MAIN ARTICLE
Long-term conviction meets nervous markets
A retreat in AI-related equities, including pronounced weakness in parts of China’s technology sector, has reopened the question of whether enthusiasm outran fundamentals. In Hong Kong, Solomon focused on a longer horizon: the deployment of AI inside companies remains far from complete. He acknowledged that advances will bring setbacks, but distinguished weekly price movements from productivity gains that may unfold over many years.
Pick offered a similarly expansive view of demand for computing. Bloomberg reported that he envisaged cumulative spending of as much as $15 trillion over the coming decade. That figure represents an estimate of future demand for chips, servers, data centers and electricity, not a guarantee that every participant will earn an attractive return. The gap between technological possibility and the price investors pay for it lies at the heart of the market’s discomfort.
The infrastructure bill must ultimately meet a revenue test
Behind an apparently weightless AI service sits expensive physical infrastructure. Goldman Sachs research published in August projected that worldwide AI investment would exceed $1 trillion in 2026, with around $581 billion in the United States.













