(Sept 25): The global market for artificial intelligence (AI)-related products is ballooning and will hit as much as US$990 billion (RM4.09 trillion) in 2027, as the technology’s quick adoption disrupts companies and economies, Bain & Co said.

The market, including AI-related services and hardware, will grow 40% to 55% annually from US$185 billion last year, the consulting firm said in its fifth annual Global Technology Report released Wednesday. That will lead to revenue of US$780 billion to US$990 billion, Bain said.

The growth will be fuelled by bigger AI systems and larger data centres to train and run them, driven by companies and governments using the technology to boost efficiency. Demand is rising so fast that it’ll strain supply chains for components, including chips needed to run the services, Bain said. Combined with geopolitical tensions, rising sales could trigger shortages in semiconductors, personal computers and smartphones, Bain warned.

Demand for upstream chip components such as integrated circuit design and related IP could rise 30% or more by 2026, putting pressure on manufacturers, Bain said. The cost of larger data centres could jump from US$1 billion to US$4 billion now to between US$10 billion and US$25 billion in five years, as their capacity expands to more than a gigawatt from 50–200 megawatts currently, it said.

“These changes are expected to have huge implications on the ecosystems that support data centres including infrastructure engineering, power production, and cooling,” the consultancy said in a statement.

Companies are moving beyond an experimentation phase and beginning to scale generative AI across their operations, Bain said. Small language models, similar to the large language models that led to the creation of OpenAI’s ChatGPT chatbot, but lightweight and efficient, could be favoured by enterprises and countries amid concerns surrounding costs and data privacy.

Governments including Canada, France, India, Japan and the United Arab Emirates are spending billions of dollars to subsidise sovereign AI, investing in domestic computing infrastructure and AI models created within their borders and trained on native data. But establishing successful sovereign AI ecosystems will be time-consuming and expensive, according to Anne Hoecker, head of Bain’s Global Technology practice.

Uploaded by Chng Shear Lane