Swan Chain Delivers 100,000 GPU Compute Hours to Nebula Block

Key Insights

• Swan Chain is an Ethereum L2 for decentralized AI infrastructure. It offers compute, storage, bandwidth, and payments via a global network of providers.
• Swan Chain’s AI compute marketplace uses a smart contract-based auction system to match tasks with providers. Universal Basic Income (UBI) rewards paid in SWAN ensure stable participation and broad access.
• In February 2025, Nebula Block partnered with Swan Chain to purchase over 100,000 GPU compute hours. Nebula Block reported 50-70% cost cuts compared to centralized cloud providers.
• Mission 3.0 is Swan’s quest platform tied to AI-generated decentralized identities (DIDs). Projects like T-RIZE use it to run campaigns, reward users, and grow verified Web3 communities.

Primer

Swan Chain (SWAN) is an Ethereum Layer-2 (L2) blockchain developed on the OP stack. Initially launched in 2021 as FilSwan, a decentralized Filecoin storage solution, the platform evolved into Swan Chain, expanding into decentralized AI solutions across computing, storage, bandwidth, and payments. By leveraging underutilized computing resources from a global network of data centers, Swan Chain seeks to reduce AI deployment costs and facilitate the monetization of idle compute.

Swan Chain incorporates Kubernetes, an open-source tool that automates the deployment and management of containerized applications, for container orchestration, i.e., management of application code across multiple machines. To enhance privacy and security, Swan uses zero-knowledge (ZK) proofs, which verify computations without exposing underlying data. Its ecosystem includes Lagrange, a decentralized computing platform, Swan IPFS storage (formerly MultiChain.storage) for cross-chain storage solutions, and the Swan SDK for seamless integration with its infrastructure. The platform also features a Universal Basic Income (UBI) model to ensure fair compensation for computing providers within its decentralized AI computing marketplace.

SwanFi is a staking platform on Swan Chain that tokenizes compute resources into liquid assets. Users stake SWAN tokens to receive fSWAN, earning rewards from computing providers while maintaining liquidity. As of June 1, 2025, 11 million SWAN tokens are staked on SwanFi, with an estimated annual percentage rate (APR) of 19.6%.

The SWAN token has a total supply of 1 billion and TGE’d on Dec. 16, 2024. Users pay Market Providers in SWAN for storage and compute resources. SWAN is also used as collateral by Market Providers to ensure they do not misbehave. Additionally, SWAN is the governance token for Swan DAO, which manages Swan Chain. Major investors include Binance Labs, Chainlink, and Filecoin. For a full primer on Swan Chain, refer to our Initiation of Coverage report.

Website / X (Twitter) / Discord

Shift to Decentralized AI

AI development is increasingly dominated by centralized compute infrastructure, where a few cloud providers control access, and costs continue to rise. This reality has led to greater interest in decentralized systems that distribute compute workloads and reduce reliance on centralized services.

Additionally, decentralized computing is suitable for edge computing, which allows models to run directly on or near devices, enabling fast inference without sending data back and forth to the cloud or a data center. This reduces latency, lowers bandwidth usage, and allows real-time decision-making in environments with limited or intermittent connectivity.

Following its transition from FilSwan, a Filecoin-based storage provider, Swan Chain has shifted its focus to decentralized AI infrastructure. The AI compute marketplace, where Compute Providers lend their available compute for user task requirements, is Swan’s core AI offering. Compute Providers earn SWAN tokens for completed work. Swan’s GPU usage hit 44% on May 25, 2025, the highest usage rate in 2025. GPU usage is a reflection of the percentage of the GPU’s that are currently being utilized during inference or training.

To ensure stable participation, Swan Chain uses a Universal Basic Income (UBI) allocation curve to distribute SWAN rewards based on relative contribution to the network. The UBI model prioritizes providers with higher-quality hardware and sustained uptime while ensuring lower-tier providers receive a minimum baseline reward. Rewards (paid job income) increase with contribution, but at a diminishing rate, preventing a small number of high-end providers from capturing a disproportionate share. This design encourages broad participation and maintains decentralization by balancing incentives between high and low-capacity providers. The curve adjusts dynamically based on network conditions, helping stabilize supply and reward distribution over time.

For Swan Chain, lower costs for AI tasks are a key value proposition of decentralized computing. Swan achieves this with the Swan Orchestrator, which schedules and assigns AI tasks to available Compute Providers. It uses a smart contract-based task auction system, where providers bid on tasks based on resource availability and pricing. The orchestrator selects providers through a scoring mechanism that accounts for reputation, hardware specs, and network latency. This process ensures tasks are matched efficiently while maintaining quality and cost control across the network. Jarod Zhao, co-founder of Nebula Block, reported that using the Swan Chain network reduced AI inference costs by approximately 50% compared to traditional providers.

Recent Partnerships

Nebula Block

On Feb. 4, 2025 Nebula Block announced a partnership with Swan Chain to acquire over 100,000 GPU compute hours. GPU hours refer to the amount of time a GPU is used for computational tasks, such as AI inference or model training. With Swan Chain’s GPU hours and AI infrastructure, Nebula can offer inference at 50–70% lower prices than centralized providers like AWS and Google. Swan Chain lowers costs by using idle resources and dynamic pricing to match supply with demand more efficiently than centralized cloud providers.

Swan Chain’s compute layer includes Edge Computing Providers (ECPs), which handle low-latency inference near end users, and Fog Computing Providers (FCPs), which manage orchestration and large-model training. Swan Chain’s compute infrastructure spans 24 global cities, with over 100 Computing Providers. This setup enables Nebula to allocate workloads efficiently dynamically across providers and geographies and deploy models such as Llama-70B and DeepSeek at reduced cost.

