Business investment has been disappointingly weak across the OECD since the Global Financial Crisis (GFC), which has contributed to slow potential growth and stagnating living standards, particularly in large parts of Europe (Draghi 2024, Andre and Gal 2024). Some of this weakness may be explained by subdued demand and elevated uncertainty (Dlugosch et al. 2025).

Yet one component has been anything but sluggish: investment in digital assets, such as hardware, software, and databases. In our new paper (Gal et al. 2026), we use detailed national accounts data by asset type and sector for 32 OECD countries (1995–2023) to document striking differences in the investment dynamics across digital and non-digital assets. In addition, we find an important divergence across countries, with the US and a few other countries from the Baltics and Central and Eastern Europe showing relatively strong performance.

Digital investment has pulled ahead of other asset classes

Across the OECD, real investment in digital assets has risen strongly for two decades and was notably resilient during both the GFC and the Covid-19 shock (Figure 1). Since 2007, average real digital investment across OECD economies increased by over 130%. Within this broad category, ICT hardware roughly doubled while software and databases almost tripled. In contrast, real business investment in non-digital tangible assets has barely grown since the GFC, with volumes only slightly above their 2007 level. With the rising importance and rapid diffusion of artificial intelligence and related technologies (Filippucci et al. 2024), digital assets are likely to matter even more (Carpinelli et al. 2026).

Figure 1 Digital investment has outpaced other investment types across the OECD, most strongly in the US

Note: Aggregates are GDP-PPP weighted; Ireland excluded due to a 2015 series break.
Source: OECD National Accounts Database and OECD calculations, see details in Gal et al (2026). 

Digital investment is rising in all major OECD regions, but the slope differs sharply. The contrast is strongest in the US, where both ICT hardware, and software and data investment accelerate markedly after the mid-2010s, reaching 3-4 times higher volumes by 2023 than in 2007. Western European economies (EU-15 in Figure 1) also expanded digital investment, but at a slower pace, with only a 1.5–2-fold increase over the same period. Japan lags further behind. A plausible interpretation is that the mid-2010s acceleration coincided with the commercialisation of cloud computing and the broader diffusion of data-intensive and AI-enabled applications, which raised demand for software, data management, and computing capacity.

Nearly all of the US outperformance in investments is due to digital assets

A decomposition of business investment growth by asset type shows that digital assets account for almost the entire gap between US business investment growth and that of the rest of the OECD over the past decade (Figure 2, Panel A). From 2014–2023, US business investment grew about 2 percentage points faster per year than the rest of the OECD, and almost all of that gap reflects faster digital investment. Robustness checks suggest this is not driven by price-measurement choices or by a shift from buying to renting digital services. 

Figure 2 Digital investment accounts for almost 90 % of the gap in business investment growth between the US and other OECD countries over the past decade

Note: Average annual real business investment growth (2014–23). Panel B shows the US–OECD (ex-US) growth gap split by asset type; digital is split into ICT and non-ICT sectors. OECD aggregates are GDP-PPP weighted.
Source: OECD National Accounts Database and OECD calculations, see details in Gal et al (2026).

When decomposing the ‘investment-growth gap’ between US and the rest of the OECD,  we find that stronger growth in digital investment explains nearly 90% of the outperformance of the US. Moreover, most of this gap reflects higher digital investment outside the tech sector in the US – consistent with an economy-wide broader diffusion of digital technologies there. Splitting the digital component of the gap into ICT sectors versus the rest of the economy, we find that only about 40% of the digital-investment gap comes from the ICT-producing and ICT-service sectors themselves (Figure 2, Panel B).

Digging deeper, in our firm level analysis we find that the so-called ‘Magnificent 7’ Big Tech companies accounted for a quarter of total digital investment in the US – a significant but not a dominant share. Both sector- and firm-level evidence points to the prevalence of digital intensity in the US. This in turn suggests that economy-wide differences in the incentives and capabilities to invest in, adopt, and scale-up digital technologies play an important role.

Digital capital is diverging across countries

Investment is a flow; what shapes productive capacity is the resulting capital stock. Digital assets – think of computers and software – depreciate rapidly, at a rate of more than 25% per year compared to 5-20% for non-digital assets such as buildings and machinery (OECD, 2025). Therefore, strong gross investment is required to build and maintain a sizable ‘digital capital’ base. Despite faster depreciation, digital capital accumulation has indeed outpaced that of most other assets, given the rapid rise of digital investment. However, this accumulation has been very uneven across countries.

