HTX Research's latest report interprets US stock AI assets: technology is still in its early stages, while capital expenditure and valuations have entered the later stage of the cycle

Recently, HTX Research, a research department under Huobi HTX, released the latest research report titled "The Industrialization of Intelligence and the Bubble Cycle: Token Economics, Capital Expenditure, and the Revaluation of AI Assets in the US Stock Market." The report analyzes the valuation status, cyclical position, and risk-return structure of various segments of the AI industry chain from the perspective of token economics. The report argues that the AI industry and AI stocks are currently not in the same time cycle. The technological diffusion of large models, agents, and multimodal products is still in its early stages, while the corporate application cycle is in the early to mid-stage; however, the capital expenditure cycle of hyperscalers has already entered the later stage, with stock valuations, market concentration, and trading sentiment closer to the latter half of a bull market.

Variables Determining Stock Prices

Over the past three years, the market first traded the supply scarcity of GPUs, HBM, servers, networks, and data centers, followed by the capabilities leap of frontier models, inference models, and coding agents. After entering 2026, the core variables determining stock returns are shifting from model parameters and capital expenditure scale to the production costs of tokens, task success rates, user consumption intensity, corporate workflow penetration rates, and whether massive AI investments can ultimately be converted into free cash flow.

This shift is backed by a significant change in the scale of capital expenditure. J.P. Morgan Asset Management estimates that the capital expenditure of five US hyperscalers will be approximately $69.7 billion in 2026, with AI capital expenditure accounting for an estimated 93% of these companies' operating cash flow, up from about 33% in 2023. When capital expenditure consumes most of the operating cash flow, the market's focus will inevitably shift from revenue growth to capital return rates.

The Bubble Exists in Financial Structures, Not in the Industry Itself

The report provides a more precise breakdown of the term "AI bubble." Cloud revenue, coding agent usage, chip revenue, and corporate demand are all genuinely growing; AI technology itself is not a false narrative. However, capital expenditure, external financing, data center projects, private model valuations, and some overvalued second-tier targets have already shown clear signs of bubble characteristics.

The method for assessing the cost-effectiveness of targets also needs to be adjusted. Surface-level price-to-earnings ratios cannot directly represent true valuation levels: Alphabet's P/E ratio is distorted by investment returns, and Amazon's current accounting profits do not represent normalized valuations. What truly offers cost-effectiveness are companies that best match normalized valuations, competitive barriers, cash flow, and AI optionality.

At the current price and cyclical position, the report believes that Alphabet's comprehensive odds are the most prominent, possessing both full-stack capabilities and multiple growth engines; Microsoft's fundamentals have the highest win rate, but valuation expansion space is limited; Meta's surface valuation matches its revenue growth rate well, but capital expenditure risks are significant; TSMC is the most stable "shovel seller" in the supply chain, needing to bear geopolitical tail risks; NVIDIA remains the most attractive core semiconductor target after growth adjustments; Amazon has significant AWS and Trainium optionality. Oracle and Micron are high-odds but low-win-rate cyclical assets, while AMD, Arista, and Vertiv have excellent business quality, but current prices demand near-perfect execution.

The AI Mainline is Changing the Asset Allocation of Crypto Users

The influence of AI as a common mainline in the global capital market is reflected not only in US stock pricing but also in changing the asset allocation behavior of crypto users. When core AI targets like NVIDIA, Micron, TSMC, Broadcom, Meta, and Alphabet enter the daily investment portfolios of crypto users alongside gold, oil, ETFs, and pre-IPO assets, the boundaries between the two markets begin to blur. Funds can flexibly switch between BTC, ETH, leading AI stocks, gold, and ETFs based on macro environments, industry trends, and risk preferences. An increasing number of users have begun to understand crypto and US stocks as different allocation directions within the same global risk asset system.

Huobi HTX is one of the earlier platforms systematically promoting this direction in crypto trading. According to data disclosed in August 2026, the platform's TradFi contract segment has accumulated a trading volume exceeding $2.5 billion, supporting over 170 TradFi-related assets, covering US stocks, ETFs, gold, silver, oil, AI chips, storage, aerospace, and pre-IPO thematic assets like OpenAI and Anthropic.

The key to this model lies in the platform already having a large number of crypto users who have completed registration, identity verification, and asset deposits. These users typically hold stablecoins like USDT directly, without needing to open traditional securities accounts or transfer funds to another financial system, allowing them to trade TradFi assets within the same account. When risk preferences decline, they allocate to gold, ETFs, or large tech stocks; when risk preferences rise, they increase the proportion of crypto and high beta AI stocks, with funds always remaining on the same platform.

The Competitive Boundaries of Trading Platforms are Changing

The growth of TradFi business indicates that future competition among trading platforms will expand from spot, contracts, liquidity, and listing speed to comprehensive competition around crypto, US stocks, ETFs, commodities, pre-IPO, wealth management, and AI investment tools. Platforms with true long-term competitiveness will upgrade their core capabilities from single trading ability to global asset distribution capabilities.

This also confirms a larger judgment proposed in the report: AI is changing not only model capabilities and computing demands but also the flow of funds, asset allocation methods, and the organizational forms of financial products. Huobi HTX's early layout in the TradFi direction resonates with HTX Research's ongoing tracking of the AI mainline and cross-market capital flows—identifying cyclical positions, judging capital flows, and understanding the interconnections between different assets are not only the core of investment research work but also constitute a source of first-mover advantage in actual business decisions. As AI drives the global financial market into a new phase of integration, institutions that can simultaneously understand industry cycles and capital flows are more likely to occupy advantageous positions in the next round of competition.

Note: The content of this article does not constitute investment advice, nor does it constitute any offer, solicitation, or recommendation of investment products.

About HTX Research

HTX Research is the exclusive research department under Huobi HTX, responsible for in-depth analysis of a wide range of fields including cryptocurrencies, blockchain technology, and emerging market trends, writing comprehensive reports, and providing professional assessments. HTX Research is committed to providing data-driven insights and strategic foresight, playing a key role in shaping industry perspectives and supporting informed decision-making in the digital asset space. With rigorous research methods and cutting-edge data analysis, HTX Research consistently stands at the forefront of innovation, leading the development of industry thought and facilitating a deeper understanding of the ever-changing market dynamics.