Gokul Hariharan, co-head of TMT industry research at J.P. Morgan Chase Asia, said that token consumption for AI applications is growing rapidly, while the willingness of large technology companies to invest in AI has not changed. They have the ability to use operating cash flow or seek financing to support AI investments, so related demand is strong. However, the supply chain is relatively slow to respond, and the investment bias is cautious, and the shortage of components such as memory is likely to continue until early 2027. He added that in the non-AI sector, demand for related terminal products is recovering slowly, such as consumer electronics products, automobiles, etc., but their supply chains are being squeezed by the AI supply chain, which in turn causes prices to rise, and the price increase trend is likely to continue for most of next year.

Gokul Hariharan, co-head of TMT industry research at J.P. Morgan Chase Asia, said that token consumption for AI applications is growing rapidly, while the willingness of large technology companies to invest in AI has not changed. They have the ability to use operating cash flow or seek financing to support AI investments, so related demand is strong. However, the supply chain is relatively slow to respond, and the investment bias is cautious, and the shortage of components such as memory is likely to continue until early 2027. He added that in the non-AI sector, demand for related terminal products is recovering slowly, such as consumer electronics products, automobiles, etc., but their supply chains are being squeezed by the AI supply chain, which in turn causes prices to rise, and the price increase trend is likely to continue for most of next year.