Other Partnerships

• Spheron Network (Sept. 6, 2024): Spheron is a decentralized GPU marketplace for renting and monetizing compute. It is integrated with Swan Chain as a compute provider, allowing Spheron nodes to serve AI workloads on Swan. This expands Swan’s GPU capacity, while Spheron gains access to Swan’s blockchain infrastructure.
• Rivalz (Sept. 6, 2024): Rivalz builds tokenized AI agents (“rAgents”) and agent collectives (“swarms”) that perform onchain tasks. Its ROME protocol now runs on Swan Chain, using Swan’s compute network to power AI agents.
• Parasail Network (Dec. 10, 2024): Parasail is a restaking protocol that underwrites DePIN services with onchain economic guarantees. The integration with Swan lets compute providers on Parasail access cheaper SWAN staking and use nodes as loan collateral.
• AEON (Feb. 21, 2025): AEON is a crypto payment protocol that abstracts crypto payment complexity with a “scan-to-pay” feature using QR codes. Swan Chain integrated AEON’s tools to enable flexible payments across its ecosystem so that users can pay for compute, storage, or services in any supported token. This improves accessibility for developers and users, helping Swan extend into real-world use cases without building native payment rails.
• Memo Labs (March 5, 2025): Memo Labs provides decentralized storage for large-scale AI data. The partnership allows Swan developers to store training data on Memo and compute on Swan in a unified workflow, creating a decentralized stack for AI applications.
• Lilypad (May 10, 2025): Lilypad is a decentralized compute network optimized for AI and machine learning (ML) workloads. Swan and Lilypad share compute infrastructure, allowing workloads to move between networks. Swan gains overflow compute capacity, while Lilypad gains additional compute power for inference requests.
• Titan Network (May 28, 2025): Titan is a decentralized cloud utilizing idle compute and bandwidth across over 650,000 nodes. The integration with Swan connects Titan’s resources to Swan’s AI compute marketplace, expanding capacity for Swan’s AI workloads while giving Titan access to blockchain-based infrastructure.

Roadmap

Mission 3.0

Mission 3.0 is Swan Chain’s community engagement platform to connect users with Web3 projects through verifiable quests and rewards. Unlike earlier campaigns focused on airdrop eligibility, Mission 3.0 is a permanent system where other projects can launch and manage quest-based campaigns. Users earn points, badges, and reputation by completing onchain and offchain tasks tied to an AI-generated decentralized identity (DID).

The beta launch was announced on March 21, 2025. Initial quests are hosted by T-RIZE, a real-world asset tokenization platform allowing administrators to control quests while users own their data.

Future Citizen

Future Citizen is an upcoming user portal from Swan Chain focused on decentralized identity (DID) and onchain engagement. As of the time of writing, no formal launch date or feature list has been released. However, the public GitHub repository confirms the team is actively developing the project. The project is expected to consolidate key components such as DID minting, profile management, and engagement history into a single interface.

The portal will serve as a decentralized identity hub, allowing users to mint and manage AI-generated DIDs with privacy controls and cross-chain compatibility. It could also integrate Mission 3.0’s questing and rewards system, enabling users to track reputation, earn badges, and claim incentives within one dashboard. Combining user data and credentials could streamline participation across Swan’s ecosystem, from compute access to community governance.

CDN Layer Development

Swan Chain’s CDN/bandwidth layer will improve data transfer between storage and compute resources, reducing latency and increasing throughput for AI workloads. A CDN (content delivery network) is a distributed server system that delivers digital content to users based on their geographic location.

The system could integrate directly with Swan’s storage and compute layers, using distributed nodes to cache and route content. This would improve resilience and lower bandwidth costs, especially for repeated or location-sensitive data requests.

The CDN layer, like the compute layer, is tailored for edge AI scenarios, where models must operate across devices with limited connectivity. Its design prioritizes AI-specific performance needs, making it ideal for Swan Chain’s infrastructure.

Closing Summary

Swan Chain is uniquely positioned at the intersection of decentralized blockchain technology and AI infrastructure, providing a robust platform for comprehensive AI development. Strategic partnerships have validated its decentralized computing model, effectively reducing AI deployment costs and increasing resource efficiency. The collaboration with Nebula Block, securing over 100,000 GPU compute hours through Swan’s AI compute marketplace, exemplifies the platform's ability to deliver substantial cost savings.

Through Mission 3.0, Swan Chain advanced community engagement, enabling richer, analytics-driven interactions. With ongoing infrastructure expansions, such as the CDN layer, Swan Chain is well-positioned to maintain its momentum in decentralized AI innovation, addressing emerging needs with agility and strategic foresight.

Let us know what you loved about the report, what may be missing, or share any other feedback by filling out this short form. All responses are subject to our Privacy Policy and Terms of Service.

This report was commissioned by Swan Chain. All content was produced independently by the author(s) and does not necessarily reflect the opinions of Messari, Inc. or the organization that requested the report. The commissioning organization may have input on the content of the report, but Messari maintains editorial control over the final report to retain data accuracy and objectivity. Author(s) may hold cryptocurrencies named in this report. This report is meant for informational purposes only. It is not meant to serve as investment advice. You should conduct your own research and consult an independent financial, tax, or legal advisor before making any investment decisions. Past performance of any asset is not indicative of future results. Please see our Terms of Service for more information.

No part of this report may be (a) copied, photocopied, duplicated in any form by any means or (b) redistributed without the prior written consent of Messari®.