HTX Research, the dedicated research arm of HTX, has published a report arguing that AI technology remains in the early stages of diffusion even as capital expenditure, valuations and investor sentiment have moved significantly ahead. J.P. Morgan Asset Management estimates that five U.S. hyperscalers will spend approximately $697 billion in 2026, lifting capital expenditure from about 33% of operating cash flow in 2023 to an estimated 93%. The report says markets are shifting their focus from GPUs, model scale and spending growth toward token production costs, task-completion reliability, usage intensity, enterprise adoption and durable free cash flow. It describes AI demand and infrastructure growth as real, but says speculative risk is concentrated in financing structures, data-center projects, private-model valuations and some high-multiple second-tier equities. Alphabet is identified as offering the most compelling overall asymmetry at current prices and cycle positions, with the framework also applied to Microsoft, Meta, TSMC, NVIDIA, Amazon, Oracle, Micron, AMD, Arista and Vertiv. The report also says AI is influencing crypto investors' allocation across a wider global risk-asset system. HTX disclosed that cumulative trading volume in its TradFi perpetuals (derivative contracts tracking traditional assets) section exceeded $2.5 billion in August 2026, with more than 170 related assets available, including U.S. equities, ETFs, commodities, AI semiconductors, memory, aerospace and pre-IPO themes such as OpenAI and Anthropic. HTX Research says competition among trading platforms is moving toward multi-asset access, wealth management and AI investment tools.