Coinbase Ventures: An Overview of the Crypto x AI Stack Landscape

In the "Agent Network," AI agents will become the main driving force of economic activities.

Written by: Jonathan King, Head of Coinbase Ventures

Translated by: 0xjs@Golden Finance

The future of AI can be built on blockchain technology, as cryptographic technology helps improve accessibility, transparency, and application scenarios in emerging technology fields. The efficiency, borderless nature, and programmability of cryptocurrencies, combined with AI, have the potential to change the way humans and machines interact with the digital economy, including enabling users to have sovereignty over their personal data. This includes the rise of the "Agent Network," where AI agents operating on a cryptographic infrastructure can drive economic activities and growth.

Disclosure and Notes: Portfolio companies of Coinbase Ventures are marked with an asterisk (*) when first mentioned in this article.

So what does this situation look like? AI agents transact on cryptographic infrastructure. Software code created by AI, including smart contracts, leads to a surge in on-chain applications and experiences. Users own, manage, and profit from the AI models they contribute to. AI is used to improve user and developer experiences within the crypto ecosystem, enhance smart contract capabilities, and create new application scenarios. And so on.

As we envision this future of crypto x AI, today we will reveal the core arguments about the future of this transformative technological integration. The key points are as follows:

We believe that cryptocurrency/blockchain technology does not need to be used at every layer of the AI technology stack to enhance capabilities or address emerging challenges. Instead, cryptocurrency can play a significant role in providing broader distribution, verifiability, censorship resistance, and native payment channels for AI, while benefiting from AI mechanisms to power new user experiences on-chain.
Crypto x AI has the potential to give rise to the "Agent Network," a transformative paradigm where AI agents operating on cryptographic infrastructure can become significant drivers of economic activity and growth. We predict that in the future, agents will have their own crypto wallets to autonomously transact and fulfill user intentions, access low-cost, decentralized computing and data resources, or use stablecoins to pay humans and other agents for tasks necessary to achieve their overall objective functions.
Preliminary points supporting this argument include: (1) Cryptocurrency will become the preferred payment channel for commercial activities between agents and humans, as well as between agents; (2) Generative AI and natural language interfaces will become the primary means of interaction for users seeking to transact on-chain; (3) AI will create the vast majority of software code (including smart contracts), triggering a "Cambrian explosion" of on-chain applications and experiences.
Crypto x AI consists of two core subfields: (1) Decentralized AI (Crypto -> AI), defined as building general artificial intelligence infrastructure to inherit the characteristics of modern peer-to-peer blockchain networks; (2) On-chain AI (AI -> Crypto), defined as building infrastructure and applications that leverage AI to power new and existing application scenarios.
The landscape of the Crypto x AI ecosystem can be divided into the following layers: (1) Computing (i.e., networks focused on providing potential graphics processing units (GPUs) for AI developers); (2) Data (i.e., networks capable of decentralized access, assembly, and verifiability of AI data pipelines); (3) Middleware (i.e., networks/platforms capable of developing, deploying, and hosting AI models/agents); (4) Applications (i.e., user-facing products (B2B or B2C) that utilize on-chain AI mechanisms).

At Coinbase, we are on a mission to help update the financial system to make it safer and more reliable while improving accessibility and usability for consumers and builders. We believe that Crypto x AI will play a significant role in this. In this blog, we will delve into the reasons, methods, and next steps for Crypto x AI.

Introduction to Crypto x AI

The AI market has seen significant growth and investment, with VC firms pouring nearly $290 billion into the field over the past five years. The World Economic Forum notes that AI technology could increase the annual GDP growth of the U.S. by 0.5 - 1.5% over the next decade. AI applications are showing strong appeal, with applications like ChatGPT4 setting new records in user growth/adoption. However, as the AI market rapidly evolves, several challenges have emerged, including data privacy issues, demand for AI talent, ethical considerations, centralization risks, and the rise of deepfake technology. These challenges have driven current discussions in the Crypto x AI space, as stakeholders seek solutions that leverage the strengths of both technologies to address these emerging issues.

