Top 10 Predictions for Encryption AI in 2025

With the rapid development of the AI industry, the integration of encryption and AI has also attracted much attention. A researcher focused on this field has made 10 predictions for 2025, and the following are the main points:

1. The total market value of encryption AI tokens reaches 150 billion USD

Currently, the market value of encryption AI tokens accounts for only 2.9% of the market value of altcoins, but this proportion is expected to increase significantly. AI encompasses multiple fields ranging from smart contract platforms to memes, DePIN, Agent platforms, data networks, and smart coordination layers, and its market position is expected to rival that of DeFi and memes.

The main reasons for being optimistic about encryption AI include:

• The integration of two powerful technologies: encryption and AI.
• Possible AI surge-triggering events, such as IPOs of large AI companies, etc.
• Retail investors find it easy to understand the concept of AI and participate in investments.

2. Bittensor Revival

As a veteran project in the encryption AI field, Bittensor's digital hive mind (Digital Hivemind) is achieving breakthroughs: lower subnet registration fees, faster inference speeds compared to Web2 counterparts, EVM compatibility, etc. The dTAO, expected to launch in the first quarter of 2025, may become a turning point, with each subnet having its own independent token, and relative prices determining emission distribution.

Potential reasons for the resurgence of Bittensor:

• Market-based emission mechanism
• Investors can deploy capital to specific subnets.
• EVM integration attracts more developers

3. The calculation market becomes the next "L1 market"

The demand for computing power is growing exponentially, and traditional infrastructure is struggling to keep up. The decentralized computing layer provides raw computing resources in a verifiable and cost-effective manner. Several startups are building the relevant infrastructure, and as decentralized training of AI models becomes feasible, the overall market size will increase significantly.

Similarities with the L1 market:

• Similar competition will unfold between the calculation protocols.
• Expected to attract a portion of the traditional cloud computing market share

4. AI agents dominate blockchain transactions

By the end of 2025, it is expected that 90% of on-chain transactions will be executed by AI agents rather than manual operations. These agents will continuously rebalance liquidity pools, allocate rewards, or execute small payments based on real-time data feedback.

Reasons why AI agents dominate trading:

• Reduce human error
• Suitable for frequent small payments
• Simplify user operation process

However, ensuring that these agent-driven systems are accountable to humans will be a major challenge.

5. The Rise of Agent Clusters

In the future, AI agents will form collaborative networks to exchange information, negotiate, and make collaborative decisions. These cluster networks are expected to generate more powerful intelligence than a single AI.

Potential Applications of Agent Clusters:

• Coordinate distributed computing resources
• Real-time verification of information content

To achieve clustering, a common communication standard is crucial. Multiple projects are laying the groundwork for the emergence of Agent clusters.

6. Popularization of Human-Machine Hybrid Work Teams

AI agents are expected to become true work partners, possessing autonomy, responsibility, and even compensation. Companies across various industries are testing human-machine hybrid teams.

Advantages of human-machine collaboration:

• Significantly enhance productivity
• Establishing trust mechanisms through smart contracts
• Promote the evolution of relevant social norms

7. 99% of AI Agents will be eliminated

Only Agents that can generate sufficient value can survive. Utility-driven Agents will thrive, while Agents lacking practical use will gradually disappear. This elimination mechanism is beneficial for industry development, encouraging developers to focus on productive use cases.

8. Synthetic Data Surpassing Human Data

Synthetic data provides a scalable, ethical, and privacy-friendly alternative to human data. Its advantages include:

• Unlimited scalability
• Protect privacy
• Can be customized according to needs

It is expected that "micro-laboratories" will emerge to generate highly specialized synthetic datasets tailored for specific use cases.

9. Decentralized training is more practical

Decentralized AI training technology is making breakthroughs. High-performance models that can run on edge devices and even mobile phones are expected to be available soon.

10. Ten Emerging Encryption AI Protocols with Market Capitalization Over 1 Billion USD

The encryption AI market is still in its early stages, and a large number of new protocols, token models, and open-source frameworks are expected to emerge. By the end of 2025, at least ten new projects that have not yet launched are expected to reach a market value of $1 billion.