Is AI business unprofitable? The dawn of DeAI has appeared.

Author: Zhang Feng

Artificial intelligence (AI) is undoubtedly the hottest technology trend globally, and AI technology is reshaping various industries at an unprecedented speed. However, behind the booming and noisy scene, a harsh reality is that the vast majority of AI businesses, especially startups, have not found a stable and sustainable path to profitability. They are caught in the predicament of "praised but not commercially viable," with technological prosperity coexisting with commercial losses.

I. Why "Losing Money to Gain Popularity"?

The profitability dilemma of AI businesses does not stem from the failure of the technology itself, but rather from the structural contradictions caused by its centralized development model.

Specifically, this can be attributed to the following three main reasons: Extreme Centralization: Exorbitant Costs and Oligopolistic Monopolies. Current mainstream AI, especially large-scale models, is a typical "asset-heavy" industry. Its training and inference processes consume massive amounts of computing power (GPUs), storage, and electricity. This leads to polarization: on one hand, there are tech giants with substantial capital (such as Google, Microsoft, and OpenAI), capable of investing hundreds of millions or even billions of dollars; on the other hand, there are numerous startups that have to "contribute" the vast majority of their funding to cloud service providers to obtain computing power, resulting in extremely squeezed profit margins. This model has created a "computing power oligopoly," stifling innovation. For example, even OpenAI, in its early development, heavily relied on Microsoft's massive investment and Azure cloud computing resources to support the research and operation of ChatGPT. For the vast majority of players, the high fixed costs make it difficult to achieve profitability at scale. Data Dilemmas: Quality Barriers and Privacy Risks. Data is the fuel of AI. Centralized AI companies typically face two major challenges in acquiring high-quality, large-scale training data. First, data acquisition is extremely costly. Whether through paid collection, data labeling, or utilizing user data, it involves huge investments of money and time. Second, data privacy and compliance risks are enormous. With the tightening of global data regulations (such as GDPR and CCPA), the collection and use of data without explicit user authorization can easily lead to legal action and hefty fines. For example, several well-known technology companies have faced hefty fines for data usage issues. This creates a paradox: AI cannot develop without data, but acquiring and using data is extremely difficult. Imbalance in Value Distribution: Contributors and Creators Excluded from Revenue. In the current AI ecosystem, value distribution is extremely unfair. AI model training relies on countless user-generated behavioral data, content produced by creators (text, images, code, etc.), and open-source code contributed by developers worldwide. However, these core contributors receive almost no reward from the enormous commercial value created by AI models. This is not only an ethical issue but also an unsustainable business model. It dampens the enthusiasm of data contributors and content creators, and in the long run, erodes the foundation for the continuous optimization and innovation of AI models. A typical example is that many artists and writers have accused AI companies of using their work for training and profit without providing any compensation, sparking widespread controversy and legal disputes. II. A New Profit Paradigm DeAI (Decentralized AI) is not a single technology, but a new paradigm that integrates blockchain, cryptography, and distributed computing. It aims to restructure the production relations of AI through decentralization, thereby specifically addressing the three major pain points mentioned above and opening up possibilities for profitability. DeAI uses a "crowdsourcing" model to distribute computing power demand to idle nodes (personal computers, data centers, etc.) globally. This is similar to "Airbnb for GPUs," forming a global, competitive computing power market that can significantly reduce computing power costs. Participants receive token incentives by contributing computing power, achieving optimized resource allocation. DeAI utilizes technologies such as federated learning and homomorphic encryption to achieve "the model moves while the data remains stationary." It avoids centralizing raw datasets, instead distributing models to various data sources for local training, aggregating only encrypted parameter updates. This fundamentally protects data privacy while legally and compliantly utilizing the value of decentralized data. Data owners can independently decide whether to provide data and profit from it. DeAI constructs a transparent and fair value distribution system through token economics and smart contracts. Data contributors, computing power providers, model developers, and even model users can automatically receive corresponding token rewards through smart contracts based on their contributions. This transforms AI from a "black box" controlled by giants into an open economy built, governed, and shared by the community. III. Transforming into a Three-Layer Architecture Migrating traditional centralized AI businesses to the DeAI paradigm requires a systematic restructuring across three levels: technology, business, and governance. (I) Technological Restructuring from Centralized to Distributed Systems (I) Computing Power Layer: Relying on Decentralized Physical Infrastructure Network (DePIN) projects, such as Akash Network and Render Network, a flexible and low-cost distributed computing power pool is built to replace traditional centralized cloud services. Data Layer: Federated learning is adopted as the core training framework, combined with cryptographic technologies such as homomorphic encryption and secure multi-party computation to ensure data privacy and security. Establish a blockchain-based data marketplace, such as Ocean Protocol, to enable data transactions under the premise of established ownership and security. The model layer involves deploying trained AI models on the blockchain in the form of "AI smart contracts," making them transparent, verifiable, and permissionless. Every use of the model and the resulting revenue can be accurately recorded and distributed. (II) Business Restructuring from Selling Services to Building an Ecosystem From SaaS to DaaS (Data as a Service) and MaaS (Model as a Service), enterprises are no longer simply selling API call counts, but rather acting as ecosystem builders, incentivizing community participation in network construction by issuing utility tokens or governance tokens. Revenue sources have expanded from single service fees to include token appreciation resulting from ecosystem value growth and transaction fee dividends. Therefore, building a decentralized task platform that publishes tasks such as data annotation, model fine-tuning, and application development for specific scenarios in the form of "bounties" allows global community members to undertake these tasks and receive rewards, greatly reducing operating costs and stimulating innovation. (III) From Corporate System to DAO: Governance Restructuring Based on Community Governance. By holding governance tokens, community participants (contributors, users) have the right to vote on key decisions, such as the direction of model parameter adjustments, the use of treasury funds, and the development priority of new features. This achieves true "users are owners." Based on openness and transparency, all code, models (some of which can be open-source), transaction records, and governance decisions are put on the blockchain to ensure the process is open and transparent, establishing a trustless collaborative relationship. This in itself is a powerful brand asset and a guarantee of trust. Taking the transformation of traditional logistics data platforms to DeAI as an example, the dilemma of traditional logistics data platforms is that although they aggregate data from various parties such as maritime transport, land transport, and warehousing, participants are "unwilling to share" due to concerns about the leakage of trade secrets, resulting in data silos and limited platform value. The core of the DeAI transformation is to release the value of data and provide fair incentives without exposing the raw data: Technically, a trusted computing network is built. The platform no longer centrally stores data but transforms into a blockchain-based coordination layer. Employing technologies such as federated learning, AI models are "airdropped" to the local servers of various enterprises (such as shipping companies and warehouses) for training, aggregating only encrypted parameter updates to collectively optimize the global prediction model (such as cargo ship arrival times and warehouse overload risks), achieving "data remains still, value moves." In terms of business, data assetization and token incentives are implemented. The platform issues a practical points system where logistics companies earn points by contributing data (model parameters) to "mine" for rewards. Downstream customers (such as cargo owners) pay tokens to query high-precision "predictions" (e.g., the on-time rate of a certain route for the next week), rather than purchasing raw data. Rewards are automatically distributed to data contributors via smart contracts. In terms of governance, the platform is built around an industry-wide DAO (Data Access Organization). Key decisions (such as new feature development and fee adjustments) are governed by a vote of token holders (i.e., core participants), transforming the platform from a privately owned company to an industry community. The platform has transformed from a centralized institution attempting to extract data intermediary fees into a nervous system for the entire logistics industry chain to co-build, co-govern, and share resources. By solving trust issues, it has greatly improved industry collaboration efficiency and risk resistance. IV. Compliance and Security While DeAI has a promising future, its development is still in its early stages and faces a series of challenges that cannot be ignored. Compliance and legal uncertainty. Regarding data regulations, even if data is not moved, models such as federated learning, when processing personal data, must still strictly comply with the requirements of GDPR and other regulations regarding "purpose limitation," "data minimization," and user rights (such as the right to be forgotten). Project teams must design compliant data authorization and exit mechanisms. Regarding securities regulations, tokens issued by projects are highly likely to be identified as securities by regulatory agencies in various countries (such as the US SEC), thus facing strict regulatory scrutiny. How to avoid legal risks when designing a token economic model is crucial for the project's survival. Regarding content liability, if an on-chain DeAI model generates harmful, biased, or illegal content, who is responsible? Is it the model developer, the computing power provider, or the governance token holder? This presents new challenges to the existing legal system. Regarding security and performance challenges, model security—that is, models deployed on public chains may face new attack vectors, such as exploits of smart contract vulnerabilities or malicious attacks on federated learning systems through data poisoning. Performance bottlenecks, namely the blockchain's inherent transaction speed (TPS) and storage limitations, may not be able to support high-frequency, low-latency inference requests for large models. This necessitates an effective combination of Layer 2 scaling solutions and off-chain computation. Collaboration efficiency, while fair, may result in lower decision-making and execution efficiency compared to centralized companies. Finding a balance between efficiency and fairness is an art that DAO governance needs to continuously explore. DeAI, as a revolution in production relations, through distributed technology, token economics, and community governance, has the potential to break the monopoly of giants, release the value of idle computing power and data globally, and build a more equitable, sustainable, and potentially more profitable new AI ecosystem. V. Current Exploration Directions The development of current AI tools still has a long way to go before achieving ideal decentralized artificial intelligence. We are currently in an early stage dominated by centralized services, but some explorations have already pointed the way forward.

