6), latest year 2022-2024, CC BY 4.0.
- ILO ILOSTAT - Employment by occupation (ISCO-08) and informal employment rate, CC BY 4.0.
- World Bank Open Data - GDP per capita (current USD), CC BY 4.0.
- WorldJobsData methodology - how risk velocity and AI exposure are built, and what they cannot tell you.
Spotted an error or know a better source? Report it and we will check it against the source data.