From Vitalik Buterin's blog on Crypto x AI

Crypto x AI combines the decentralized infrastructure of blockchain with AI's ability to mimic human cognitive functions and learn from data, creating a synergy that could fundamentally change various industries. Blockchain redefines system architecture, data/transaction verification, and distribution methods. AI enhances data computation, analysis, and provides new content generation capabilities. This intersection has sparked excitement and skepticism among developers in both technology communities, driving exploration of new application scenarios that could accelerate adoption in both fields in the long run. While both cryptocurrency and AI are umbrella terms covering various technologies and topics, we believe the intersection of the two fields can be broken down into two core subfields:

Decentralized AI (Crypto -> AI) enhances AI capabilities through the permissionless and composable infrastructure of cryptocurrency. This unlocks application scenarios such as democratized access to AI resources (e.g., computing, storage, bandwidth, training data), collaborative, open-source model development, verifiable reasoning, or immutable ledgers and cryptographic signatures for content provenance and authenticity.
On-chain AI (AI -> Crypto) brings the advantages of AI into the crypto ecosystem, improving user and developer experiences through large language models and natural language interfaces, or enhancing smart contract capabilities. Two pathways for on-chain AI adoption include: (1) Developers integrating AI models or agents into their smart contracts and on-chain applications; (2) AI agents utilizing cryptographic channels (e.g., self-custody wallets, stablecoins) for payments and delegating decentralized infrastructure resources.

While both subfields are still in their infancy, the potential for "cryptocurrency to integrate with AI" or "AI to integrate with cryptocurrency" is immense and is expected to unlock a range of previously unimagined application scenarios, especially as computing infrastructure and intelligent speed continue to improve.

Crypto x AI: Key Unlocking Factors for the "Agent Network"

One particularly exciting aspect we find in the intersection of cryptocurrency and AI is the concept of AI agents operating on cryptographic infrastructure. This integration aims to create the "Agent Network," a transformative paradigm capable of enhancing security, efficiency, and collaboration in an AI-driven economy, based on robust incentive structures and cryptographic primitives.

We believe that AI agents can become significant drivers of economic activity/growth and the primary "users" of applications (both on-chain and off-chain), gradually replacing human users in the medium to long term. This paradigm shift will force many internet-native companies to rethink their core assumptions about the future and provide the necessary products, services, and business models to better serve an agent-centric economy. That said, we do not believe that cryptocurrency/blockchain technology needs to be used at every layer of the AI technology stack to enhance capabilities or address emerging challenges. Instead, cryptocurrency can play a significant role in providing broader distribution, verifiability, censorship resistance, and native publishing channels for AI, while benefiting from AI mechanisms to power new user experiences on-chain.

Preliminary points supporting this argument are as follows:

Cryptocurrency will become the preferred payment channel for commercial activities between agents and humans, as well as between agents: Cryptocurrency is the internet-native programmable currency, offering various advantages for driving an agent-based economy. As AI agents become more autonomous and engage in microtransactions at scale (e.g., payments for reasoning, data, API access, decentralized computing, or data resources), the efficiency, borderless nature, and programmability of cryptocurrency will make it a superior medium of exchange compared to traditional fiat channels. Additionally, agents will require unique, verifiable identities (i.e., "Know Your Agent") to ensure compliance with regulatory rules and requirements when transacting with businesses and end-users. Low-fee blockchains, smart contracts, self-custody wallets (like Coinbase AI Wallets), and stablecoins can help simplify and reduce the costs of complex financial agreements between agents, while the verifiability and immutability of decentralized networks will ensure the significance and auditability of AI agent transactions.
Generative AI and natural language interfaces will become the primary means of interaction for users seeking to transact on-chain: As natural language processing speeds and AI's contextual understanding of cryptocurrency improve, interacting on-chain through conversational interfaces will become the default norm and expectation for users, consistent with current Web2 trends (e.g., ChatGPT). Users will simply describe their intended transaction in natural language (e.g., "swap X for Y"), and the AI agent will translate these intentions into verifiable smart contract code, providing the most efficient and cost-effective transaction execution path.
AI will create the vast majority of software code (including smart contracts), triggering a "Cambrian explosion" of on-chain applications and experiences: AI's code generation capabilities are rapidly evolving in Web2 (e.g., Devin, Replit) and fundamentally changing the software development paradigm. We believe this shift will soon take center stage in the crypto space, with a recent focus on significantly lowering the entry barriers for new and existing developers. However, the future state is that AI "software agents" will generate smart contracts and highly personalized applications from scratch in real-time based on user preferences, storing and verifying them on-chain.

