Skip to main content

GPU.net / GPUNET

GPU.net / GPUnet (GPU) — a reference card for a crypto project: what the product does and how verifiable the project’s key claims are; also — alignment with the Decentralized AI thesis (DecAI Fit).

GPU.net (GPUNET) — a DePIN / AI/Compute project: a decentralized compute platform for renting GPUs (“on‑demand GPUs”) and an ecosystem around GAN Chain (described in the project materials as an L1 on Substrate with Provider/Validator/Queen roles and its own consensus mechanisms). The $GPU token is described as the “backbone” of the economy: compute rental, staking (Validator/Queen), governance/voting for subnets and a deflationary model (burn gas fees) — per the official documentation.

Disclaimer

Important: this page is a reference overview of the project. Within DecAIHub, we separately record a verification based on open sources within the AI thesis (what is real AI/onchain/why a token, and which sources this is based on).
This is not financial advice and not a recommendation to buy/sell/hold any asset. We do not assess price, returns, or entry/exit timing.
Conclusions are based on publicly available materials found as of the verification date. The absence of confirmation does not imply the absence of the technology.

TL;DR (for readers who are not into AI)

  • What it is: a marketplace/network for renting GPUs and running computations (including for AI) + its own chain infrastructure (GAN Chain).
  • Who it’s for: developers/teams that need GPU resources; providers that have idle GPUs; validators/node operators.
  • Why onchain/token: onchain is needed for the accounting of nodes/orders/rewards and for governance; the GPU token — payment for compute and staking/voting (per the docs).

How it works (very briefly)

  1. The compute consumer does: “rents” a GPU via the dApp/marketplace.
  2. The network does: matches demand with providers (Provider Nodes) and records/validates the work via the validator/queen mechanics (per the GAN Chain description).
  3. The consumer receives: access to computations; the provider/validator receives rewards (per the token docs/Investor docs).
  4. Where the blockchain is: GAN Chain is described as a ledger for the node network and compute proofs; the $GPU token exists onchain (in this card the onchain SSOT of the token is Ethereum ERC‑20 per the explorer link).
  5. Token role: payment for GPU rental, staking (Validator/Queen), governance/voting on subnets; gas fees burn (per the $GPU Token Launch).

Quick links: AI thesis fit

Project passport

Field Value
Project name GPU.net / GPUNET
Also known as (aliases) GPUnet, GPU Network, GAN Chain, $GPU
Logo (logo) ../assets/232_gpunet_logo.svg
Token symbol (ticker) GPU
Category AI/Compute | DePIN
Primary chain (chain) Ethereum
Type non-EVM
Network status mainnet
Consensus (consensus) Proof of Authority / Proof of Availability / Proof of Compute
Launch date (launch_date) 2025-06-06 (TGE in the Investor Docs)
Project status (status) active (website + docs + token)
Token (standard) ERC-20
Token contract 0x79D464248516Bc6977cA2069ba15d8D1044479D8
Token model (token_type) utility (staking + governance)
Max supply Etherscan “Max Total Supply”: 1,750,011 GPU
Issuance fixed supply (as claimed) + burn gas fees (as claimed) + “doubling/halving” issuance model (as claimed)
Tokenomics (tokenomics, brief) 80% community / 10% investors / 10% liquidity (per the $GPU Token Launch); deflationary model via burn gas fees
Governance governance via staking/voting (per the docs)
Admins/keys —
Audit “auditing all smart contracts” is claimed (GAN Chain page), but no reports are provided
Bug bounty —
Team (team) —
Funding (investors/funding) 10% investor allocation (per the token launch); round details — in the Investor Docs (without primary confirmations)
AI component (ai_component) GPU compute + AI agents / inference
Use‑case (use_case) GPU rental for AI/ML
DecAI Fit — AI thesis fit 5/6 — high alignment with the Decentralized AI thesis
DecAI Fit component summary 2 / 3 / 3 / 3
Last verified 2026-02-11

General project overview

GPU.net describes GAN Chain as infrastructure that “brings high‑power GPUs within everyone’s reach”: a network of nodes (Provider/Validator/Queen) and consensus mechanisms for trust in the availability and correctness of computations. In the official docs, the $GPU token is launched with a fixed supply of 200M and is claimed as a token for compute consumption/supply, staking and governance/voting for subnets.

Additionally, the Investor Docs list network/product metrics (data centers, providers, validators, subnets) and a TGE plan for June 6, 2025. At the same time, the onchain supply data on Etherscan conflicts with the claimed supply — this is noted in the passport and in Evidence.

Features/modules

  • GAN Chain: a network of Provider/Validator/Queen nodes and multi‑component consensus (GAN Chain page).
  • On‑demand GPU rentals: a dApp marketplace for renting GPUs (token launch docs with a link to the dapp marketplace).
  • Subnets: a “sub‑network” reward model with voting and an issuance of 1440 $GPU/day per subnet (token launch docs).
  • $GPU token model: fixed supply, allocations, burn gas fees, staking roles (token launch docs).
  • AI/agents: the Investor Docs mention AI agents and AI enhancements (as product directions).

