Crypto x AI: Top 10 cutting-edge trends to look forward to in 2025

Author: Archetype Source: X, @archetypevc Translation: Shan Ouba, Golden Finance

1. Interaction between agents

The transparency and composability of blockchain make it an ideal substrate for interaction between agents. In this scenario, agents developed by different entities for different purposes can interact with each other seamlessly. Currently, many experimental applications have emerged between agents, such as fund transfers between agents and joint launch of tokens. We look forward to seeing how the interaction between agents can be further expanded, including the creation of new application areas (for example, new social scenarios driven by agent interaction), and the improvement of current cumbersome enterprise workflows, such as platform authentication and verification, micropayments, and cross-platform workflow integration.

— Danny, Katie, Aadharsh, Dmitriy

2. Decentralized Agent-Based Organizations

Coordination of large-scale multi-agent systems is another exciting area of ​​research. How do multi-agent systems collaborate to complete tasks, solve problems, and govern systems and protocols? In his article "The Promise and Challenges of Crypto+AI Applications" in early 2024, Vitalik proposed using AI agents for prediction markets and arbitration. He believes that multi-agent systems have extraordinary "truth" discovery capabilities and autonomous governance potential when operating at scale. We look forward to seeing how the potential of "multi-agent systems" and forms of "swarm intelligence" can be further explored and experimented with.

As an extension of inter-agent coordination, agent-human coordination also offers interesting room for design—particularly how communities interact around agents, or how agents organize humans for collective action. We expect more experiments to emerge, especially with agents whose objective functions involve large-scale human coordination. This will require some kind of verification mechanism, especially if the human work is done off-chain, but it has the potential to produce some strange and interesting emergent behaviors.

— Katie, Dmitriy, Ash

3. Intelligent Multimedia Entertainment

The concept of Digital Personas has been around for decades. For example, Hatsune Miku (2007) sold out 20,000 seats at a concert, and virtual influencer Lil Miquela (2016) has over 2 million followers on Instagram. Newer examples include AI virtual streamer Neuro-sama (2022), who has over 600,000 subscribers on Twitch, and anonymous K-pop virtual boy band PLAVE (2023), who have accumulated over 300 million views on YouTube in less than two years.

With advances in AI infrastructure and the integration of blockchain for payments, value transfer, and open data platforms, we expect to see how these agents become more autonomous, potentially unlocking a whole new category of mainstream entertainment by 2025.

— Katie, Dmitriy

4. Generative/Agentic Content Marketing

In the previous category, the agent itself was the product, while here the agent can complement the product. In the attention economy, the continuous production of engaging content is critical to the success of any idea, product or company. Generative/agentic content is a powerful tool that teams can use to build a scalable 24/7 content production pipeline. The development of this area has been driven by the discussion around "what is the difference between meme coins and agents". Even if the current meme coins are not strictly "intelligent", agents have become an important tool for them to obtain distribution channels.

Another example is that games often need to become more dynamic if they want to maintain user engagement. A classic way to create game dynamism is to foster user-generated content; purely generative content (including in-game items, NPCs, and even fully generated game levels) is likely the next stage of this evolution. We are curious to see to what extent the boundaries of traditional distribution strategies can be pushed by the capabilities of intelligent agents in 2025.

— Katie

5. Next Generation Art Tools/Platforms

In 2024, we launched IN CONVERSATION WITH, a series of interviews with crypto artists in music, visual arts, design, curation, and more. A key observation I made during this year’s interviews is that artists interested in crypto tend to also have a broad interest in cutting-edge technologies and tend to make these technologies a core or aesthetic focus of their creative practice, such as AR/VR objects, code-based art, and live coding.

Generative Art has long had a natural synergy with blockchain, which makes it clearer as a potential vehicle for AI art. It is extremely difficult to display and present these artistic mediums on traditional art display platforms. ArtBlocks provides us with a window into how blockchain can be used to present, store, monetize, and protect digital artworks in the future - while improving the overall experience for artists and audiences.

Beyond display, AI tools even expand the ability of ordinary people to create art. We look forward to seeing how blockchain can further expand or support these tools in 2025 to empower art creators and enthusiasts.

— Katie

6. Data Marketplace

Since Clive Humby coined the phrase “data is the new oil” 20 years ago, companies have taken strong steps to monopolize and monetize user data. Today, users are aware that their data is the foundation upon which these multi-billion dollar companies are built, but they have minimal control over it and little to no share in the profits it generates. This tension becomes increasingly critical as the development of powerful AI models accelerates. If part of the opportunity in data marketplaces is to reduce the exploitation of user data, another part is to address the problem of data shortage, as increasingly powerful AI models are exhausting the easily accessible data resources on the internet, and new sources of data are urgently needed.

The design space for how to use decentralized infrastructure to return control of data to users is very broad, and innovative solutions are urgently needed in multiple areas. Some of the most pressing challenges include:

• Where to store data and protect its privacy (during storage, transmission, and computation)

• How to objectively evaluate, screen, and measure data quality

• Mechanisms for data attribution and monetization (especially tracing value back to the source after inferring results)

• How to orchestrate or retrieve data in a diverse model ecosystem.

