CITIC SEC Technology Sector Investment Strategy for the second half of 2026: If liquidity shocks lead to a valuation correction, the opportunities outweigh the risks.

CITIC SEC releases the investment strategy for the second half of 2026 in the technology sector. Global technology stock valuations are high, and crowded positions are the current biggest challenge. Additionally, the potential listings of SpaceX, Anthropic, and OpenAI in the next six months to a year could impact market liquidity, making it a significant event and observation window for global technology stocks. Considering the possibility of a marginal tightening of macro liquidity in the second half of this year compared to the second half of 2025, market interpretation will rely more on the continuous realization and validation of performance. However, as long as the industrial trend remains unchanged, if liquidity shocks lead to a valuation correction, the opportunities will outweigh the risks. Continued optimism in: 1) Model iterations leading to application explosions and continuous growth in ARR providing investment opportunities for base model companies; 2) Semiconductor equipment companies benefiting from the continuous expansion of domestic storage and advanced process production, as well as advanced process companies with the potential to breakthrough and increase operational efficiency; 3) Leading companies in cloud, storage, optics, CPUs, PCBs, power supplies, etc., where prices are rising due to exploding computing demand, with supply chain stability; 4) Siasun Robot & Automation and related companies in the autonomous driving sector, where technology is evolving rapidly and entering a period of capitalization with mass production imminent; 5) Undervalued stocks in the Hang Seng Technology Index, such as NETDRAGON, offering configuration value; 6) Investment opportunities in AI services and data governance companies due to the implementation of enterprise-level AI. Key points from CITIC SEC: Market review: The profitability-driven characteristics of the technology sector have been established, but absolute valuations and fund allocations are at high levels. According to Wind data, the Sci-Tech 50 Index has risen by 31.69% since the beginning of the year, significantly outperforming the Shanghai Composite Index by 6.9%, with a 25.05% increase in April leading the market. In comparison, the hardware side of electronics and communications have risen by 41.17% and 52.39% respectively since the beginning of the year, while software areas like computers and media have only seen increases of 8.63% and 1.46%. Progress brought by AI Coding has consistently exceeded market expectations, driving the continuous momentum of tech stock trading. The rise of the Sci-Tech 50 and the Nasdaq index since the beginning of 2026 has been mainly driven by systematic upward revisions of EPS, with PEs only slightly increasing by 4% and remaining stable, as the logic of valuation expansion has shifted to profit realization. From a trading perspective, the TMT sector holds a 36.4% weight in active equity funds, with performance-wise, CITIC's electronics industry saw a 47.4% year-on-year increase in net profit in Q1, and leading companies in communications also exhibited high performance growth. Computing power: Domestic computing power represented by domestic models and chips presents the biggest opportunity, with overseas demand continuing to rise. 1) Domestic computing power will become a major growth opportunity, with accelerated production expansion in advanced logic and storage expected in 2026, leading to a potential increase in the domestic production rate of devices from 30% to 40%. The full year order growth rate for storage equipment companies is expected to be revised upward to 50%. At the same time, a surge in Token usage has caused a shortage of domestic and overseas computing power, prompting domestic large-scale models to actively embrace local chips such as Ascend and Cambricon, driving the acceleration of domestic computing power. Overall, the expansion of domestic wafer fabs, the increase in domestic production rates, and the surge in AI computing demand will create a triple catalyst, with each link in the domestic supply chain expected to experience continued growth. 2) Looking at the demand for computing power: In 2026, under the drive of Agents, the rapid increase in global Token usage will continue to bring about incremental growth in Capex. Economic conditions in North America are expected to continue to improve: Q1 revenues of the three major cloud companies in North America accelerated, with Google Cloud Platform (GCP) experiencing a year-on-year surge of 63%, while Meta, Google, and Amazon have revised their 2026 Capex guidance to $125-145 billion, $180-190 billion, and approximately $200 billion, respectively. In comparison, the absolute values and growth rates of Capex for Chinese cloud companies Alibaba, Tencent, and Baidu in 2026 are significantly lower than those in North America, with Capex-to-revenue ratios ranging from 11% to 17%, indicating ample payment capabilities and significant room for future Capex increases. 3) Overseas computing power conditions are expected to continue to improve, with a sustained wave of price increases. Core industries like optical modules, storage, PCBs, and others are experiencing further tightening supply and demand relationships, with leading companies showing high growth rates and increasing long-term contract constraints. The absolute scale increase in the AI market has driven an improvement in economic conditions and a trend of ongoing capacity scarcity extension upwards. Models and applications: Increased frequency of model iterations, with Chinese model manufacturers accelerating their pace. The increase in Anthropic Token usage has led to rapid revenue growth, with ARR surpassing $44 billion as of May 2026. In addition, in Q1 2026, model manufacturers at home and abroad have broken the quarterly iteration rhythm, with OpenAI continuously iterating on the GPT-5 series, while Google and Anthropic have simultaneously released flagship versions like Gemini 3.1, Claude Opus 4.6/4.7. Meanwhile, Chinese model manufacturers have released heavyweight versions such as DeepSeek V4, Qwen3.6, Dou Bao 2.0, intensifying competition in State-of-the-Art (SOTA) models. Looking ahead, Chinese model manufacturers will accelerate their catch-up efforts, and as AI applications deepen and Chinese enterprises become more digitally intelligent, the deep integration of Chinese models and industries will create greater industrial value. Similarly, new propositions and sub-directions such as Physical AI, World Models, and Video Generation Models have entered the early stages of commercialization and are expected to open up a new wave of commercialization in AI beyond the baseline models. Risk factors: Changes in geopolitical conditions may prevent Mainland Chinese IC design companies from using advanced process capacity in overseas wafer foundries for chip fabrication. Additionally, AI applications seen so far have only yielded economic benefits in limited areas, with larger-scale applications depending on the fusion process of AI large models with existing digital infrastructure. Investments in AI by enterprises are therefore subject to potential cyclical fluctuations, indicating the need for investors to fully consider the risks that such investment cycles may bring to the technology industry. CITIC SEC also highlights the risks of macroeconomic recovery progress falling short of expectations, related industry policies not meeting expectations, core technologies of companies not progressing as expected, AI application deployment not meeting expectations, and cloud provider capital expenditures falling short of expectations.