Typical scenarios

  • Scenario 1 (compute consumer): rents a GPU via the dApp/marketplace → runs a workload (AI/rendering/computations).
  • Scenario 2 (provider): connects a GPU as a Provider Node → receives rewards (per the docs).
  • Scenario 3 (validator/queen): stakes $GPU → participates in securing the network and in governance/voting (per the docs).

Getting started (wallet/network)

  • What you need: access to the dApp (marketplace) and a wallet for operations with $GPU.
  • Network/chain: Ethereum (the $GPU token); GAN Chain is mentioned for the infrastructure (the RPC is listed in the GVEX roadmap docs).
  • Fees: Ethereum fees for ERC‑20 operations; gas fees burn is mentioned as part of the model (docs).

Architecture/tech (non‑AI)

  • Components: a network of nodes (provider/validator/queen), marketplace/dApp, subnets, the $GPU token (docs/website).
  • Participant roles: compute consumers; GPU providers; validators/queen; subnet owners.
  • Where logic runs: compute ordering/execution — offchain + onchain accounting/proofs (as claimed); the token is on Ethereum.
  • Dependencies: Ethereum + GAN Chain infrastructure (Substrate) and RPC (rpc.gpu.net is listed in the GVEX roadmap docs).

Token/economics (expanded, no forecasts)

  • Token utility: GPU rental on the marketplace; staking (Validator/Queen) + governance/voting; support for PoW subnets (per the token launch docs).
  • Protocol necessity drivers: if $GPU is the means of payment for compute and the collateral for network roles, the claimed economy does not work without it.
  • Who pays fees: users/network participants (gas fees).
  • Where fees go: gas fees burn (per the token launch docs).
  • Issuance/burn: fixed supply 200M (claimed), but the supply in the explorer conflicts; the issuance dynamics are described as “doubling then halving” with a circulating limit until 2028 (Investor Docs).
  • Security/incentive mechanisms: staking and slashing risks are explicitly mentioned (token launch docs); a reward model for subnets (1440 $GPU/day at the start).
  • Distribution and token issuance schedule: allocations 80/10/10 (token launch docs); detailed circulating dynamics in the Investor Docs (without onchain proof).
  • Treasury: —.

Governance

  • Who makes decisions: participants staking $GPU (Validator/Queen; voting power) — per the token launch docs.
  • How protocol parameters change: via voting (details not disclosed).
  • Where decisions are discussed/made: docs + Discord/Telegram/X.
  • Keys/multisig: —.

Security and incidents

  • Audits: the website claims that the contracts are audited, but links to reports are absent in links_*.
  • Bug bounty: —.
  • Incidents: —.
  • Threat model: risks of provider/validator centralization, the quality of PoCompute/PoAvailability, risks of the token contract and supply inconsistency.

Project status and timeline

  • Current status: the docs and products are active; the token is on Ethereum; GVEX is described as a live deployment (GPU/USDT pairs) — per the roadmap.
  • Key releases/milestones: TGE 2025‑06‑06 (Investor Docs).
  • Roadmap: https://docs.gpu.net/gvex/roadmap

Ecosystem/integrations

  • Integrations: dapp marketplace (dapp.gpu.net); RPC (rpc.gpu.net) is mentioned in the docs; GitHub code for the chain/contracts.
  • Partners: the Investor Docs list enterprise/partners (not verified by separate primary links).
  • Where it’s actually used: compute rentals + subnets (per the docs; without independent KPIs).

AI component of the project

  • What is real AI here? The main AI focus is providing GPU compute for AI/ML and mentions of AI agents/LLM inference improvements (Investor Docs); own “models” as a product are not verified.
  • What is truly onchain? the ERC‑20 token $GPU (Etherscan) + the described GAN Chain infrastructure (website/docs).
  • Why is a token needed? payment for compute and staking/governance (token launch docs).
  • Red flags (based on the information found): supply inconsistency (Etherscan vs docs/CMC), the absence of public audit reports, some KPIs are marketing claims.

DecAI Fit component summary

  • DecAI Fit: 5/6 — high alignment with the Decentralized AI thesis
Component Level Score
AI reality medium 2
Onchain necessity high 3
Token necessity medium 3
Evidence Tier‑1 present 3
Details

AI reality

AI_0 — AI intent (Tier‑1 roadmap) (1/0)

  • AI_0 (decision): 0
  • Claim (one line): There are signs of AI reality per Tier‑1 (see AI_1..AI_4), so the AI_0 flag (intent‑only) does not apply.
  • Evidence: Tier‑1 — see Evidence/links in AI_1..AI_4.
  • Why sufficient / why not (one line): AI_0=1 is admissible only when AI_1..AI_4=0; here ai_score without AI_0 = 2.
  • Conflict note (if any): none.

