Buy Crypto Markets Spot FuturesBTC Earn Event Center Rewards Hub

Overview Dolphin is an AI infrastructure project focused on decentralized model inference and distributed GPU computing. Its flagship product, Dolphin Network, is designed to aggregate idle GPUOverview Dolphin is an AI infrastructure project focused on decentralized model inference and distributed GPU computing. Its flagship product, Dolphin Network, is designed to aggregate idle GPU

Learn/Project Spotlight/What Is Dolphin (POD)? How Idle GPUs Become Decentralized AI Inference Capacity

What Is Dolphin (POD)? How Idle GPUs Become Decentralized AI Inference Capacity

Author: MEXC

Beginner

Oct 9, 2026MEXC

9 min

#Beginners

Overview

Dolphin is an AI infrastructure project focused on decentralized model inference and distributed GPU computing. Its flagship product, Dolphin Network, is designed to aggregate idle GPU resources from gamers and other hardware owners and use that capacity to process AI inference workloads. Instead of relying entirely on centralized data centers, the network aims to coordinate distributed computing resources while using encryption, randomized validation, and cryptoeconomic mechanisms to protect the integrity of submitted work.

POD is the Base-based token currently listed by MEXC under the Dolphin project, with a disclosed total supply of 500 million tokens. MEXC opened POD/USDT spot trading in the Innovation Zone on October 2, 2026. Dolphin has also continued developing its distributed inference infrastructure, including a major Network V2 upgrade focused on automated updates, GPU utilization, routing, and load balancing.

Key Takeaways

Dolphin is building a decentralized AI inference network that uses distributed GPU resources.
GPU owners can contribute otherwise idle computing capacity to network workloads.
Dolphin Network uses encryption, randomized validation, and cryptoeconomic mechanisms to help verify distributed computation.
The current POD token is issued on Base with a disclosed total supply of 500,000,000 tokens.
MEXC supports POD/USDT spot trading in the Innovation Zone.

What Is Dolphin?

Dolphin is an AI project focused on two connected areas: AI model development and distributed inference. The project’s broader goal is to make GPU computing resources available to AI workloads without requiring all inference capacity to come from centralized cloud infrastructure.

Dolphin Network is the project’s main infrastructure product. It is designed to connect GPU owners with computational workloads so that hardware that would otherwise remain unused can contribute to AI inference. A gamer, workstation owner, or other GPU operator can potentially make computing resources available to the network, while Dolphin coordinates how workloads are distributed and processed.

The project also develops AI models and tools alongside the network itself. Its public channels describe Dolphin as an AI lab working on models and distributed inference, making the project broader than a simple GPU marketplace. The infrastructure layer is intended to support model execution, while the project’s AI work explores how the available distributed compute can be used for inference, data generation, and model-related workloads.

How Does Dolphin Network Work?

Distributed GPU Inference

AI inference is the process of running a trained model to generate an output from a new input. Large language models and other generative AI systems can require substantial GPU memory and processing capacity, particularly when many requests are handled simultaneously.

Dolphin Network approaches this problem by aggregating distributed GPUs. Instead of requiring every GPU to remain continuously dedicated to the network, the model is designed around hardware owners contributing otherwise unused capacity. Network workloads can then be distributed across participating resources.

This gives Dolphin a different infrastructure model from a conventional centralized AI API. In a centralized environment, one provider owns or rents the majority of the servers. In a decentralized network, computing capacity can come from a broader set of independent operators, making coordination and verification central technical problems.

Validation and Network Integrity

Distributed computing creates an important question: how can the network determine whether a remote GPU actually performed the requested work correctly?

Dolphin’s public project description says network integrity is supported through a combination of encryption, randomly sampled validation, and cryptoeconomic bonding. Rather than automatically trusting every participating machine, the architecture is intended to introduce verification and economic accountability into the processing layer.

This distinction matters because a decentralized AI network is not simply a marketplace for renting GPUs. Reliable inference also requires the network to coordinate requests, identify available resources, validate computation, and maintain service quality when participating machines have different hardware specifications or availability.

Routing and GPU Utilization

Routing is another important component of distributed inference. Different AI models may require different amounts of memory, GPU architecture, and compute capacity, while individual nodes may enter or leave the network.

Dolphin Network therefore needs to match workloads with suitable computing resources and distribute requests efficiently. The project’s Network V2 update specifically highlighted improved routing and load balancing as mechanisms for increasing network throughput and GPU utilization.

The same upgrade rebuilt the network architecture in Golang and introduced automatic updates intended to reduce manual node migrations. Dolphin also added NVFP4 as the default format for Blackwell GPUs, reflecting an effort to optimize inference for newer NVIDIA hardware.

Why Idle GPU Capacity Matters for AI

Demand for AI computation has expanded rapidly as larger models and inference-intensive applications become more widely used. At the same time, a substantial amount of privately owned GPU capacity is not continuously utilized. Gaming GPUs, workstations, research machines, and other hardware may remain idle for significant periods.

Dolphin attempts to connect those two conditions. Rather than treating idle hardware as unavailable capacity, the network is designed to aggregate it into a usable inference layer. For GPU owners, this creates a potential mechanism for contributing computing resources. For developers and AI applications, the longer-term goal is to provide an additional source of inference capacity outside traditional centralized infrastructure.

The practical challenge is reliability. Consumer and independently operated GPUs differ in memory, bandwidth, operating environment, and uptime. A decentralized network therefore needs more than raw compute supply: scheduling, verification, routing, software compatibility, and workload management are all necessary for the model to work at scale.

Dolphin Network V2 and Recent Development

Dolphin announced Network V2 as its first major infrastructure upgrade after the network’s initial launch. According to the project’s official updates, the V2 architecture was rebuilt in Golang and introduced automatic node updates, improved routing and load balancing, and additional optimization for newer GPU architectures.

