Global money laundering losses have fallen for the first time in the three-year history of Napier AI’s AML Index. The drop is real, but it is fragile. Compliance budgets are levelling off, and the industry must now catch more criminals without spending more to do so.
According to Napier AI, the 2026–2027 edition of the Index covers 42 markets and assesses money laundering losses, compliance costs, AML effectiveness and AI regulation. It estimates that $5.09tn was laundered worldwide last year, equal to 4.30% of global GDP. This is a year-on-year fall of $203bn, or 3.84%.
The United Nations Office on Drugs and Crime has long put the share of global GDP laundered each year at between 2% and 5%.
Napier AI argues that the absolute fall is what counts. Losses as a share of GDP fell from 4.81% to 4.30%, but nominal global GDP grew by 7.58% over the same period, which flatters the ratio. The report also includes this unattributed comment: “Losses have fallen, but not by enough, and economic growth is doing some of the work. Geopolitical disruption is handing criminals new opportunities. Governments that treat money laundering as recoverable value, not just regulatory cost, stand to gain most.”
Compliance spending fell as well. Global AML compliance cost an estimated $227.95bn, down $11.79bn, or 4.9%, from $239.74bn. It is the first year in which losses and spending have fallen together, and the cost per dollar laundered improved from 4.53 cents to 4.48 cents. The Index expects spending to stay broadly flat at about $227.8bn through 2027. Napier AI reads this as a shift towards efficiency rather than a pause before costs rise again.
That plateau is why AI matters. The Index estimates that $143.08bn in AI-driven savings remains unrealised across the markets it studied. It also finds that $3.05tn could be returned to economies each year if every market matched the AML performance of the strongest performer. Napier AI says the route to that value is better detection, not bigger budgets, and that this depends on AI that is explainable, auditable and accountable to people.
Napier AI chief data scientist Dr Janet Bastiman said, “Over $3 trillion in AML recovery offers huge potential for economies and organisations, and our data shows that spending more on compliance isn’t the only answer. Better detection and more robust underlying data, underpinned by AI, give compliance teams a more accurate picture of their financial risk, allowing them to focus on the largest money laundering threats.”
She added, “But that AI must be explainable, and every automated decision must be explainable to a regulator. It requires human judgement both to own the decisions and to understand the patterns that criminals are using to stop illicit flows.”
The improvement is uneven across markets. Scores run from one to ten, and lower is better. The Czech Republic leads with a score of 1.81, and Canada ranks second. Greece, new to the Index this year, has the strongest AML effectiveness score in the world. The United States ranks 16th, with $658.43bn in losses (2.1% of GDP) and the largest unrealised AI opportunity of any market at $27.69bn. The UK climbs ten places to 24th, but losses of $206.72bn equal 5.16% of GDP.
Its compliance bill is $7.87bn, and a further $4.11bn in AI savings remains untapped, the sixth largest opportunity in the Index. Russia moves the other way: its losses rose by $89.82bn to $278.09bn, which is more than the increase for the whole of Europe.
Copyright © 2026 RegTech Analyst
Copyright © 2026 RegTech Analyst
