Global counterfeit trade
OECD and EUIPO estimate counterfeit and pirated goods represented up to 2.3% of global imports in 2021.
OECD/EUIPO 2025 ↗Digital systems coordinate enormous economic value. Their weakest link is often not the record itself, but whether the person, product or authority connected to that record is real.
Counterfeit products, impersonation, compromised credentials, fraudulent payments and untrusted supply chains appear in different budgets and regulatory categories. Economically, they share a common structure: a digital claim is accepted without a sufficiently strong connection to the relevant person, object or physical event.
Physical verification cannot remove every form of fraud or cyber risk. It can, however, change the economics of selected transactions by increasing the cost of substitution, narrowing unauthorized pathways and improving the evidence available when an institution must decide whether to permit an action.
As intelligent action branches across institutions and machines, trustworthy execution still requires an accountable point of origin.
INTELLIGENCE / AUTONOMY / PHYSICAL REACHAs intelligence branches into more decisions, physical verification becomes a common control point.
4SI POSITION / THE INTELLIGENT AGE
These figures measure different populations and should not be added together. Read collectively, they show the scale at which identity, authenticity and authorization failures already create economic friction.
OECD and EUIPO estimate counterfeit and pirated goods represented up to 2.3% of global imports in 2021.
OECD/EUIPO 2025 ↗The FBI received 859,532 complaints for 2024; reported losses increased 33% from 2023.
FBI IC3 2024 ↗FTC data show reported U.S. fraud losses increased 25% in 2024 while report volume remained broadly stable.
FTC 2024 ↗IBM’s 2025 study reports the global average breach cost and identifies weak AI access controls as a material governance gap.
IBM 2025 ↗A verification failure becomes economically meaningful when it changes allocation: who receives access, which product enters a supply chain, what transaction executes or which component is installed.
Select a domain to see where the cost appears, what creates the exposure and where physical verification could change a decision.
Counterfeit goods can move the cost from lost revenue into health, liability and operational continuity.
The economic case for a physical trust layer should be evaluated workflow by workflow, not through a universal market-size claim.
Counterfeit acceptance, unauthorized access, fraudulent approval, disputed custody or untrusted maintenance.
Frequency, value at risk, detectability, reversibility and downstream externalities.
Identify the physical moment where an independent verification can change the decision.
Control cost, integration burden, false decisions, avoided loss and operational value.
A single trillion-dollar claim would be false precision. The defensible conclusion is more consequential: AI lowers the cost of intelligent action, autonomy increases its frequency, and robotics gives digital authority a physical reach.
Approximate decline in the inference cost of a GPT-3.5-level system between November 2022 and October 2024, reported by Stanford’s 2025 AI Index.
Historical doubling time measured by METR for the length of software tasks frontier models complete at 50% reliability. Extrapolation remains uncertain.
Industrial robots installed worldwide in 2024 — more than twice the annual volume recorded ten years earlier, according to IFR.
AI does not need to become AGI before control risk becomes physical. Once agents can instruct robots, vehicles, industrial machinery, laboratory systems or logistics infrastructure, an authorization error can propagate from software into the real world.
The relevant exposure is therefore not a single market-size figure. It is a multiplying control surface:
As each factor grows, verification becomes infrastructure: a way to require accountable human, asset or custody authority before high-consequence execution.
The cited figures describe different geographies, years, reporting systems and loss categories. They cannot be summed into a 4SI total addressable market. Reported loss also differs from total economic cost, and a verification control will address only a subset of each category.
4SI’s inference is narrower: as digital claims become cheaper to create and economic systems become more automated, the value of controls that can independently bind selected decisions to physical reality is likely to increase. The size of that value must be demonstrated through pilots, threat models and measured deployment outcomes.
Work with 4SI to define the loss event, exposed pathway, physical control point and evidence required to judge the economics.