The Director of the Monetary and Capital Markets Department of the IMF, Tobias AdrianAnd Pierre-Olivier Gourinchaswho directs research within the institution, delivered a particularly interesting critical analysis of the dynamics of global growth, and the prospects outlined for 2026.
They underline the unexpected nature of current economic resilience, in an environment marked by the multiplication of trade tensions and by geopolitical fragmentation on a scale not seen in several decades.
This apparent solidity is due to the dynamic instilled by AI. Investment in tech today plays a role as a major macroeconomic shock absorber and compensates for the persistent weakness of the manufacturing industry. A positive effect which nevertheless constitutes a significant point of structural fragility.
In two years, the rapid rise of AI has profoundly changed the composition of productive investment. In the United States, IT spending now represents the highest share of GDP observed since the early 2000s. This dynamic contrasts with the sluggishness of the rest of investment, particularly industrial, and largely explains the ability of the American economy to absorb the shock of trade restrictions and geopolitical tightening. The ripple effect goes beyond the national framework. Technology value chains irrigate Asia, particularly economies exporting advanced components and equipment, strengthening the sector’s contribution to global growth.
This stabilizing role is based on a widely shared belief in the transformative potential of AI. Companies anticipate significant productivity gains, increased automation of complex tasks and lasting improvement in margins. These expectations have supported the rise in financial markets since the end of 2022, in parallel with the diffusion of the first large-scale generative AI tools. The stock market valuation of the technology sector has increased rapidly, facilitating access to financing and accelerating investment programs.
This dynamic is, however, not without its counterparts. As the expansion continues, debt financing takes on an increasing role, increasing the financial leverage of technology companies. The future profitability of these investments depends closely on the materialization of the expected productivity gains, but also on more prosaic industrial parameters. The rate of obsolescence of computing infrastructures, the frequency of renewal of advanced processors and depreciation choices have a direct impact on margins. An intensification of technology upgrade cycles could mechanically compress results and increase financing needs.
The comparison with the period of the Internet bubble is naturally obvious, but it must be handled with caution. IT investment today represents a share of GDP comparable to that observed at the end of the 1990s, but its rise has been more gradual. Stock market valuations have certainly increased rapidly, but price-to-earnings ratios remain more contained, due to high real profits. However, systemic risk has not dissipated, but it has changed in nature.
The first source of vulnerability is the extreme concentration of stock market performance. A growing share of indices is now carried by a reduced number of players directly exposed to AI. This concentration increases the sensitivity of the markets to any revision of sectoral expectations. A targeted correction could quickly spread to the entire market.
The second weakness lies in the financing structure. Many strategic AI players are not listed, but rely heavily on debt to finance their infrastructures. This segment partly escapes traditional market discipline mechanisms. A deterioration in financial conditions could weaken these companies and generate tensions which would not immediately manifest themselves on the stock markets, but would weigh on the financial system as a whole.
Finally, the size reached by financial markets in relation to the real economy amplifies the risks of macroeconomic transmission. In the United States, market capitalization now far exceeds the level observed in the early 2000s as a proportion of GDP. In this context, even a moderate correction in technology assets could have a significant impact on consumption and investment, well beyond just the sector concerned.
AI thus acts as a short-term stabilizer, supporting investment and confidence in a degraded environment. But it simultaneously concentrates risks, geographically, sectorally and financially. This duality explains why the current resilience of global growth must be interpreted with caution. It is based less on a generalized diffusion of economic dynamics than on the performance of a technological core whose promises remain, in part, to materialize.
In a world marked by geopolitical tensions and their commercial impacts, this concentration increases the vulnerability of the global cycle. The trajectory of AI will not only determine the future of productivity, it will also condition the capacity of the global economy to prevent its main cyclical shock absorber from becoming, in the long term, a factor of systemic destabilization.