The project has also published network activity updates as it stress-tests distributed inference and data-generation workloads. One official update reported more than 1,000 GPUs online, approximately 49 TB of aggregate VRAM, and large-scale token generation using Qwen-family models. These figures represent project-reported network activity and should be understood in the context of an infrastructure network that continues to evolve.

Dolphin has separately indicated plans or development work around broader hardware support, including AMD GPUs, CPUs, macOS, and Windows environments, as well as public inference APIs, image and audio generation, and sharded inference. These items should be treated as planned or under development where they have not yet been publicly released as production features.

What Is the Difference Between Dolphin Network and POD?

Dolphin Network is the decentralized AI infrastructure. It encompasses GPU nodes, workload routing, inference processing, validation mechanisms, and the software required to coordinate distributed compute.

POD is the Base-based crypto asset currently associated with Dolphin in MEXC’s listing information. The two should not be treated as interchangeable: the network describes the underlying infrastructure, while POD is the tradable digital asset.

There is also an important terminology issue in currently available public materials. Dolphin’s project description states that GPU contributors can earn “DPHN” tokens that may later be used for inference or sold, while the asset currently listed by MEXC is POD. The public official materials reviewed for this article do not clearly explain whether DPHN is an earlier name, a separate reward unit, or how its described network functions relate to POD.

For this reason, functions explicitly described for DPHN should not automatically be presented as confirmed POD utility until Dolphin publishes clearer token documentation.

POD Tokenomics and Utility

MEXC’s official listing announcement confirms that Dolphin (POD) operates on Base and has a total supply of 500,000,000 POD.

Item	Details
Project	Dolphin
Token	POD
Network	Base
Sector	AI
Total Supply	500,000,000 POD
MEXC Spot Pair	POD/USDT

A complete official public allocation table covering team, community, ecosystem, treasury, investors, node incentives, market making, and other categories was not identified in the official materials reviewed for this article. A comprehensive POD vesting and unlock schedule was also not available.

Users should therefore avoid assuming an allocation structure or unlock mechanism that has not been formally disclosed. The same applies to POD-specific staking, governance, fee-sharing, burn, or network-payment functions.

The clearest confirmed information is that POD is the Base token associated with Dolphin’s current market listing. Dolphin’s network description separately discusses DPHN in connection with GPU rewards and inference usage, so those functions should remain distinguished until the project provides an explicit explanation of the relationship between the two token names.

Why Decentralized AI Inference Is Different From GPU Rental

At first glance, decentralized inference may appear similar to a conventional GPU rental marketplace, but the requirements are different. Renting a GPU generally gives a user access to computing hardware for a defined period. An inference network must instead coordinate incoming requests and continuously determine where each workload should be processed.

This requires a layer for routing, availability, model compatibility, validation, and load balancing. Nodes may also differ substantially in their capabilities. A consumer GPU suitable for one model may not have enough memory for another, while newer data-center GPUs may support different numerical formats and higher-throughput workloads.

Dolphin’s approach therefore focuses not only on attracting GPU supply but also on turning distributed hardware into coordinated inference capacity. If successfully implemented at scale, the relevant product is not simply access to a machine; it is the ability to send an AI workload to a distributed network and receive a usable result.

Dolphin’s Role in the Decentralized AI Infrastructure Market

Decentralized AI infrastructure has developed as projects explore alternatives to relying exclusively on centralized cloud-computing providers. Different networks focus on areas such as model training, inference, GPU marketplaces, data generation, verification, or distributed storage.

Dolphin’s current emphasis is distributed inference, with idle GPU utilization serving as the supply side of the network. The project has also used its available capacity for data-generation and model-development workloads while building toward broader inference access.

For users evaluating Dolphin, the important measures are therefore not only token-market activity but also whether the network can attract reliable compute capacity, support relevant AI models, maintain workload integrity, and convert distributed hardware into useful inference services.

Where and How to Trade POD on MEXC

MEXC opened POD/USDT spot trading in the Innovation Zone on October 2, 2026. Users who want to trade POD can generally follow these steps:

Create or sign in to an MEXC account.
Search for Dolphin or POD.
Confirm that the selected asset is Dolphin (POD) on Base.
Deposit USDT or use another currently available funding method.
Open the POD/USDT spot market.
Choose a market or limit order according to the available trading options.
Review the price, quantity, and order details before confirming.
If withdrawing POD, verify the supported network and destination wallet before submitting the withdrawal.

Users should verify the token name and supported blockchain rather than relying only on the ticker, particularly because identical or similar ticker symbols can exist across the crypto market.

Conclusion

Dolphin is building decentralized infrastructure intended to turn distributed GPU resources into usable AI inference capacity. Its approach combines idle GPU participation with routing, validation, encryption, load balancing, and cryptoeconomic mechanisms, while Network V2 represents a significant technical upgrade focused on making the infrastructure easier to operate and more efficient.

POD is the Base-based asset currently associated with Dolphin’s market listing and has a disclosed total supply of 500 million tokens. However, users should distinguish confirmed POD information from network documentation that refers to DPHN for rewards and inference. Before trading POD/USDT, users should verify the latest Dolphin token documentation, supported network, current MEXC market information, and any future clarification of POD’s specific role within Dolphin Network.

Risk Disclaimer: This article is for informational and educational purposes only and does not constitute financial, investment, legal, or tax advice. Cryptocurrency prices are highly volatile, and users may lose part or all of their invested capital. Project features, token utility, network architecture, roadmaps, and trading availability may change over time. Always verify information through official Dolphin and MEXC channels and conduct independent research before making any trading decision.