Uneasy Acceleration

The global artificial intelligence boom has entered a phase that feels both historic and unsettling. On one hand, the technology is maturing at a pace few could have imagined even three years ago. On the other, the money rushing into the sector ~ across chips, models, data centres, and speculative partnerships ~ has begun to resemble the financial manias that have punctuated earlier waves of technological change. When industry leaders themselves caution that elements of the current frenzy veer into irrationality, it is worth stepping back to examine the trajectory. The core tension lies in AI’s dual identity as both a general-purpose technology and a speculative magnet.

The former is undeniable: AI is already reshaping creative work, logistics, healthcare, and education. The latter is visible in the staggering valuations placed on companies whose revenues remain dwarfed by the scale of investments pledged in their name. In some corners of the sector, the ratio of hype to fundamentals mirrors that of the late-1990s internet surge. That episode ultimately led to a painful correction, even though the underlying technology went on to transform the global economy. Drawing that parallel does not imply a perfect replay, but it does signal that exuberance and disillusion often travel together in the early years of powerful innovations. Yet one must also acknowledge the real strategic competition unfolding among the world’s largest technology companies.

Advertisement

The players that control the full chain ~ from semiconductor design to proprietary datasets, from massive compute clusters to consumer-facing platforms ~ enjoy a structural advantage that can buffer them from market volatility. Their long-term horizon allows them to invest in frontier science and infrastructure even if the broader market cools. Smaller firms, by contrast, depend on venture capital cycles and are far more vulnerable to sudden contractions. Beyond questions of valuation, the physical footprint of AI is emerging as a challenge. The energy required to train and deploy the latest models is no longer a marginal issue. AI systems now consume a measurable share of global electricity, and the expansion of data centre capacity will intensify this demand. Governments eager to host AI investment must grapple with the reality that inadequate energy infrastructure can become a hard brake on digital ambition.

This introduces a paradox: societies want the productivity gains promised by AI, yet the energy required to power these gains risks delaying climate targets and straining power grids. The disruption to labour markets adds another layer of complexity. Jobs are unlikely to vanish, but few professions will remain untouched. The workers who learn to incorporate AI tools into daily practice will thrive; those who resist adaptation may find the ground shifting under their feet. This transition will require thoughtful policy to ensure that opportunities outweigh dislocations. What is unfolding is not simply a technology cycle but the early architecture of a new economic era. The turbulence may be unavoidable, but so too is the need for sober governance and realistic expectations.