India Trains a Large Pool Of AI Talent

The figure behind every chart and every row in this piece is a country of education, not a nationality. Revelio Labs builds its dataset from public professional profiles, roughly 616 million people across the global workforce as of its September 2025 snapshot, from which it identifies about 1.6 million technical AI professionals. For each of them, interface's analysis records where they earned their first university degree, tracked over a movement window running from April 2024 through September 2025.

A Proxy, Not a Nationality

Studying in India does not require Indian citizenship, and holding Indian citizenship does not require studying there. interface's own note on the data puts the scale of that gap in context: roughly 8% of OECD bachelor's students study outside their home country, so any national total already mixes citizens who studied at home, citizens who studied abroad and non-citizens who studied domestically. That is why this piece writes 'educated in India' throughout rather than 'Indian' or 'Indian-born,' even where the shorter phrase would read more naturally.

The same caveat applies everywhere else in this piece, and it is why the dataset's biggest blind spot, covered further down, is a coverage gap rather than a finding: professional-profile platforms are not used evenly around the world.

Nine of the ten countries in the hero chart saw India's share of their AI workforce grow in a single year, and the size of the gains varies a lot. Ireland moved the most, from 22.4% to 30.3%, a rise of 7.9 percentage points. The United Arab Emirates was close behind at 5.1 points, from 28.5% to 33.6%, already the highest share of any country tracked here in both years.

Canada rose 3.7 points (16.3% to 20.0%), the United Kingdom 3.6 points (12.7% to 16.3%), Australia 3.3 points (13.7% to 17.0%) and the United States 3.1 points (17.5% to 20.6%). Luxembourg, Germany and Sweden moved less than a single point each, a reminder that a widening share means something different in a market that already had a large India-educated base than in one still building one.

Singapore's share fell from 13.2% to 12.1%, a drop of 1.1 percentage points and the only decline among the ten countries charted. It would have been easy to leave Singapore out of a piece built around a rising trend, and that is exactly why it stays in. A single-year drop in one country does not undo a rise recorded in nine others, but a chart that only ever shows increases invites the reasonable question of what got left out.

Singapore's own AI workforce is smaller than the United States' or India's, which makes a single year's reading more sensitive to hiring swings at a handful of companies than the same movement would be in a bigger market. That is a reason for caution reading too much into one country's single-year number, not a reason to remove it from the comparison.

The ten countries on the chart are not the whole picture. India ranks among the top five countries of education in 33 destinations in this dataset, and the ten shown are simply the ones where its share is largest. Four of the 33 sit above 20% and twelve at or above 10%, while fifteen are below 5%. The spread runs from the United Arab Emirates at 33.6% down to Brazil at 0.4%. This is not a uniform global shift. It is a steep curve with a small group of countries bunched at the top of it.

Movement below the chart runs the same way. New Zealand rose from 9.3% to 12.0% and Japan from 7.9% to 10.5%, both crossing 10% between the two years, while Saudi Arabia climbed from 3.6% to 5.9%. Eleven of the 33 did fall, and Singapore is not the steepest of them: Finland dropped 1.4 points, from 7.0% to 5.6%, and Estonia fell 1.1 points. So the pattern is wide rather than universal, which is the distinction the full table below sets out row by row.

China does not appear on the chart, and interface is explicit about why: the Revelio Labs dataset lacks comprehensive coverage of China because of limitations in its underlying sources, which have far lower penetration there than in most of the countries measured here. That is a statement about which data exists, not about how much AI talent China has.

The absence does not inflate anything on the chart. Each figure is a share of that country's entire AI workforce, not a share of the few origins named beside it. For the United States the five listed origins account for 91.4% of the workforce, for Ireland 78.9%, for the UAE 67.8%. The rest sits in a remainder that is simply not broken out. Leaving one country off the display does not hand its people to another, so India's 20.6% in the United States reads 20.6% whether or not China is drawn.

interface's own report points to independent estimates using other sources that put combined tech talent pools in Beijing and Shanghai above 500,000 people. That figure is not built the same way as the 1.6 million AI professionals identified in this dataset, so the two cannot be set against each other, and this piece does not try. What can be said plainly is this: nothing here supports a claim about the size of China's own AI workforce, and nothing here supports a claim that Chinese-trained talent is being replaced by India-trained talent anywhere.

The pattern holds up without needing its exceptions edited out. India is a top-five source of AI talent in 33 of 46 tracked destinations, and the share widened again in 22 of those 33 between 2024 and 2025, Singapore's single decline included. It rose in the largest AI workforce on the chart and in some of the smallest, which is what makes this look like supply rather than one country's recruitment drive.

None of that makes it a story about winning or losing. Ireland's 7.9-point jump and the UAE's 5.1-point jump say more about how those two economies are hiring than about any contest between countries. And China's absence from the chart is a reminder that a supply story built on one dataset can only describe what that dataset measures: one country's graduates, turning up in a growing share of the AI workforce almost everywhere there is data to check.