DGrid Genesis Achieves Over $23 Million in Revenue in Six Months: Decentralized AI Enters Real Paid Verification Stage

The boundaries of AI capabilities are still being explored, and its commercial implementation is being validated through real paid services in the crypto world.

In the first half of 2026, DGrid Genesis's membership program achieved cumulative revenue exceeding $23 million, with over 15,000 paid members. The significance of these figures lies not only in the revenue scale itself but also in the fact that decentralized AI is transitioning from conceptual narratives to a commercial stage that can be validated by real users paying for it.

The key to DGrid's success in this step is that it is not merely a decentralized computing power platform or a tool for model invocation matching; rather, it is a "verifiable intelligent network hub" that connects AI developers, enterprise users, and crypto-native users.

Unlike projects that focus on decentralized training or simply provide a computing power market, DGrid focuses on two things: making the inference process more verifiable and ensuring that the paid closed loop truly works. The former addresses trust issues, while the latter validates commercial demand. This is also the core significance of its underlying Proof of Quality (PoQ) mechanism.

Genesis Program: Proving Commercial Viability with Real Payments

The Genesis membership program is currently DGrid's most representative commercial achievement.

Users can join the Genesis program by paying an annual fee of $1,580, which grants them a monthly model usage credit of $300, exclusive NFT rights, one-click deployment capabilities via DClaw, and rewards in DGAI token mining. Calculated on an annual fee basis, the model invocation rights enjoyed by members each month are equivalent to approximately 44% of the official price, making it highly competitive in the market.

Unlike traditional Web2 AI subscription services, DGrid's membership payment records are based on the blockchain, possessing inherent public transparency. This $23 million in revenue is not merely a narrative on a PowerPoint presentation but is built on credible data derived from real payment behaviors.

More importantly, these over 15,000 paid members represent real and sustained market demand. Their willingness to pay for AI services inherently validates the actual value of the product—something that many purely token narrative projects have never been able to prove.

Beyond the membership system, DGrid allows users to actively participate in blind testing and scoring of models through AI Arena, accumulating high-quality human preference data. This data feeds back into intelligent routing optimization, further enhancing the service quality of Gateway. The validation of paid demand and continuous optimization of product experience together form DGrid's positive cycle.

It is worth mentioning that DGrid's seed round financing amounted to $5 million—leveraging $5 million in financing to achieve $23 million in revenue is quite rare in the entire sector, further validating the health of its business model.

Of course, whether this cycle can be sustained in the long term depends on user retention, credit consumption rates, model cost control, and whether the DGAI incentive mechanism can continue to operate. But at least from the current results, DGrid has taken a significant step in the commercialization of decentralized AI.

Product Matrix: Making AI Truly Usable

Supporting the revenue of the Genesis program is DGrid's product architecture built around users' real pain points.

AI Gateway provides a one-stop intelligent routing entry, automatically invoking over 200 mainstream models such as Claude, GPT, Gemini, MiniMax, and GLM based on cost, speed, and historical performance, allowing users to access high-quality AI services at a lower cost through a single API without needing to connect to different vendors' interfaces separately.

AI Arena collects human preference data on different model outputs through blind testing and scoring. Users anonymously judge the quality of responses from two models, and this data can be used to optimize intelligent routing and further accumulate into commercializable labeled assets. Currently, over 300,000 users are participating in the Arena.

DClaw supports users in rapidly deploying local AI assistants within minutes, allowing direct invocation of top models without the need for key configuration, and provides persistent memory and hot-swappable skill plugins, compatible with mainstream platforms like Telegram, WeChat, and Discord, suitable for long-term stable usage scenarios.

The Model Marketplace supports free listing, self-pricing, and asset tokenization of AI models, allowing model providers to directly settle in and participate in market competition, providing them with a more direct revenue path.

Additionally, DGrid has launched the Dori intelligent recommendation agent, allowing users to describe their needs in natural language, which in turn recommends suitable models and invocation schemes, further lowering the barriers to using multi-model AI services.

The focus of this product matrix is not merely on stacking functions but on addressing a core question: how to enable users to use AI at a lower cost, with more trust, and greater convenience.

Deepening BNB Chain: Endowing Agents with On-Chain Identity and Payment Capabilities

If the product matrix addresses the question of "Is AI usable?", then DGrid's integration with BNB Chain addresses another question: Can AI Agents operate autonomously on-chain?

