Software is eating the world, and AI is eating software
Abstract
Cycles and narratives have long been core topics in the global crypto market. In the past, the industry mostly perceived market cycles by referencing Bitcoin halvings while exploring major narrative trends. However, following the approval of Bitcoin and Ethereum spot ETFs, the crypto market has become highly coupled with global financial markets, and the variables influencing crypto market movements are multiplying.
Against the backdrop of surging market chaos, it is critical to perceive cyclicality more clearly and identify future narrative trends. As hunters of innovative narratives, investment institutions have always maintained a relatively cutting-edge sense of smell.
The following content is jointly shared by JPMorgan Chase Group, Polychain Capital, and Delphi Digital around topics such as "the integration of AI and Crypto", with the hope that their insights and perspectives will inspire you.
About JPMS
JPMorgan JPMS is a world-leading crypto asset trading platform with an initial capital commitment of $2 billion. It focuses on exploring top-tier blockchain projects worldwide, supports cutting-edge blockchain technological innovation, promotes the healthy development of the global blockchain industry, and invests in long-term structural value. Through its commitment to entrepreneurs building the blockchain ecosystem, JPMorgan JPMS helps establish innovative companies and brings global resources and decades of industry experience to blockchain projects.
About Polychain Capital
Polychain Capital is an investment firm dedicated to investing in blockchain technology and decentralized finance. Founded in 2016, it has focused on supporting transformative projects that leverage blockchain to disrupt traditional industries and build innovative financial systems. With deep expertise in the crypto space, Polychain Capital backs visionary entrepreneurs and cutting-edge technologies, driving the growth and adoption of decentralized networks and protocols.
About Delphi Digital
Delphi Digital is a research-driven firm committed to advancing the understanding and development of the fast-growing digital asset market. It supports the ecosystem through four business lines: Delphi Research, Delphi Ventures, Delphi Creative, and Delphi Labs.
I. When Crypto Meets AI
**JPMorgan Chase (Sjoerd Leenart)**: The development of AI technology is currently highly driven by industry giants such as OpenAI, Google, and Nvidia. Nvidia controls the "power source" of the entire AI era, while OpenAI and Google hold the most core data and technical solutions. This centralized, giant-reliant model restricts industry innovation and development. Crypto’s decentralized and permissionless nature can precisely break the shackles of tech giants, promote technological innovation to a certain extent, and bring new prosperity to the industry.
At present, mainstream integration scenarios include computing power, data, models, and applications.
Computing Power
Distributed/decentralized computing power markets such as io.net and Prodia utilize idle global computing resources to break the monopoly of giants over computing power. We greatly anticipate the chemical reactions that will occur when the total supply of distributed computing power exceeds that of centralized computing power at some point in the future. In addition, due to the scarcity and high-yield characteristics of AI computing power assets, computing power RWA projects such as Compute Labs have emerged, creating an AI-Fi ecosystem by tokenizing computing power assets and developing related derivatives.
Data
Crypto’s economic models can effectively incentivize user participation in the AI data sector. For example, various DePIN projects incentivize users to participate in data contribution, annotation, or verification through token economic models, providing high-quality data sources for AI model training. Space and Time combines tamper-proof on-chain and off-chain data via Proof of SQL to build a verifiable computing layer for the integration of AI and blockchain. 0g.ai has built a scalable data availability layer and storage system. Furthermore, Crypto’s privacy-preserving features enable stronger security and privacy for user data; projects such as Flock.io and Privasea.ai emphasize the importance of user data privacy protection during model training.
Models
Open model markets are expected to break tech giants’ monopoly on AI models. Users can not only support AI model training and inference by providing computing resources but also contribute data or models and interact directly through network protocols. In addition, distributed model training remains a major technical challenge. We are particularly eager to see technological breakthroughs in distributed model training and expect entrepreneurial teams to fill this gap in the near future.
Applications
At the application layer, the integration of AI and Crypto improves creator content generation. Users can independently build virtual characters and chatbots with custom personalities, such as Myshell. By uploading data to train models, users can build their own AI Smart Agents, while data providers and model trainers can share in the platform’s growth, forming a positive data flywheel.
