PM Anwar says approvals for new non-AI data centres are halted due to energy and water consumption concerns, with focus on high-value tech.
PETALING JAYA: The government has stopped approvals for new data centres unrelated to AI amid rising concerns over energy and water consumption, said Prime Minister Datuk Seri Anwar Ibrahim.
He said projections from the Energy Transition and Water Transformation Ministry indicate sufficient supply for AI-related facilities over the next one to two years.
“Energy and water consumption have risen significantly. About one and a half to two years ago, the government limited approvals for new data centres not related to AI or high-value technology.
“New applications that take advantage of lower water and energy costs have been halted,” he said during the Prime Minister’s Question Time in Parliament yesterday.
He was responding to Besut MP Datuk Che Mohamad Zulkifly Jusoh, who raised concerns that rapid AI growth has increased electricity and water usage, citing US statistics showing that data centres consume around 4.4% of national electricity and millions of litres of water for server cooling.
Anwar said short-term energy needs can still be met but the potential strain on the national grid and electricity tariffs must be taken seriously.
He highlighted long-term initiatives, including the Asean Power Grid linking Vietnam, Peninsular Malaysia and Singapore, and plans to channel energy from Sarawak to Peninsular Malaysia and Singapore.
He said while carbon footprint limits may not be directly included in the proposed AI Governance Bill, existing energy regulations could address these concerns.
On AI governance, Anwar said the Bill would prioritise ethical use, accountability and protection of citizen rights.
“In our national context, issues of citizen rights and moral and ethical considerations will be given special attention. This may differ from Western approaches, which focus more on technical and legal aspects.”
The Bill, in early drafting under the Digital Ministry, aims to cover the full AI system life cycle, from development and training to deployment, monitoring and risk management.
It would also define accountability measures and clarify the roles of developers and deployers.
He said the Bill would include a risk and classification framework, integrating harm and incident reporting to balance innovation with ethical standards and avoid overburdening local MSMEs.
