AI's true impact will become apparent in the coming year (www.economist.com)

After three years of awe around ChatGPT and the rise of generative siblings like Sora for video, the story of AI is shifting from technological wonder to real-world consequences. In 2025 large US tech firms poured roughly $400bn into data centres and other infrastructure; one estimate pegs cumulative AI-related spending at $7tn by 2030. Yet commercial returns so far are tiny by comparison—about $50bn a year—highlighting a big gap between investment and revenue and raising questions about whether the sector will deliver a sustained boom, suffer a rout, or provoke political and social backlash in 2026. That coming year matters because the heavy infrastructure build-out creates financial fragility (overcapacity and stranded assets), concentrates power in a few firms, and accelerates cultural and regulatory tensions. Expect battles over content, jobs, competition policy and national security as generative models reshape media (music, video, podcasts) and productivity. The piece signals that technical feats—advanced multimodal models and massive datacentre footprints—are no longer the main story; their economic, financial and societal ripple effects are. How investors, regulators and societies respond will determine whether AI becomes a broad-based engine of growth or a source of instability and backlash.