These points suggest that the boundaries between AI and cryptocurrency will increasingly blur in the future, creating a new paradigm of intelligent, autonomous, and decentralized systems. Based on this framework, let us delve deeper into the technology stack enabling the integration of cryptocurrency and AI.

Opportunities in the Crypto x AI Stack (Current)

The exploration of "integrating cryptocurrency into AI" or "integrating AI into cryptocurrency" has given rise to an emerging and complex field that is rapidly developing, with many builders eager to capitalize on market momentum. Today, we believe the Crypto x AI space can be divided into the following layers: (1) Computing (i.e., networks focused on providing potential graphics processing units (GPUs) for AI developers); (2) Data (i.e., networks capable of decentralized access, orchestration, and verifiability of AI data pipelines); (3) Middleware (i.e., networks/platforms capable of developing, deploying, and hosting AI models/agents); (4) Applications (i.e., user-facing products (B2B or B2C) that utilize on-chain AI mechanisms).

Computing

Both model training and inference execution for AI require substantial computational GPU resources. As AI models become increasingly complex, the demand for computation is rising, leading to shortages of advanced GPUs like those from NVIDIA, resulting in longer wait times and increased costs. Decentralized computing networks are emerging as potential solutions to these challenges in the following ways:

Establishing permissionless markets for purchasing, leasing, and hosting physical GPUs
Building GPU aggregators that allow anyone (e.g., Bitcoin miners) to contribute their excess GPU computing power to perform on-demand AI tasks and receive token incentives in return
Financializing physical GPUs by tokenizing them as digital assets on-chain
Developing distributed GPU networks for compute-intensive workloads (e.g., training, inference)
Creating infrastructure that enables AI models to run on personal devices (similar to decentralized Apple devices)

These proposed solutions aim to increase the supply and accessibility of GPU computing while providing highly competitive pricing. However, given that most participants in this field vary in their support for advanced AI workloads and face challenges related to the lack of co-location with GPUs, as well as, in some cases, a lack of developer tools and uptime guarantees comparable to centralized alternatives, we believe these products are unlikely to achieve mainstream adoption in the near to medium term. Emerging fields and example projects being built in this layer include:

General computing: Decentralized computing markets providing GPU computing resources for various applications (e.g., Akash, Aethir)
AI/machine learning computing: Decentralized computing networks providing GPU computing resources for specific services (e.g., GPU aggregators, distributed training and inference, GPU tokenization, etc.) (e.g., io.net, Gensyn, Prime Intellect, Hyperbolic, Hyperspace)
Edge computing: Computing and storage networks powering large language models for personal, contextual inference (e.g., PIN AI, Exo, Crynux.ai, Edge Matrix)

Data

Scaling AI models requires continuously growing training datasets, with large language models trained on trillions of words of human-generated text. However, the currently available public, human-generated data is limited (Epoch AI estimates that high-quality language/data sources may be exhausted by 2024), raising the question: Will the lack of training data become a major bottleneck for AI model performance, potentially leading to stagnation? Therefore, we believe data-focused crypto x AI companies have the following opportunities to address these challenges:

