Anisuzzaman Chowdhury is an experienced businessman with interests in various industries including pharmaceuticals, construction and energy. This article will look at artificial intelligence (AI) and its impact on corporate operations, automating processes while simultaneously enhancing decision-making capabilities.
Over the course of the last three decades, the World Wide Web has advanced exponentially. Seen as little more than a super-charged fax machine at its launch, the internet has gained trillions of dollars in value, with its meteoric rise today recognised as a game-changing technology supercycle that has transformed businesses in an array of market sectors, impacting daily life for millions of people all over the world. Just as the internet’s impact was underestimated at the outset, experts warn of the risk of miscalculating the transformative power of the ongoing AI supercycle.
AI is transforming not just businesses but entire industries as companies deploy systems to act with more autonomy, rethinking software, operating models and data governance to compete in the digital world. Businesses of all sizes are grappling with the next phase of the AI supercycle.
As models continue their rapid advancement, AI has reached a new inflection point, with companies clamouring to deploy AI at scale. Agentic AI systems have come into their own in terms of planning and executing complex tasks, in turn driving a sharp increase in consumption of compute and tokens while imposing new demands on the software, infrastructure and data required to support AI’s massive expansion. The attached infographic features some interesting statistics on agentic AI deployment in the United States in 2025.
As impressive as AI agents are, with one MIT study reporting productivity gains of circa 60%, researchers suggest that they are just part of the AI supercycle, with the next frontier moving from AI on digital devices into the real world. Early signals of this phase include Amazon’s industrial robots, Google and Meta’s innovations with smart glasses, and Waymo’s robotaxis. In terms of infrastructure, this shift requires a ‘future-back’ perspective of AI demands, recognising that AI is already outgrowing the infrastructure created for the internet in terms of connectivity and computational power.
AI models are trained and deployed at data centres. As these vast facilities grow larger and denser, they are creating unprecedented power demands. With AI workloads already consuming tens of gigawatts globally, experts warn that by the end of the decade consumption could approach hundreds of gigawatts, with AI pushing against the limits of existing power infrastructure. The embedded video delves deeper into this issue, exploring moves by the United States government to increase power supplies for AI infrastructure.