By integrating Agent Registry and x402 payment capabilities, DGrid enables AI Agents to have on-chain identities and pay-per-use invocation capabilities.

On one hand, AI Agents deployed by DGrid can register on-chain, becoming discoverable, composable, and callable on-chain entities; on the other hand, when AI Agents invoke models on DGrid, they can also complete service settlements through pay-per-request methods.

This means that AI Agents are no longer just off-chain tools but gradually possess on-chain identities, service invocation, and autonomous payment capabilities. It is on this basis that Agents can potentially evolve from a single tool into intelligent entities capable of participating in on-chain economic activities.

PoQ Mechanism: Filling the "Verifiable" Gap for Decentralized AI

The PoQ mechanism aims to address the underlying trust issues of decentralized AI.

In centralized AI services, users typically assume that the platform will complete real invocations, accurate billing, and stable delivery. However, in decentralized networks, model providers, node operators, developers, and users are distributed across different roles, and without a verification mechanism, issues such as low-quality models masquerading as legitimate ones, false billing, and untraceable task execution processes may arise.

The Proof of Quality (PoQ) proposed by DGrid revolves around this issue. It verifies the service quality of model providers through an independent sampling mechanism and records the verification results on-chain—it's important to emphasize that PoQ checks whether the provider honestly delivered the promised model service, using the platform's own test set without touching the user's invocation data or putting user data on-chain.

This is particularly important for decentralized AI. As AI services transition from a single platform to an open network, trust no longer comes solely from brand endorsement but needs to stem from a verifiable, accountable, and billable mechanism. The value of PoQ lies in providing a foundational trust layer for such open AI service networks.

The DGrid team is also continuously researching directions such as PoQ and Optimistic TEE-Rollups. According to public information, its core members have doctoral backgrounds from institutions like Stony Brook University and have published four related papers:

• Proof of Quality: https://arxiv.org/abs/2512.16317
• Optimistic TEE-Rollups: https://arxiv.org/abs/2512.20176
• Cost-Aware Proof: https://arxiv.org/html/2601.21189v1
• PoQ-Judge: https://arxiv.org/pdf/2606.11196

From Technical Validation to Service Network

PoQ initially addressed the question of how a single inference task could be verified, but what DGrid aims to build is not just a "verifiable invocation tool" but a network of AI services connecting model providers, node operators, developers, and end users.

In this network, invocation, verification, billing, and settlement need to be integrated into the same closed loop. Users pay fees, nodes complete services, the system records key processes, and developers and model providers receive revenue. Only when this process is sufficiently transparent can the network transition from early trials to larger-scale real adoption.

This is also one of the important reasons why DGrid has attracted the attention of external institutions. Its seed round attracted participation from Waterdrip Capital, IoTeX, Paramita VC, Zenith Capital, CatcherVC, and others. Waterdrip Capital CEO Jademont pointed out that if decentralized networks prevent builders from understanding the data processing process, they will quickly face execution bottlenecks.

This statement actually highlights a core challenge of decentralized AI: the more open the network and the more participants there are, the higher the trust costs become. If the model inference, data processing, and node execution processes are all opaque, then after scaling up, the system is more likely to fall into inefficiency and untrustworthiness.

DGrid's path is to advance AI services from "callable" to "verifiable, billable, and accountable" through PoQ, on-chain records, and payment mechanisms. This is also the key distinction from purely computing power markets or model aggregation platforms.

Conclusion

DGrid has demonstrated one thing with over $23 million in Genesis membership revenue and a gradually maturing product matrix: verifiable decentralized AI is not just a distant concept; it has begun to enter the stage of real payments and commercial validation.

Of course, this is still a path that requires long-term refinement. How to reduce latency and costs while ensuring verification capabilities, how to smoothly integrate into existing enterprise systems, and how to maintain a balance between membership growth and token incentives will all determine how far DGrid can go in the future.

But it is worth noting that DGrid did not wait until the product matured to add the "trustworthy" narrative; rather, it has embedded "verifiability" into the system design from the ground up.

The market is no longer satisfied with how big a story "AI + Crypto" can tell but is more concerned with how much real revenue it can generate, how many real users it can accumulate, and how many real problems it can solve.

$23 million is the first answer DGrid has provided. What truly needs to be validated next is whether this model can extend from the Genesis membership program to a larger-scale developer ecosystem, model marketplace, and enterprise applications.