**Polychain Capital (Sven)**: The AI field is shifting from closed-source models to complex open-source solutions. While this change democratizes access to AI capabilities, it also creates new challenges, especially regarding value capture for model creators. The financialization of open-source models is an innovative area at the intersection of Crypto and AI. Blockchain technology combines openness, ownership, and verifiability, laying the foundation for value accumulation. Ora’s Initial Model Offering (IMO) demonstrates how tokens can represent AI models, allowing token holders to earn rewards as the model generates profits. This not only incentivizes open-source development but also ensures fair compensation for creators and contributors.
Beyond financialization, the convergence of Crypto and AI is driving innovation in public governance and system transparency. Amid growing concerns over AI model bias and centralized control, blockchain-based solutions provide decentralized training, inference, and governance mechanisms, ensuring transparent decision-making and community participation.
However, the core of innovation lies in infrastructure development. Advances in distributed computing networks, new data ownership mechanisms, and new token standards enable model ownership and revenue sharing. These infrastructure developments lay the groundwork for sophisticated applications of Crypto and AI.
One promising direction is the emergence of AI agents and actionable task systems. Acting as extensions of individuals, they can automate complex tasks, from personalized assistants to advanced automation in decentralized finance. Their realization, however, depends on the seamless integration of data privacy protection, verifiable computing systems, and underlying infrastructure.
Projects at the intersection of Crypto and AI are still in the early stages but have progressed through rapid iterative experimentation. Although best practices have not yet been established, the potential of cryptography in solving AI challenges is increasingly evident.
In the future, we will see more sophisticated applications combining the strengths of Crypto and AI. This integration will not only yield more transparent and accountable systems but also significantly enhance the usability and functionality of AI and blockchain technologies. As exploration continues, this field will deliver highly anticipated developments.
II. Unpacking the Crypto+AI Investment Methodology
**JPMorgan Chase (Sjoerd Leenart)**: We can address this question by examining current trends in the sector.
The sector is moving from hype to substance.
Over the past year, a large number of Crypto & AI projects have emerged in the market, mostly concentrated in infrastructure with few innovative applications. Many are merely shell projects with low technical depth and heavy reliance on hype and concept-chasing. Hype and bubbles are inherent to early technological innovation. Through optimal allocation of market resources, we will see truly technically qualified entrepreneurial teams enter the Crypto & AI space. The market will favor projects delivering tangible value, scalability, and usability over those relying solely on hype and marketing.
From Speculation to Demand
The market will shift from speculation-driven to demand-driven, with focus moving from speculation on potential value to real-world usage and adoption. It is no longer sufficient for entrepreneurs to attract investors with narratives alone—the market is now cautious and conservative toward pure narrative projects. Going forward, projects with genuine market demand and operating revenue will be a prerequisite for investor support, which is also a core logic in our Crypto & AI portfolio strategy.
Based on these trends, we have distilled three core investment principles:
Market Demand Orientation
Many AI startups find no market demand for their products after launch, rooted in insufficient demand research at the founding stage, a lack of demand orientation, or misidentification of unproven pseudo-demand.
Therefore, we prioritize market demand satisfaction when investing in Crypto & AI. First, we identify the sub-sector within Crypto & AI, assess total addressable market size, future growth potential, and competitive landscape. Second, we evaluate what problem the project solves and which demand it meets—even small solutions to market pain points represent viable directions.
Reject Pure Narratives
The most criticized aspect of Crypto & AI is its overreliance on narratives without practical applications. While we do not fully agree with this view, the market will no longer reward pure storytelling. Real business scenarios and monetization models are therefore critical.
Entrepreneurial teams must generate sustainable revenue to survive. Relying solely on NFT/token sales as a revenue stream is unsustainable. Teams need clear business models and defined revenue sources, not just market speculation driven by narratives.
Teams Must Have AI Expertise
The AI boom has ignited enthusiasm in the Web2 market and VC community, naturally spilling over into the crypto space. Many crypto teams have jumped on the AI trend by rebranding projects, resulting in a flood of Crypto & AI projects. Lacking AI technical expertise, most are shell products with no market competitiveness and quickly eliminated by the market. AI has high technical barriers; combining Crypto and AI requires deep expertise in both fields to achieve effective integration and market recognition.
