One picture to understand: Comparison data of the entire AI industry chain between China and the United States

Breakthrough in chip computing power technology

China: Accelerating domestic substitution, chips such as Cambrian, Huawei Ascend, and Baidu Kunlun are rapidly rising, gradually narrowing the gap with international leading levels. These domestically produced chips have demonstrated strong competitiveness in specific application scenarios.

Beauty: Holding high-end AI chips such as NVIDIA, NVIDIA’s GPU plays a crucial role in AI model training and inference, providing powerful computing power support.

Data Advantage: The Game between Quality and Scale

China: Relying on a large user base and diverse application scenarios, it has the world’s largest data pool. The massive amount of data provides rich materials for the training of AI models, promoting the rapid development and application of AI technology.

Beauty: High data quality, but strict privacy supervision. High quality data provides a solid foundation for the training of AI models, ensuring their accuracy and reliability.

Big Model Track: Dual Competition of Performance and Cost

Chinese: The catching up speed of domestic large models is astonishing, the performance gap has narrowed to 0.3%, and the cost is as low as 1/10 of GPT-4. The rapid progress of domestic large-scale models has made AI technology more popular and cost-effective.

Beauty: OpenAI and Google are still leading in the field of big models. Models such as GPT-4 have significant advantages in performance and wide applicability.

Application Expansion

Chinese: AI is penetrating into hardware and terminal scenarios, such as AI phones, AI PCs, AIoT, etc.

Beauty: AI is widely used, especially in fields such as autonomous driving, smart speakers, and smart homes.

Industrial changes

China: The combination of AI and innovation has exploded, and fields such as content creation, e-commerce retail, healthcare, and intelligent manufacturing are actively exploring the possibility of “AI+”.

The United States: The maturity of AI applications in fields such as finance, healthcare, and education is relatively high, such as Palantir’s applications in data analysis and defense.

opportunity

Semiconductors: Micron (MU), TSM, Nvidia (NVDA), etc.

Server: Credo Technology (CRDO), Dell, Supermicro Computer (SMCI), etc.

Cloud computing: AWS (AMZN), Azure (MSFT), Google Cloud (GOOG), etc.

Data centers: Equinix (EQIX), Digital Realty (DLR), etc.

Software: Palantir (PLTR), Meta (META), Snowflake (SNOW), etc.