Author: Dolphin
Compiled by: Yuliya, PANews
Editor's Note: Recently, the AI narrative within the Base ecosystem has experienced explosive growth, with the privacy-first generative AI platform Venice ($VVV) and its ecosystem projects attracting the most attention. ( Related reading: The "VVV" concept, which has surged 9 times in six months, is the new main theme of AI in the Base ecosystem .) As a co-developer of Venice's core default model, the Dolphin network and its token $POD performed remarkably in May, with its market capitalization soaring from $12.2 million to $192 million, an increase of over 14 times. This article details the Dolphin network's unique "peer-to-peer" economic model, token value capture mechanism, and innovative design for ensuring network security through staking and penalty mechanisms. The following is a detailed breakdown and analysis of the mechanisms:
Peer-to-Pool Economic Model Design
The Dolphin network is designed as a peer-to-pool system to reuse idle GPUs. Each AI model runs on a GPU provided by the network.
This differs from most AI DePINs. In other networks, buyers typically rent a node directly from the provider, establishing a one-to-one "session."
On the supply side, nodes running the same model form a "pool" to collectively process task requests from others. The system randomly assigns tasks based on node availability, and there is no direct connection between the requester and the node provider. The sole criterion for nodes to earn rewards is the number of AI computing tasks (i.e., inference tokens) they process, with rewards paid out in POD tokens issued by the protocol treasury.
On the demand side, users using the API directly purchase quotas from the protocol. The Dolphin network accepts payments in various cryptocurrencies, including $POD, $ETH, $BTC, $USDC, $XMR, and $ZEC.
All revenue received by the protocol will be used to buy back POD tokens on the market – this directly offsets any increase in token issuance.
Buyers and sellers are separate entities, which means that the POD reward issued to a node can be more or less than the POD we earn from income.
To illustrate this more clearly, let's look at a specific example of running the Qwen3.6-35B model on a Dolphin network:
The current cost of running datagen.dphn.ai is $0.50 per million tokens processed.
The cheapest competing product on OpenRouter costs $1.00 per 1 million tokens.
Dolphin charges users $0.70.
Dolphin pays the node $0.50.
Net buyback funds: $0.20 is generated for every 1 million tokens created.
In other words, the Dolphin network not only prices its tokens 30% lower than the cheapest centralized providers, but also generates $0.20 in pure profit for every million tokens generated, which is used to buy PODs on the market.
Why is this the best use case for DePIN?
This model is considered a highly promising application area in the DePIN field, mainly for the following reasons:
Extremely high demand for AI inference: The market's desire for AI inference computing power is in a period of explosive growth.
A massive pool of idle computing power : The supply of idle gaming GPUs capable of running local AI models is enormous. This network model feels similar to the old-school Proof-of-Work (PoW) cryptocurrency mining, but because it produces truly commercially valuable AI computing, its earning potential is much greater.
Ignoring geographical limitations: Unlike many DePIN networks, the geographical location of AI inference is not critical, thus avoiding coverage issues. Because AI inference is highly flexible geographically, a latency of a few hundred milliseconds has a negligible impact on user experience. This allows the Dolphin network to connect consumers and computing resources globally, significantly improving the scalability and utilization of each node.
The necessity of liquidity pooling computing : It's the only way to unlock the largest GPU supply group (gamers and PC enthusiasts). It allows nodes to go online or offline at any time, without requiring a fixed online time like P2P node rentals. Previous GPU DePIN projects required a one-to-one binding between consumers and nodes, which was simply unworkable for idle GPUs like those in gaming PCs or data center graphics cards, as the owner might want to take the computer back for personal use at any time. After all, nobody wants a rented GPU that suddenly goes offline after being taken back by its owner.
Token Mechanism and Value Accumulation
POD is the only valuable asset in the Dolphin ecosystem. 100% of all revenue generated by the network is automatically used to buy back POD on the market. Furthermore, Dolphin has no external equity structure based on shareholders, and never will.
For POD holders, staking tokens into the xPOD vault can grant multiple exclusive benefits:
Receive direct, automatic compound interest dividends from network token buybacks.
Get daily AI inference credits, allowing you to use all models online for free.