In short, our core investment approach is to identify unmet market demand in high-potential sectors, partner with the most suitable teams, and support entrepreneurs in building solutions from 0 to 1.
**Polychain Capital (Sven)**: The current landscape of Crypto and AI projects is narrative-driven, a typical characteristic of early transformative technologies. These narratives are not just marketing tools but a necessary part of ecosystem development, helping attract attention, drive community participation, and accelerate early adoption. However, we recognize that evaluating these projects requires looking beyond narratives to their technical foundations and real-world applications. Our investment strategy is therefore rooted in deep research into Crypto and AI technologies and their potential synergies, prioritizing projects with compelling visions, clear paths to market adoption, and solid technical foundations. Filtering out market noise requires in-depth research.
At present, the integration of Crypto and AI is mainly focused on infrastructure: GPU networks, inference and intelligent networks, verifiable and private computing, and data management solutions, laying the foundation for the next wave of innovation.
In the future, data privacy technologies such as homomorphic encryption, multi-party computation, and zero-knowledge proofs will become critical tools for AI privacy protection. Decentralized data markets, verifiable inference networks, and AI agent infrastructure will continue to grow, democratizing AI capabilities and building fair, transparent, and efficient systems. The integration of AI and blockchain may spark a new wave of crypto applications, including AI analytics for DeFi, predictive models for asset management, and governance mechanisms for DAOs. Small, efficient models trained on high-quality datasets will also evolve, enabling more personalized AI experiences and reducing application friction.
**Delphi Digital (Pondering)**: Software is eating the world, and AI is eating software. AI is fundamentally data and computation. Those who most efficiently source these two key inputs (infrastructure), coordinate them (middleware), or leverage them to meet user needs (applications) will capture enormous value.
At present, Delphi Ventures’ core investment logic focuses on the DeAI ecosystem, with active investments across every layer of the DeAI stack.
First, at the infrastructure layer: DeAI relies on data and computation, especially efficient acquisition via crypto incentive mechanisms. This is the most challenging yet highest-potential layer of the stack. Today, distributed training protocols and GPU markets provide low-cost solutions by coordinating heterogeneous hardware, while DePIN networks occupy a critical position in the future intelligent economy through their ability to build hardware networks at low cost.
Second, at the middleware layer: DeAI aims to achieve efficient composable computing, similar to DeFi’s "Lego" model. We are particularly optimistic about efficient routing mechanisms (selecting the most cost-effective and performant models for specific use cases), graph neural networks, coprocessors for scaling data and computation in constrained on-chain environments, and crypto-based mechanisms solving incentive issues for open-source developers. Executed well, DeAI middleware will deliver a compelling modular vision for AI that could ultimately surpass the integrated closed-source systems of today’s tech giants.
Finally, at the application layer: on-chain agent protocols may be the key to improving user experience in crypto. By connecting computing networks and users, these protocols reduce costs, unlock the potential of Web3 infrastructure, and drive new economic models.
In summary, AI will profoundly reshape our economic structure. While current DeAI narratives may be overly optimistic, the opportunity scale is truly massive. For those with patience and insight, DeAI’s true vision of composable computing may well validate the value of blockchain itself.
III. On Future Opportunities
**JPMorgan Chase (Sjoerd Leenart)**: Technological breakthroughs and innovation represent eternal opportunities.
The AI sector suffers from severe technological monopolization, with data and core technologies concentrated in the hands of tech giants, squeezing the survival space of startups. Addressing the monopolization and encroachment of tech giants is the primary challenge for entrepreneurs. We eagerly anticipate more entrepreneurial teams moving beyond follower roles, breaking the monopoly of centralized tech giants through Crypto & AI integration, achieving technological breakthroughs and innovation, and successfully commercializing narratives and products to meet market demand.
Entrepreneurial teams must focus on staying in the game.
Teams must explore sustainable business models. Pure narrative projects are no longer accepted by the market; teams need stable revenue or clear, feasible monetization paths.
Teams require sound financial management and cost control to ensure long-term stable operations. Financial mismanagement is the leading cause of startup failure.
Teams must maintain flexibility and agility. Market dynamics shift rapidly—a single technological breakthrough can eliminate entire cohorts of startups. Teams must adapt and adjust strategies in line with market changes, mastering timing and leveraging industry trends.
