This could happen, for example, when more members of the baby-boomer generation retire or if there is a sharp upturn in unemployment if many jobs are displaced by technology, leading to a reduction in contributions to retirement funds and/or to emergency withdrawals. In an alternative scenario, investor sentiment could turn against leading AI companies, if doubts take hold over whether the huge investments in AI capital expenditure (capex) will pay off. Total spending on AI worldwide is estimated at $1.5 trillion in 2025 and is projected to rise to $2 trillion in 2026, with the main segments being generative AI (genAI) smartphones, AI-optimized servers, AI services, AI application software, AI processing semiconductors and AI infrastructure software.68 The data centre capex of the top eight US hyperscalers (very large cloud services providers) alone amounted to $258 billion in 2024 and is projected to more than double to $525 billion in 2032.69 However, current and future revenues linked to these AI capex investments are difficult to estimate; there may ultimately be many losers alongside a few winners. Some companies will be undercut by providers of similar services at cheaper prices, while others may find that some key technological inputs, notably graphics processing units (GPUs), become quickly outdated. The vulnerability of the companies that are investing heavily today will depend not only on the revenues that materialize, but also on how they have financed their outlays. The largest hyperscalers have until recently drawn heavily on their own cash. But increasingly the AI buildout is also being financed via relatively opaque special- purpose vehicles and/or with debt.70 It is possible that the strategic decisions made by today’s leading technology companies will pay off, particularly with support from governments, given AI’s strategic geopolitical value and the vast opportunities across sectors. However, if investor concerns about funding mechanisms and debt levels start to outweigh excitement about uncertain future revenues, that could trigger an asset bubble burst. Other possible triggers to watch for include a societal backlash against the AI buildout; for example, if concerns emerge around data centre water usage,71 unemployment, or, more broadly, inequality. Longer term, quantum technologies could potentially upend entire data centre-based business models. Boomerang inflation According to the IMF, inflation is projected to fall to 4.2% globally in 2025 and to 3.7% in 2026, albeit with above-target inflation in the United States and subdued inflation in most other countries.72 In the immediate term, inflation is thus expected to remain largely under control, although the figure masks an acute cost-of-living crisis in many countries following the significant global inflation spike in 2021–2022. There are several risks that could worsen the inflation outlook. Rising prices of natural resources if geoeconomic confrontation intensifies are of concern. Further, the inflationary pressures associated with higher tariffs should not be underestimated. Sustained, broad tariffs could lead to widespread inflationary pressures, particularly for the United States and closely linked economies including Canada and Mexico.73 Uncertainty is the defining feature of the outlook; specific policy design and the level of sector-specific targeting of tariffs are critical in determining inflationary impacts. Another source of inflation risk may emerge from disruptive paradigm changes in monetary policy. As governments seek ways to stimulate growth and manage growing debt servicing burdens, some may also increase pressure on central banks to run more accommodative monetary policies. Central- bank independence could be further eroded in Falco Negenman, Unsplash Global Risks Report 2026 44