Each year, about €260 billion of value added escapes the European economy due to purchases of cloud services and software from American providers.
The consulting firm Asterès reported this estimate in April 2025, as part of a study commissioned by Cigref. While based on extrapolations from a small sample rather than exhaustive macroeconomic data, the figure has been widely cited since then.
Asterès continued by examining the consequences of technological dependencies from the angle of price increases for cloud-software services. It argues that over the 2026-2030 period, these increases could, each year, amount to an average extra cost of €140 billion and deprive the European economy of €107 billion in value added.
Inputs from About Fifty CIOs
The study relied on a questionnaire answered by 54 European CIOs who are members or partners of Cigref, spanning 14 sectors and three countries (France, Belgium, the Netherlands). Asterès adjusted the results to reflect the weight of each sector in the European economy by revenue. It acknowledges a self-selection bias, with the most affected companies presumably being the most likely to respond.
Prior to the survey, there was a review of the literature on the topic. Asterès cites mainly five sources:
- Two reports by its economists Guillaume Moukala and Charles-Antoine Schwerer
On one side, Cloud computing: preserving competition to support a competitive economy (2023, for Google).
On the other side, Cloud computing: between growth contribution and anti-competitive practices (2022, for the Concorde Foundation). - Two analyses by economist Frédéric Jenny for the CISPE association
On one side, Cloud Infrastructure Services: An analysis of potentially anti-competitive practices (2021).
On the other, Potential Market Distortions in the Cloud Infrastructure Services Market (2023). - A white paper from the consulting firm Elée: Explosion des coûts logiciels: comment reprendre la main sur vos budgets ? (2026)
According to Asterès, this literature illustrates limited competition in this market, keeping users in a position of dependence on their suppliers. Those suppliers, by virtue of their market power, can impose inflation higher than the average.
Asterès cites a 2022 survey conducted with Vanson Bourne. It found that 90% of French companies were affected by headline price increases in the order of 3% to 6% per year. And that 70% had already been offered bundled packages with an average price differential of 26%.
An Average Annual Inflation of 8.7%
Based on the testimonies of the 54 respondents, headline price increases are the most common inflation mechanism. Other effects were observed (volume, cost structure, scope…), rooted in potentially abusive business practices:
- Vendor lock-in (40% of respondents say they have been victims in the last three years)
- AI by design (34%)
- Planned obsolescence (30%)
- Bundling (26%)
- Reduction of metrics (21%)
Respondents report an average 8.7% increase in the cost of cloud-software services over the past three years.
Across the entire sample, the digital budget accounts for 4.1% of revenue (or, for public administrations, of GDP). It is directed 28% toward cloud-software services. Under the current inflation trajectory, that share could reach 42% by 2030.
To absorb the surcharges, 47% of respondents cut other digital expenditures. Primary targets: external ESN services (57%), hardware purchases (27%), and IT human resources (13%).
33% plan to increase their overall digital budget, at the expense of R&D (32%), cash flow and margins (26%), recruitment (21%), and salary increases (12%).
€140 Billion of Annual Overcost from 2026-2030
In last year’s study, Asterès estimated €400 billion as the annual European spend on cloud-software services. Building on this figure, it projected the net inflation overcost over five years, at two levels: for buyers and for the European economy. Regarding the latter, it assumed that 34% of the price increases would be reinvested. That assumption also came from last year’s study, which showed that for every €1 spent on cloud-software, 17 cents went directly to European suppliers and 17 cents more were reinvested in Europe by non-European suppliers.
Despite the 8.7% observed by CIOs, Asterès retained an average inflation rate of 12% for its projections. This mainly stems from a mechanical continuation of the trend: the 2022 Vanson Bourne survey had shown 5.8%. Another factor: AI acting as a price-raising lever. The recent announcements by GitHub and Anthropic, moving to usage-based billing, attest to this.
With these parameters, the firm concludes that price rises would cost European companies and public administrations an average €140 billion per year from 2026 to 2030, of which €93 billion are net losses at the European economy level.
€65 Billion per Year Not Reinvested in the European Economy
By distributing the cost increases by funding mode and factoring in the share of imports, Asterès estimated the crowding-out effect. In other words, the total spending and investments that firms forgo to finance the price hikes. As mentioned above, this corresponds to reductions in other digital expenditures or reductions in other spending to support the higher digital budget.
Of the €140 billion in costs, €108 billion would be financed by these levers. Details of unspent expenditures:
- €37 billion for ESN services
- €26 billion for human resources
- €24 billion for R&D
- €18 billion for hardware purchases
- €3 billion for training
Asterès estimates that 60% of the foregone expenditures would have been directed to European suppliers or kept in-house. This yields €65 billion of crowding-out effects in Europe (60% of €108 billion). Of which €23 billion for ESN services, €17 billion for HR, €16 billion for R&D, €7 billion for hardware, and €2 billion for training.
1.4 Million Jobs by 2030
Asterès assessed the consequences of the crowding-out effect in terms of lost activity across three dimensions:
- Direct effect on businesses and public administrations
- Indirect effect on first-tier suppliers
- Spillover effect on the rest of the European economy
It translated each into revenue, value added, and employment.
Regarding the direct effect, Eurostat’s national accounts were translated into foregone recruitment expenditures based on the average wage in the European economy, then into value added via labor productivity and into turnover via the value-added rate.
For the indirect effect, the foregone spending by users corresponds to a loss of turnover for their suppliers. That loss was then translated into value added and employment using sectoral ratios from the national accounts.
For the induced effect, Asterès used an impact model based on OECD input-output tables.
Bottom line: price increases could cost the European economy €107 billion of value added per year (about 0.6 percentage points of GDP). And 1.4 million jobs by 2030. More precisely:
- Turnover: €44 billion direct losses, €48 billion indirect, €113 billion in the cascade
- Value added: €22 billion direct, €27 billion indirect, €58 billion in the cascade
- Jobs: 311k direct, 318k indirect, 737k in the cascade
Three Alternative Scenarios
The study only delves into the baseline trajectory where inflation continues to accelerate at the pace observed over the past six years, with modest offset from productivity gains. Nevertheless, Asterès identifies at least three other plausible scenarios.
In one scenario, a productivity gain of 0.8% per year (roughly two days’ work) driven by AI would offset the price increases. This rate matches the upper bound of academic estimates, which range from less than 0.1% to 1.3%.
Another possibility: a deflationary pressure. It could result from the entry of an actor determined to undercut prices, as Free did in mobile (prices fell by 45% in five years). Or from the diffusion of GenAI, in a classic “creative destruction” scenario, to borrow Asterès’ phrasing. Users would benefit from cost reductions and could reinvest the gains, but the software sector would contract sharply.
A third scenario, already mentioned in the 2025 study: a proactive European policy to support the digital industry. It would limit eviction effects on the European economy. For example, a 15-point increase in the market share of European suppliers could preserve 120,000 jobs and €9 billion of value added per year.