Regarding solving the bottleneck of data supply, the key is not just to replicate existing data annotation platforms (such as Scale AI) with tokens, but to understand how we can create competitive solutions through technological advantages, whether in scale, quality, or incentive mechanisms, to produce high-value data products. Especially in the context of the demand side mainly coming from Web2 AI, thinking about how to combine the mechanism of smart contract execution with traditional service level agreements (SLAs) and tools is an important area worthy of attention.

— Danny

7. Decentralized computing power

If data is a fundamental building block for AI development and deployment, then computing power is another. Over the past few years, the old paradigm dominated by large data centers (such as exclusive access to specific sites, energy, and hardware) has largely defined the development trajectory of deep learning and AI. However, this dynamic is being challenged as physical limitations emerge and open source technologies develop.

The v1 version of decentralized AI computing power looks like a replica of the Web2 GPU cloud, with no real advantage in supply (hardware or data center) and a lack of natural market demand. In the v2 version, some teams are developing technology stacks to build competitiveness through capabilities such as orchestration, routing, and pricing of **heterogeneous high-performance computing resources (HPC)**, and introducing proprietary features to attract demand and combat profit compression, especially in reasoning tasks. In addition, teams are beginning to compete in a differentiated manner around different application scenarios and market entry strategies (GTMs), with some teams focusing on using compiler frameworks to improve reasoning routing efficiency across diverse hardware, while others are pioneering distributed model training frameworks on the computing power networks they have built.

We are even beginning to see the beginnings of an AI-Fi market, proposing new economic primitives to turn computing power and GPUs into yield-generating assets, or to leverage on-chain liquidity to provide data centers with alternative sources of funding for hardware. A key question is to what extent will decentralized AI (DeAI) rely on decentralized computing power for development and deployment? Or will it be like the storage market, where the gap between ideals and actual needs is never bridged and the full potential of the concept is ultimately not realized?

— Danny

8. Computing power accounting standards

Associated with incentivizing decentralized high-performance computing power networks, a major challenge in coordinating heterogeneous computing power is the lack of a set of recognized computing power accounting standards. The unique output characteristics of AI models introduce complexity to high-performance computing markets, such as different model variants, quantization techniques, and tunable randomness through temperature and sampling hyperparameters. In addition, different AI hardware (such as GPU architectures and CUDA versions) further lead to output differences. Ultimately, this requires the establishment of standards for how models and computing markets account for their capabilities in heterogeneous distributed systems. Due to the lack of standards, this year we have seen multiple cases in both Web2 and Web3 where models and computing markets were unable to accurately account for the quality and quantity of their computing power. This has led to users having to run their own model benchmarks, audit by comparing performance results, and even verify true performance by limiting the workload of computing markets (Proof-of-Work). Given the core principle of "verifiability" in the crypto field, we hope that in 2025, the combination of crypto and AI will have an advantage over traditional AI in terms of verifiability. Specifically, ordinary users should be able to compare the output of models or computing clusters like by like to audit and benchmark system performance.

— Aadharsh

9. Probabilistic Privacy Primitives

In "The Promise and Challenges of Crypto + AI Applications", Vitalik proposed a unique challenge facing the integration of encryption and AI:

"In cryptography, open source is the only way to achieve security, but in AI, open source models (even training data) greatly increase the risk of adversarial machine learning attacks."

Although privacy is not a new research field for blockchain, the rapid development of AI will further accelerate the research and application of privacy-enhancing technologies. This year, we have seen significant progress in privacy technologies such as zero-knowledge proofs (ZK), fully homomorphic encryption (FHE), trusted execution environments (TEEs), and multi-party secure computation (MPC), which can be used for private shared computation on encrypted data and are suitable for general-purpose applications. At the same time, we have also seen centralized AI giants such as NVIDIA and Apple using proprietary TEE technology to enable federated learning and private AI reasoning in systems with consistent hardware, firmware, and models.

Given this, we will closely monitor the development of how to maintain privacy in random state transitions and how these advances can accelerate the implementation of decentralized AI applications in heterogeneous systems, including decentralized private reasoning, storage and access pipelines for encrypted data, and fully autonomous execution environments.

— Aadharsh

10. Agent Intentions and Next-Generation User Trading Interfaces

The use of AI agents in autonomous on-chain trading is one of the most realistic use cases. However, there has been a lot of fuzzy definitions over the past 12-16 months around concepts such as “intent”, “agent behavior”, “agent intent”, “solver”, and “agent solver”, especially how to distinguish them from the development of traditional “trading robots” in recent years.

In the next 12 months, we expect to see more advanced language systems combined with different data types and neural network architectures to advance the overall design space.

• Will agents use current on-chain systems to trade, or will they develop their own tools/methods?

• Will Large Language Models (LLMs) continue to serve as the backend for these agent trading systems, or will something completely different emerge?

• On a user interface level, will users start using natural language to trade?

• Will the long-held assumption of “wallets as browsers” finally come to fruition?

These questions will be our focus.

— Danny, Katie, Aadharsh, Dmitriy