Incentivizing users to share their private/proprietary data (e.g., "Data Decentralized Autonomous Organizations (Data DAOs)"—on-chain entities where data contributors can see economic benefits from contributing their private data from social platforms and manage the use and monetization of that data)
Creating tools for generating synthetic data assets from natural language prompts or providing user incentives to scrape data from public websites
Incentivizing users to help preprocess datasets for training models and maintain data quality (e.g., data labeling/reinforcement learning from human feedback)
Establishing multifaceted, permissionless data markets where anyone can be compensated for contributions

These opportunities have given rise to many emerging participants we see today in the data layer. However, it is worth noting that existing centralized enterprises in the AI model lifecycle have established network effects and proven data compliance regimes, which traditional companies value, potentially leaving little room for decentralized alternatives. Nevertheless, we believe the data layer of decentralized AI presents a significant long-term opportunity to address the "data wall" challenge. Emerging fields and example projects being built in this layer include:

Data markets: Decentralized data exchange protocols designed for data providers and consumers to share and trade data assets (e.g., Ocean Protocol, Masa, Sahara AI)
User-owned/private data (including DataDAOs): Networks designed to incentivize the collection of proprietary datasets (including private user-owned data) (e.g., Vana*, NVG8)
Public and synthetic data: Networks/platforms for scraping data from public websites or generating new datasets through natural language prompts (e.g., Dria, Mizu, Grass, Synesis One)
Data intelligence tools: Platforms and applications for querying, analyzing, visualizing, and providing actionable insights on on-chain data (e.g., Nansen, Dune, Arkham, Messari*)
Data storage: File storage networks for long-term data storage/archiving and relational database networks for managing frequently accessed and updated structured data (e.g., Filecoin, Arweave, Ceramic, Tableland*)
Data assembly/tracing: Networks and platforms optimizing data ingestion pipelines and processing for AI and data-intensive applications, ensuring correct provenance tracking and verifiable authenticity of AI-generated content (e.g., Space and Time, The Graph*, Story Protocol)
Data labeling: Networks and platforms improving AI models' reinforcement learning and fine-tuning mechanisms by incentivizing distributed human contributors to create high-quality training datasets (e.g., Sapien, Kiva AI, Fraction.AI)
Oracles: Networks using AI to provide verifiable off-chain data for on-chain smart contracts (e.g., Ora, OpenLayer, Chainlink)

Middleware

To realize the full potential of an open, decentralized AI model or agent-based ecosystem, new infrastructure needs to be built. Some high-potential areas that builders are exploring include:

Utilizing publicly available large language models to power on-chain AI application scenarios while building foundational models capable of quickly understanding, processing, and acting on on-chain data
Distributed training solutions for large foundational models (e.g., 100B+ parameters); due to various technical complexities, this is often seen as an unattainable dream, but recent breakthroughs from Nous Research, Bittensor, and Prime Intellect are attempting to change that
Implementing privacy-preserving, verifiable inference using zero-knowledge or optimistic machine learning (i.e., zkML, opML), trusted execution environments (TEEs), or fully homomorphic encryption (FHE)
Achieving open, collaborative AI model development through resource coordination networks or building agent networks/platforms that leverage cryptographic infrastructure channels to enhance the potential of AI agents in on-chain/off-chain application scenarios

While some progress has been made in building these foundational infrastructure primitives, production-ready, on-chain large language models and AI agents are still in their infancy, and we do not expect this situation to change until the maturity of computing, data, and model infrastructure. Nevertheless, we believe this category is very promising and is a core focus of Coinbase Ventures' investment strategy, driven by the long-term implicit growth and demand for AI services. Emerging fields and example projects being built in this layer include:

