As of April 20, 2026, 09:57, the National Index for Information Technology Innovation (CN5075) surged by 1.98%, and the Innovation ETF E Fund (159540) rose by 2.13%.
As of 10:06, the CSI Artificial Intelligence Theme Index (930713) surged by 1.92%, and the Artificial Intelligence ETF E Fund (159819) increased by 1.99%, marking a third consecutive rise with a transaction volume of 329 million yuan.
As of April 17, the Artificial Intelligence ETF E Fund (159819) has seen a scale increase of 964 million yuan since the beginning of this month, achieving significant growth.
A report released by International Data Corporation (IDC) in 2025 indicated that global AI computing power expenditure is expected to exceed 130 billion USD by 2026, with the inference segment accounting for over 65%, becoming a core part of data center computing power consumption. In the second quarter of 2026, the supply-demand mismatch in the AI industry infrastructure chain will continue to worsen, highlighting the upward cost pressure across the entire computing power industry chain. On the computing power side, global high-end GPU rental prices have skyrocketed, with the spot price of NVIDIA's next-generation Blackwell chip soaring by 48% in two months to 4.08 USD/hour, and the contract price for H100 has increased by nearly 40% over six months; domestic cloud vendors have raised prices intensively, with Alibaba Cloud and Baidu Smart Cloud's AI computing power products seeing maximum increases of 34%. Major model vendors like Zhipu AI have frequently adjusted prices in the short term, making high-end computing power hard to obtain, with orders already scheduled until after 2027.
Guotai Junan Securities pointed out that in the AgenticAI inference era, a large number of CPU-centric tools have been introduced into the execution process, resulting in a high proportion of time consumption on the CPU side. The paradigm of AI workloads is evolving from simple text generation to complex agents and reinforcement learning, and the insufficient number of CPUs will lead to wasted GPU resources. The ratio of CPUs to GPUs may increase, and the demand for CPUs may continue to rise.
GF Securities believes that domestic AI large models are facing development opportunities. Anthropic's recent implementation of an identity verification mechanism, combined with its previous service restrictions on companies with Chinese capital backgrounds, indicates that the supply constraints of overseas large models to China are further strengthening. In the short term, this will raise the threshold and uncertainty for Chinese users to use overseas models; in the medium to long term, it will force domestic companies to accelerate their transition to local models and local computing power systems. This incident will continue to strengthen market expectations for the limitations of advanced overseas models and the acceleration of self-control, and continue to be optimistic about the industrial opportunities of domestic large models and AI computing power infrastructure.
The Innovation ETF E Fund (159540) closely tracks the National Index for Information Technology Innovation, which selects listed company securities from the Shanghai and Shenzhen North Exchanges that belong to the fields of basic hardware, basic software, application software, information security, and other information technology innovation themes as samples, to reflect the price changes of securities of listed companies in the information technology innovation theme, providing corresponding investment tools for investors
The E Fund Artificial Intelligence ETF (159819) closely tracks the CSI Artificial Intelligence Theme Index, which selects 50 listed company securities involved in providing fundamental resources, technology, and application support for artificial intelligence as index samples, reflecting the overall performance of listed companies in the artificial intelligence theme.
The E Fund Innovation ETF (159540), off-exchange connection (Class A: 020403; Class C: 020404).
The E Fund Artificial Intelligence ETF (159819), off-exchange connection (Class A: 012733; Class C: 012734)

159819

515070

159248

159540

As of April 20, 2026, the Xinchuan ETF E Fund rose by 2.13%, and the Artificial Intelligence ETF E Fund increased by 2%. The CNI Information Technology Innovation Theme Index rose by 1.98%, while the CSI Artificial Intelligence Theme Index increased by 1.92%. AI computing power expenditure is expected to exceed USD 130 billion by 2026, with increasing cost pressures across the entire computing power industry chain. Domestic cloud vendors have raised prices, with AI computing power products increasing by 34%. Guojin Securities and GF Securities are both optimistic about the development opportunities for domestic AI large models

腾讯新闻 - 财经

E Fund Artificial Intelligence ETF

ChinaAMC CSI Artificial Intelligence ETF

Wanjia CSI Artificial Intelligence ETF

E Fund CNI Information Technology Innovation Theme ETF

NVIDIA

Alibaba

BABA-W

Baidu

BIDU-SW

KNOWLEDGE ATLAS

SINOLINK SECURITIES

GF SECURITIES

GF SEC

NVIDIA Corporation

BABA-WR

Alibaba HK SDR 5to1

BIDU-SWR

The Xinchuan ETF E Fund rose by 2.13%, while the Artificial Intelligence ETF E Fund surged by 2%. Institutions are optimistic about the opportunities in domestic large models and AI computing power infrastructure industry