Temasek is betting on the next generation chip and AI model architecture, predicting that AI energy demand will undergo disruptive changes.

Singapore's state-owned investment company Temasek Holdings Pte. has pointed out that the continuous progress in the field of Artificial Intelligence (AI) is expected to significantly change the energy demand patterns of the industry and currently, this industry is putting increasing pressure on the global power grid. Russell Tham, head of the Global Investment Emerging Technology Department at Temasek, stated that with the application of more efficient AI architectures, combined with continuous innovations in materials science and chip manufacturing processes, the overall impact of the AI industry on the environment is expected to be greatly reduced. Tham stated at the Bloomberg Sustainable Business Summit held in Singapore on Tuesday, "In the future, there may be a radical change in the energy structure required to generate AI tokens." He further pointed out that the existing AI architectures, "though impressive and developing at a rapid pace, are fundamentally inefficient architectures." Earlier this year, Temasek had indicated that it might be difficult to achieve its goal of halving its portfolio's carbon emissions by 2030 compared to 2010 levels, partly due to the continuous increase in energy demand from the AI industry. The strain on the power grid and the rising electricity prices are prompting data center operators to actively explore new power supply solutions. It is predicted that if current trends continue, by 2035, the electricity consumption of US data centers will account for about 20% of the total national electricity consumption, far higher than the current 5.9%. Companies like CuspAI in the UK, funded by investors like Temasek and Bezos Expeditions, are working on improving semiconductor production processes to reduce or even replace the use of some rare metals. Tham revealed, "We have invested in highly innovative materials and chip architectures that are significantly more energy efficient than current technologies. At the same time, we have also laid out some promising new AI model architectures that have not been fully validated but have the potential to achieve a significant increase in energy efficiency."