Open-source large language models: Open-source AI models that are publicly accessible, allowing anyone to use, modify, and freely distribute (e.g., LLama3, Mistral, Stability AI)
On-chain model creators: Networks and platforms capable of creating foundational large language models for on-chain application scenarios (e.g., Pond*, Nous, RPS)
Training and fine-tuning: Networks and platforms capable of implementing incentivized, verifiable training or fine-tuning mechanisms on-chain (e.g., Gensyn, Prime Intellect, Macrocosmos, Flock.io)
Privacy: Networks and platforms employing privacy-preserving mechanisms for the development, training, and inference of AI models (e.g., Bagel Network, Arcium*, ZAMA)
Inference networks: Networks and platforms using cryptographic techniques/proofs to verify the correctness of AI model outputs (e.g., OpenGradient*, Modulus Labs, Giza, Ritual)
Resource coordination networks: Networks designed to facilitate resource sharing, collaboration, and coordination for AI model development (e.g., Bittensor, Near*, Allora, Sentient)
Agent networks and platforms: Networks and platforms for facilitating the creation, deployment, and monetization of AI agents in on-chain/off-chain environments (e.g., Morpheus, Olas, Wayfinder, Payman, Skyfire)

Applications

In the cryptocurrency space, AI agents are beginning to emerge, with early examples such as Dawn Wallet (a crypto wallet that uses AI agents to send transactions on behalf of users and interact with protocols), Parallel Colony* (an on-chain game where players collaborate with AI agents that have their own wallets and can create their own paths in the game), or Venice.ai (a generative AI application/natural language prompt with verifiable reasoning and privacy protection mechanisms). However, application development remains largely experimental and opportunistic, with a plethora of application ideas emerging amid the hype in this field. Nevertheless, we believe that advancements in AI agent infrastructure and frameworks will enable the crypto industry to transition from primarily reactive smart contract applications to more complex proactive applications in the medium to long term. Emerging fields and example projects being built in this layer include:

AI Companions: Applications for creating, sharing, and monetizing personalized and context-aware AI models and agents owned by users (e.g., MagnetAI, MyShell, Deva, Virtuals Protocol)
Natural Language Processing Interfaces: Applications that use natural language prompts as the primary interface/entry point for interacting with and executing on-chain transactions (e.g., Venice.AI, Veldt)
Development/Security Tools: Developer-focused applications/tools that leverage AI models/agents to enhance the on-chain developer experience and security mechanisms (e.g., ChainGPT, Guardrail*)
Risk Agents: Services that utilize machine learning models or AI agents to help protocols dynamically adjust and respond to on-chain risk parameters in real-time (e.g., Chaos Labs, Gauntlet, Minerva*)
Identity (Personhood Proof): Applications that use cryptographic proofs and machine learning models to verify users' personhood proofs (e.g., Worldcoin*)
Governance: Applications that utilize AI agents to execute transactions based on human-driven governance decisions/feedback (e.g., Botto, Hats)
Trading/Decentralized Finance: AI-driven trading infrastructure and decentralized finance protocols that utilize AI agents to automatically execute on-chain transactions (e.g., Taoshi, Intent.Trade)
Gaming: On-chain games that leverage intelligent non-player characters or AI mechanisms to drive core gameplay mechanics (e.g., Parallel*, PlayAI)
Social: Applications that utilize AI mechanisms to drive on-chain social experiences (e.g., KaiKai, NFPrompt)

Conclusion

While the Crypto x AI stack is still in its infancy, we believe there will be significant progress in decentralized AI infrastructure, on-chain AI applications, and the emergence of "agent networks," where AI agents will become the primary drivers of economic activity. Although challenges remain in areas such as computational infrastructure and data availability, the synergy between cryptocurrency and AI could accelerate innovation in both fields, leading to more transparent, decentralized, and autonomous systems. As this field continues to rapidly evolve, driven by newly funded teams and more mature teams dedicated to finding product/market fit, it will be crucial for internet-native companies and developers to adapt to changing paradigms and embrace the potential of Crypto x AI to create previously unimaginable new applications and experiences.

Overall, Coinbase Ventures is excited about the future potential and opportunities of Crypto x AI, and we are actively investing in every layer of this architecture.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。