Figure 3 contrasts the cross-country dynamics of total capital per worker (Panel A) with digital capital per worker (Panel B). Total capital shows a familiar convergence pattern: economies with lower initial capital per worker tend to grow faster, catching up over time. Digital capital does not. The United States starts from a high level and continues to accumulate quickly, pulling further ahead. Moreover, countries with smaller initial digital capital bases often record slower subsequent digital capital growth (e.g. Germany, Canada), implying ‘falling behind’ dynamics rather than catch-up in terms of digital capital. As digital capital becomes increasingly central to productivity growth – a plausible scenario given the diffusion of AI, cloud computing, and data-intensive business processes – divergence in digital capital per worker could translate into more persistent divergence in productivity and incomes.

Figure 3 Lack of convergence in digital capital stocks per worker

a) Growth over the past ten years and initial levels: total capital per worker

b) Growth over the past ten years and initial levels: digital capital per worker

Note: Capital–labour ratios are calculated using the working-age population, defined as individuals aged 16–64. Capital stock is converted into USD using GDP-based purchasing power parities (PPPs), as no specific PPPs for capital stocks exist.
Source: OECD National Accounts Database and OECD calculations.

What role for government policies?

A few policy areas stand out as likely candidates to explain cross-country differences in investment performance, especially through stronger potential impacts on digital investments.

Access to finance

A large part of digital assets are intangible (software and databases) and have limited collateral value, yet frequently require sizable upfront spending and rapid replacement. That combination can make digital investment particularly sensitive to financing conditions and to the structure of capital markets. Policies that deepen equity financing, support venture capital, and more improve the financing environment have been shown to be effective regarding intangibles (Demmou and Franco 2021).

Labour markets and complementary intangibles

Digital assets are productive only when complemented with a range of intangibles such as skills, organisational capital, data governance, and (often) R&D (Pisu et al 2021, Sorbe et al. 2019). Cross-country differences in digital investment are likely to reflect differences in these complements. Policies that raise digital skills, reduce adoption costs and do not penalise experimentation with new technologies through excessive labour regulation can encourage technology diffusion through new investments across firms and sectors.

Scale, competition, and market fragmentation

Digital technologies tend to feature high fixed costs and scale economies (McMahon et al. 2024). Barriers to scaling up – including across borders – and weak competitive pressure can reduce incentives to undertake large, risky digital investments. These factors seem particularly relevant in explaining the US-EU investment performance gap, given the more unified US internal market (Cerdeiro and Rotunno 2026).

Digital investment has become the bright spot in an otherwise subdued business investment environment. It has also been a leading factor driving cross-country differences and in particular the growing digital capital gap between the US and the rest of the OECD. A better understanding of the reasons and policy levers is critical, given that investment in new technologies is a key driver of sustained stronger productivity growth and, in turn, rising living standards.

References

André, C and P Gal (2024), “Reviving productivity growth: A review of policies”, OECD Economics Department Working Paper No. 1822.

Carpinelli, L, F Natoli  and M Taboga (2026), “From AI investment to GDP growth: An ecosystem view”, VoxEU.org.

Cerdeiro, D A and L Rotunno (2026), “EU barriers to scaling up: The case of fragmented product markets”, VoxEU.org.

Demmou, L and G Franco (2021), “Mind the financing gap: Enhancing the contribution of intangible assets to productivity”, OECD Economics Department Working Paper No. 1681.

Dlugosch, D, M Glanville, J Hooley, F Ozturk and B Westmore (2025), “Understanding the weakness in business investment”, OECD Economics Department Working Paper 1836.

Draghi, M (2024), The Future of European Competitiveness – A Competitiveness Strategy for Europe.

Gal, P, J Hooley, F Ozturk and F Unsal (2026), “Business investment in the face of the digital transformation: Initial evidence”, OECD Economics Department Working Paper No. 1859.

McMahon, M, S calligaris, E Doyle and S Kinsella (2021), “Scale, market power and competition in a digital world: Is bigger better?”, OECD Science, Technology and Industry Working Paper No. 2021/01.

OECD (2025), “Annual capital formation by economic activity and Fixed assets by economic activity and by asset”, National Accounts Database.

Pisu, M, C von Rüden, H Hwang and G Nicoletti (2021), “Spurring growth and closing gaps through digitalisation in a post-COVID world: Policies to LIFT all boats”, OECD Economic Policy Paper No. 30.

Schoefer, B (2025), “Eurosclerosis at 40: Labor Market Institutions, Dyanamism, and European Competitiveness”, NBER Working Paper No. 33975.

Sorbe, S, P Gal, G Nicoletti and C Timiliotis (2019), “Digital Dividend: Policies to Harness the Productivity Potential of Digital Technologies”, OECD Economic Policy Paper No. 26.