【State Council: Promote the “Internet of Everything” for smart terminals, vigorously develop smart connected vehicles, AI smartphones and computers, smart robots, smart homes, smart wearables, and other new generation smart terminals】
The State Council has issued opinions on the in-depth implementation of the “AI +” initiative. The opinions propose to promote the “Internet of Everything” for smart terminals, cultivate an ecosystem of smart products, and vigorously develop smart connected vehicles, AI smartphones and computers, smart robots, smart homes, smart wearables, and other new generation smart terminals, creating an integrated smart interaction environment that covers all scenarios. Accelerate the integration of AI with technologies such as the metaverse, low-altitude flight, additive manufacturing, and brain-computer interfaces, and explore new forms of smart products.
(Tonghuashun Red Letter News)
Why are edge computing chips considered “shovel stocks”?
1. Fundamental and Necessary (“In a gold rush, everyone needs a shovel”):
Whether it is smartphones, smart cars, AR/VR glasses, smart home devices, or various IoT devices, to become “smart” and achieve real-time response, privacy protection, low latency, and other functions, AI computing capabilities must be deployed on-device.
Edge computing chips (including CPU, GPU, NPU/TPU and other AI acceleration cores) are the “shovels” that provide this fundamental computing power. Without them, any smart terminal is just an empty shell.
2. Leverage Effect of Industry Explosion (“The more intense the gold rush, the more expensive and numerous the shovels sold”):
Currently, we are at a critical juncture where generative AI (AIGC) is penetrating into terminal devices. Large models are shrinking, striving to run on smartphones, computers, cars, and other devices.
This “edge AI gold rush” has an unprecedented demand for computing power, directly leveraging the demand for high-end edge AI chips. Chips need to support more complex models and provide higher TOPS (trillions of operations per second) computing power, which leads to an increase in chip prices and value.
3. Technical Barriers and Moats (“Building a good shovel has a high technical threshold”):
Designing an advanced SoC (system-on-chip), especially one that integrates powerful AI acceleration capabilities, has very high technical barriers. It involves cutting-edge process technologies (such as 3nm, 4nm), complex IP core integration, advanced packaging technologies, and hardware-software collaborative optimization.
This high barrier creates a deep “moat” for leading chip companies, making it difficult for newcomers to enter easily, ensuring the sustainability and profit margins of the “shovel selling” business.
1. Summary: Regardless of which smart terminal application ultimately explodes, they all rely on the most fundamental edge computing capabilities.
Key recommendations: Broadcom Integration, Rockchip, Quectel, Espressif, Yanshan Technology, Guanghetong, Meige Intelligent, Runxin Technology, Zhongke Chuangda, Hengxuan Technology,