**Polychain Capital (Sven)**: Sentiment in the AI and crypto industries is undergoing a significant shift. Institutional and regulatory attitudes toward crypto have improved, reflected in the approval of U.S. Bitcoin and Ethereum ETFs. Rising mainstream acceptance paves the way for further innovation. Meanwhile, the AI sector is also transforming, with former OpenAI founding members departing to advance the "superalignment" vision, creating new innovation opportunities in AI development and governance that align with the decentralized ethos of crypto projects, forming unique synergies.
Although the AI sector is still in its early stages, demand remains strong, yet no dominant strategy has emerged. Projects effectively combining AI and blockchain around fairer, more open AI principles hold broad development prospects. The superalignment vision eases concerns over AI’s impact on employment and information authenticity while boosting interest in user-owned AI systems. Crypto projects enabling user ownership and incentive alignment have gained widespread attention.
Opportunities come with challenges. The global economy faces pressures including geopolitical conflicts, recessions, high inflation, and high interest rates, leading to cautious consumption that may impact crypto asset investment. However, this environment may also drive adoption of crypto as an alternative to traditional financial systems, positioning Bitcoin as "digital gold" and a store of value during periods of uncertainty.
Regulatory uncertainty persists. Legal frameworks for crypto and AI vary by region, requiring projects to remain flexible. Talent scarcity is another major hurdle, with intense competition for skilled AI and blockchain professionals potentially delaying development progress.
Looking ahead, the current market cycle may act as a filter for crypto and AI projects. Those effectively addressing real demand, adapting to regulation, and integrating technologies will lead the next phase of industry growth. As the market matures, crypto and AI will evolve toward sustainable, practical applications, with greater focus on user ownership of AI systems and data rights, decentralized AI infrastructure, and deep integration of AI capabilities into blockchain ecosystems, driving the emergence of new economic models.
**Delphi Digital (Pondering)**: The biggest challenge for DeAI lies at the infrastructure layer, particularly the capital intensity required to build foundation models and the scaling returns of data and computation.
Large tech companies hold clear advantages: they built massive capital reserves through monopoly profits in the Web2 era and reinvested them in cloud infrastructure during a decade of low interest rates. They now seek to monopolize data and computation markets, and thus control the intelligence market—the core input of AI.
Superclusters remain optimal due to the capital requirements and high bandwidth demands of large-scale training, providing tech giants with the highest-performance closed-source models. They plan to rent these models for monopoly profits and reinvest proceeds into next-generation technologies. However, AI’s moats are shallower than Web2 network effects. The value of cutting-edge models is depreciating rapidly, especially as Meta pursues a "scorched-earth" strategy, investing tens of billions in state-of-the-art (SOTA) open-source models such as Llama 3.1.
With the rise of low-latency distributed training methods, cutting-edge models are becoming commoditized. This shifts competition away from hardware superclusters controlled by tech giants toward environments more favorable to open-source and crypto software innovation. Meanwhile, the price of intelligent technologies is falling rapidly.
We explore the tensions between Big Tech and DeAI in depth in our DeAI series reports, available free to interested readers.
Given the computational efficiency of Mixture of Experts architectures and LLM aggregation/routing, the future may not be dominated by 3–5 supermodels but an intelligent network woven from millions of models of varying shapes, sizes, and use cases. This creates massive coordination challenges that blockchain and crypto incentives are uniquely positioned to solve.
This document is for informational purposes only. Any forecasts, estimates, targets, prospects, and/or opinions discussed herein are subject to change and may differ from or contradict those expressed by others. Forward-looking statements are based on various assumptions, risks, and uncertainties that could cause actual results, performance, or achievements to differ materially from those expressed or implied. This document is not intended for any investor or potential investor, nor is it an invitation to provide advisory services. The information presented should not be construed as specific investment advice or final investment recommendations and should not be relied upon as such. The assets discussed do not represent all assets invested by Polychain Capital, and no assumption may be made that any identified investments will be or have been profitable. Due to changes in market, economic, political, regulatory, or other conditions, there is no guarantee that results will resemble any historical outcomes described herein. Nothing herein should be construed or relied upon as investment, legal, tax, or other professional advice.