Total factor productivity (TFP) is the economist's ultimate measure of technological progress. It captures the portion of output growth that cannot be explained by increases in labour and capital inputs — the "residual" that reflects innovation, organisational improvement, and technology adoption. If AI is truly transformative, TFP should eventually show it.
So far, it has not. OECD data shows TFP growth across advanced economies averaging 0.3% annually from 2020 to 2025 — no acceleration whatsoever from the AI investment surge. This macro-level silence has led some economists to question whether AI's economic impact will match the hype. Others argue that the measurement systems themselves are failing. Both perspectives have merit, and understanding the debate is essential for investors making long-horizon capital allocation decisions.
The disconnect between AI investment and TFP growth has three structural explanations — none of which imply that AI is failing.
TFP is a macro-level metric that averages across all firms in an economy. When a small fraction of firms achieve dramatic AI-driven productivity gains while the majority are still in the investment phase (or have not adopted AI at all), the aggregate figure is diluted to near zero.
This is precisely what the data shows. Firm-level studies consistently find significant productivity gains among mature AI adopters, while aggregate TFP remains flat. The productivity gains are real — they are just not yet widespread enough to move national statistics.
As the AI productivity paradox literature documents, general-purpose technologies require complementary investments in organisational design, workforce skills, and process redesign before productivity gains materialise at scale. Historical precedent suggests this lag is 10-20 years for transformative technologies.
AI is approximately 5-7 years into serious enterprise deployment (dating from the release of transformer-based models in 2017-2018). If historical patterns hold, TFP acceleration should become visible between 2027 and 2035.
The absence of AI-driven TFP growth in macro data is consistent with historical technology adoption patterns. Electricity, computing, and the internet all showed similar lags between investment and measurable productivity impact. The question is not whether AI will affect TFP, but when — and the historical evidence suggests we are in the transition period.
Growth accounting frameworks divide output growth into contributions from labour, capital, and the residual (TFP). These frameworks were designed when capital meant machinery and buildings. AI-era capital increasingly takes the form of intangible assets — training data, organisational knowledge, model weights, algorithmic innovations — that national accounts struggle to measure.
The Corrado-Hulten-Sichel framework attempted to address this by expanding the definition of capital investment to include intangible categories. But even the CHS framework was designed before AI and does not fully capture the distinctive characteristics of AI-driven productivity improvement.
Several major economic institutions have published projections for AI's eventual impact on TFP and GDP growth. The range is wide, reflecting fundamental uncertainty about adoption speed and complementary investment.
| Source | TFP impact | GDP impact | Time horizon | Key assumptions |
|---|---|---|---|---|
| IMF (2024) | +0.5-1.5% annually | +1.0-3.0% | By 2035 | Moderate adoption, complementary investment |
| Goldman Sachs (2023) | +1.0-1.5% annually | +1.5% | By 2033 | Generative AI automation of 25% of tasks |
| McKinsey (2023) | Not specified | +$7-14T annually | By 2030 | Broad sectoral adoption |
| OECD (2024) | +0.3-0.8% annually | +0.5-1.5% | By 2035 | Conservative, policy-dependent |
| Acemoglu (2024) | +0.06% annually | +0.1% | By 2034 | Sceptical: narrow task automation |
The enormous range in projections — from Acemoglu's near-zero estimate to Goldman Sachs' transformative forecast — reflects genuine disagreement about whether AI will automate narrow tasks or transform entire industries. Investors should be wary of projections at either extreme and instead focus on leading indicators of adoption and complementary investment.
While aggregate TFP data remains flat, sector-level evidence offers early signals of AI's productivity impact.
Financial services: AI adoption in fraud detection, credit scoring, and algorithmic trading has been measurable for a decade. Sector-level TFP in financial services has outpaced the broader economy by 0.4-0.6 percentage points since 2020, with AI cited as a contributing factor.
Healthcare: AI-assisted diagnostics, drug discovery, and clinical decision support are showing measurable efficiency gains. The challenge is that healthcare output measurement is particularly poor, making TFP attribution difficult.
Professional services: Early evidence suggests 15-25% productivity improvements in specific tasks (document review, code generation, analysis) but limited evidence of firm-level or sector-level TFP impact.
For long-horizon investors — pension funds, sovereign wealth funds, private equity with 7-10 year hold periods — the macro-economic evidence points to a specific opportunity window.
Companies investing in AI and complementary organisational change now are building the capabilities that will generate productivity gains in 2027-2035. The current period — when investment is high but measured returns are low — creates a valuation discount for AI-investing companies relative to their expected future productivity.
This is analogous to the late 1990s discount for companies investing in internet infrastructure before the revenue materialised. Investors who understood the J-curve dynamic captured significant returns when the productivity gains eventually appeared.
The sectors most likely to show early AI-driven TFP gains are those with:
Financial services, logistics, manufacturing, and professional services fit these criteria. Healthcare and education have high potential but face regulatory and measurement barriers.
AI's impact on total factor productivity is not yet visible in macro data, but this absence is consistent with every previous general-purpose technology transition. Historical evidence, firm-level studies, and economic modelling all point to AI-driven TFP acceleration in the 2027-2035 window. For investors, this creates a structural opportunity: building positions in AI-investing companies during the transition period, when investment costs are visible but productivity returns are not yet priced in. The Opagio Growth Platform helps investors track intangible capital accumulation — the leading indicator of future productivity — across portfolio companies.
David Stroll is Co-Founder and Chief Scientist at Opagio. His research on productivity economics spans three decades, from the computing revolution to the AI era. He holds a PhD in productivity economics from a leading research institution. Learn more about the Opagio team.
A comprehensive guide to measuring intangible capital — the investments in R&D, software, brand, training, and organisational design that drive over 90% of enterprise value in modern economies but remain invisible on most balance sheets. Abstract visualisation of intangible capital measurement frameworks with interconnected nodes representing R&D, brand, and organisational capital
Read more →A practical guide to building an intangible capital dashboard — the KPIs, data sources, and design principles that make intangible investment visible to boards, investors, and management teams. Dashboard visualisation showing intangible capital metrics across six categories with trend lines and benchmark comparisons
Read more →The Solow Residual — the unexplained portion of economic growth — is largely explained by unmeasured intangible capital. This article traces the connection between intangible investment, TFP, and the productivity statistics that shape economic policy. Abstract visualisation of the productivity measurement gap showing intangible capital as the hidden factor in economic growth accounting
Read more →Get the latest insights on intangible asset growth and productivity delivered to your inbox.
Book a free consultation to see how the Opagio Growth Platform can help your business.