Authors: Advait Jayant, Matthew Sheldon, Sungjung Kim and Swastik Shrivastava

Compile: BeWater

With the recent launch of the lightweight Llama 1B and 3B parameter models optimized for device-side application scenarios by Meta, and the upcoming release of its new products by Apple Intelligence at the end of October, we believe that edge AI and device-side AI will become the biggest topics of 2025.

Peri Labs and BeWater have collaborated to release a report of approximately 250 pages, covering:

• The necessity of Edge AI
• Core innovation in the field of edge AI
• Why edge AI needs encryption technology
• Understanding the core framework of edge AI
• The current status of Edge AI and encryption technology

BeWater has translated this report into Chinese, and the essence of the summary is as follows:

The Rise of Edge AI

Edge AI is revolutionizing the field of artificial intelligence by shifting data processing from centralized cloud servers directly to local devices. This approach addresses the limitations of traditional AI deployment, such as high latency, privacy issues, and bandwidth constraints. By enabling real-time data processing on devices like smartphones, wearables, and internet of things sensors, Edge AI reduces response time and securely stores sensitive information on the devices themselves.

Advancements in hardware and software technology have made it possible to run complex AI models on resource-constrained devices. Innovations such as dedicated edge processors and model optimization techniques make device-side computing more efficient without significantly impacting performance.

**Key point 1: The rapid rise of AI has already surpassed Moore's Law.

Moore's Law states that the number of transistors on a microchip doubles approximately every two years. However, the rise of AI models has exceeded the speed of hardware improvements, resulting in a widening gap between computing demand and supply. This gap makes collaborative design of hardware and software essential.

Point 2: Major industry giants are increasing their investments in edge AI and adopting different strategies.

The major industry giants are making massive investments in edge AI, recognizing its ability to completely transform fields such as healthcare, autonomous driving, robotics, and virtual assistants by providing instant, personalized, and reliable AI experiences. For example, Meta recently released models optimized for edge devices, and Apple Intelligence will also launch its edge AI technology at the end of October.

The intersection of edge AI and encryption technology

Point 3: Blockchain provides a secure and decentralized trust mechanism for edge AI networks.

The Block chain ensures the integrity and tamper resistance of data through its immutable ledger, which is particularly critical in Decentralization networks composed of edge devices. By recording transactions and data exchanges on the Block on-chain, edge devices can securely perform identity verification and authorization operations without relying on centralized institutions.

Point 4: encryption economic incentive mechanisms promote resource sharing and capital expenditure

Deploying and maintaining edge networks requires a substantial amount of resources. The encryption economic model or Token incentives can be used to provide Token rewards, encouraging individuals and organizations to contribute computing power, data, and other resources to support the construction and operation of the network.

Key Point 5: The Decentralized Finance model promotes efficient allocation of resources

By introducing concepts such as stake, lending, and liquidity pools in Decentralized Finance, the Edge AI Network can establish a market for computing resources. Participants can provide computing power through stake tokens, lend out excess resources, or contribute to a shared pool to receive corresponding rewards. Smart contracts automatically execute these processes to ensure fair and efficient allocation of resources based on supply and demand, and implement dynamic pricing mechanisms in the network.

Key Point 6: Trusted Decentralization

In a Decentralization edge device network, establishing trust without centralized supervision is a challenge. In an encryption network, trust is achieved through mathematical means; this computation and math-based trust is the key to facilitating trustless interactions, a characteristic that AI does not currently possess.

Future Outlook

Looking ahead, there are still plenty of opportunities for innovation in the field of edge AI. We will see edge AI become an indispensable part of our lives in many application scenarios, such as highly personalized learning assistants, digital twins, autonomous vehicles, collective intelligence networks, and emotional AI companions. We are excited about the future!