Current Exploration and Future Challenges.Although the ideal DeAI has not yet been realized, the industry is already making valuable attempts, which helps us see the future path and the obstacles that need to be overcome.

Such as the prototype of multi-agent system collaboration.

Some projects are exploring the creation of environments where AI agents collaborate and co-evolve. For example, the AMMO project aims to create a "symbiotic network of humans and AI," with its multi-agent framework and RL Gyms simulation environment allowing AI agents to learn collaboration and competition in complex scenarios. This can be seen as an attempt to build the underlying interaction rules of the DeAI world. Another example is the initial experiment with incentive models. In the DeAI vision, users who contribute data and nodes that provide computing power should receive fair rewards. Some projects are attempting to redistribute value directly to contributors in the ecosystem through cryptographic incentive systems. Of course, how this economic model can operate on a large scale, stably, and fairly remains a significant challenge. For example, moving towards more autonomous AI: Deep Research-type products demonstrate the powerful autonomy of AI in specific tasks (such as information retrieval and analysis). They can autonomously plan, execute multi-step operations, and iteratively optimize results. This task automation capability is the foundation for the independent operation of AI agents in future DeAI networks. For AI practitioners struggling in a red ocean market, instead of being trapped in the old paradigm, it's better to bravely embrace the new blue ocean of DeAI. This is not only a shift in technological approach, but also a reshaping of business philosophy—from "exploitation" to "incentive," from "closed" to "open," from "monopoly profits" to "inclusive growth."