Enjoy premium subscription status in Dolphin's web chat rooms, bots, and other ecosystem applications.
In designing its token economics, Dolphin drew on the essence of many excellent DeFi projects and deeply integrated these aspects that best fit distributed AI inference and training networks:
Borrowing from the ETH mechanism: Node operators and validators are required to pay a deposit, which will be forfeited if malicious behavior occurs.
Drawing inspiration from the CRV mechanism, this system offers reward acceleration features for node operators. Locking POD can potentially double earnings, and based on deposit-to-return ratios on other platforms, an acceleration multiplier of 1.5 to 2x is highly competitive in the market.
Drawing inspiration from the xSUSHI/yCRV mechanism , an automatically compounding staking vault is introduced. Users do not need to manually claim rewards, meaning that xPOD (i.e., staked Dolphin tokens) can be directly used as collateral for node operators.
Drawing inspiration from the stAAVE mechanism , a reasonable withdrawal cooling-off period and withdrawal time window have been set to ensure the stability of network funds.
Drawing inspiration from vlCVX/veCRV, a "bribery market" has been established for unused xPOD computing credits each day. Users can sell unused computing credits to earn higher staking rewards.
Deposit binding, violation fines and reward doubling mechanism
In decentralized computing networks, cheating is undoubtedly the biggest threat. If left unchecked, node operators could secretly replace their AI models with smaller, crippled, or even fake ones, and still collect rewards. This would cause the output quality to collapse, those who bought computing power to run away, and the entire ecosystem's flywheel would never start spinning.
To address this challenge, the Dolphin network introduced a "deductible deposit" mechanism, deeply aligning the interests of node operators with the value of the POD token . If malicious cheating is detected, the node will have its deposit equivalent to four weeks' earnings deducted directly. This makes cheating extremely uneconomical.
By default, node operators earn POD in a "bound state" . Once a node has accumulated enough bound POD equivalent to four weeks' worth of income, they can choose at the weekly settlement whether to continue receiving bound POD or receive liquid POD that can be traded at any time.
If you choose to claim liquidity POD, the system will deduct a 20% transaction fee. This money will be directly deposited into the xPOD staking pool, which will then be distributed among other stakers and node operators who have committed their stake.
Nodes can further store xPOD into binding contracts, which not only increases their earnings but also grants them the qualification to validate other nodes in the network.
The POD reward multiplier determines how much extra money a node can earn on top of its base reward. This mechanism is inspired by Curve Finance's liquidity provider (LP) acceleration mechanism, but Dolphin has made specific modifications for decentralized AI networks, adding features such as rewards based on usage, unified calculation of deposits across all accounts, and penalties for violations.
In short:
Nodes earn basic rewards by completing AI calculations, verification tasks, and related protocol tasks.
The system will multiply your earned node rewards by a factor based on the amount of tokens linked to your account and your earnings percentage.
When calculating the earnings ratio, the system looks at your average base reward over the past few weeks and uses a smoothing algorithm that "rises quickly and falls slowly": when you take on more computing tasks, your average earnings metric will rise rapidly; but when you have less free time, it will fall very slowly.
You are eligible to become a validator if your account has maintained deposits with earnings for more than 3 months and has at least 50,000 active deposits.
If your deposit is equivalent to 6 months (26 weeks) of earnings, the system guarantees your rewards will be multiplied by at least 1.5.
If your deposit exceeds 6 months of earnings, your reward multiplier can reach up to 2x. The exact multiplier depends on your relative proportion to other participants who have exceeded their 6-month target, and the absolute amount by which you exceed the target.
All calculations are measured solely in the number of PODs; the reward system does not involve any fiat price oracles. Deposits are calculated per account (wallet), and the calculated reward multiplier applies to all nodes under your account. If you add more nodes, your total account earnings will increase, so you need to increase your active deposit proportionally to maintain the original earnings multiplier.
Finally, the Dolphin Network will release its paper "Encrypted Live-Weight Proofs for Decentralized Inference" tomorrow. This paper will detail a lightweight verification system capable of verifying whether nodes on various hardware are running the correct model, surpassing standard TEE verification which can only be used on enterprise NVIDIA graphics cards.