AI_1 — what the AI does (1/0)

  • AI_1 (decision): 1
  • Claim (one line): the platform provides GPU compute (including the mention of LLM inference improvements/AI agents) as a service/infrastructure.
  • Evidence: Tier‑1 — https://docs.gpu.net/investor-docs ; https://docs.gpu.net/usdgpu-token-launch
  • Why sufficient / why not (one line): the docs describe AI cases as a workload and as product improvements.
  • Conflict note (if any): specific models/metrics are not independently confirmed.

AI_2 — demos/examples (1/0)

  • AI_2 (decision): 0
  • Claim (one line): no Tier‑1 public demo/benchmarks of “compute rental → AI workload” with reproducible steps were found (while Tier‑1 does contain a link to the dApp marketplace).
  • Evidence: Tier‑1 — https://docs.gpu.net/usdgpu-token-launch (link to the dapp marketplace)
  • Why sufficient / why not (one line): a description without reproducible tests is not considered a demo.
  • Conflict note (if any): none.

AI_3 — metrics/quality (1/0)

  • AI_3 (decision): 0
  • Metrics: “MLPerf benchmarks” and “10% faster LLM inference” are mentioned, but without a full protocol/results in the text.
  • Context (version/date/how measured): —.
  • Evidence: Tier‑1 — https://docs.gpu.net/investor-docs
  • Why sufficient / why not (one line): metrics without details and reproducibility are not counted.
  • Conflict note (if any): none.

AI_4 — tech description/architecture/versions (1/0)

  • AI_4 (decision): 1
  • Claim (one line): node roles, consensus mechanics and the token utility for compute/staking/governance are described.
  • Evidence: Tier‑1 — https://www.gpu.net/gan-chain/ ; https://docs.gpu.net/usdgpu-token-launch
  • Why sufficient / why not (one line): sufficient for a high‑level understanding of the architecture.
  • Conflict note (if any): none.

Onchain necessity

OC_1 — contracts/addresses/chain (1/0)

  • OC_1 (decision): 1
  • Chain: Ethereum (the ERC‑20 token $GPU)
  • Addresses/identifiers: $GPU: 0x79D464248516Bc6977cA2069ba15d8D1044479D8
  • Evidence: Tier‑1 — https://etherscan.io/token/0x79D464248516Bc6977cA2069ba15d8D1044479D8
  • Why sufficient / why not (one line): the explorer confirms the contract and max total supply (there is a conflict with the docs).
  • Conflict note (if any): Etherscan max total supply != 200M from the docs/CMC.

OC_2 — onchain beyond just a token (1/0)

  • OC_2 (decision): 1
  • What useful logic is onchain: a ledger/accounting of nodes and compute proofs are claimed as onchain functions; governance/voting on subnets (per the docs).
  • Evidence: Tier‑1 — https://www.gpu.net/gan-chain/ ; https://docs.gpu.net/usdgpu-token-launch
  • Why sufficient / why not (one line): Tier‑1 sources describe an onchain layer beyond the token.
  • Conflict note (if any): specific addresses/an explorer for GAN Chain are not recorded in this card.

OC_3 — onchain is critical (1/0)

  • OC_3 (decision): 1
  • “Replace the blockchain” test: without onchain/consensus (in the claimed design), the trustless accounting of availability/compute and the distribution of rewards per the network rules break down.
  • Evidence: Tier‑1 — https://www.gpu.net/gan-chain/ ; https://docs.gpu.net/usdgpu-token-launch
  • Why sufficient / why not (one line): the project claims compute verification and governance as network onchain functions.
  • Conflict note (if any): an SSOT for the specific GAN Chain network/explorer is required.

Token necessity

TK_0 — no token (1/0)

  • TK_0 (decision): 0
  • Claim (one line): the project has a token, $GPU.
  • Evidence: Tier‑1 — https://etherscan.io/token/0x79D464248516Bc6977cA2069ba15d8D1044479D8
  • Why sufficient / why not (one line): confirmed by the explorer.
  • Conflict note (if any): none.

TK_1..TK_4 (1/0)

TK_1 — payment for the service (1/0)

  • TK_1 (decision): 1
  • Claim (one line): $GPU is used to rent real GPUs on the marketplace/dApp.
  • Evidence: Tier‑1 — https://docs.gpu.net/usdgpu-token-launch
  • Why sufficient / why not (one line): the “rent GPUs on the dapp marketplace” utility is stated directly.
  • Conflict note (if any): none.

TK_2 — incentives/collateral/penalties (1/0)

  • TK_2 (decision): 1
  • Claim (one line): staking for the Validator/Queen roles, a reward model and slashing risks are mentioned; gas fees burn (deflation).
  • Evidence: Tier‑1 — https://docs.gpu.net/usdgpu-token-launch ; https://www.gpu.net/gan-chain/
  • Why sufficient / why not (one line): the docs directly mention staking roles and slashing risks.
  • Conflict note (if any): slashing/contract parameters are not provided.

TK_3 — governance (1/0)

  • TK_3 (decision): 1
  • Claim (one line): $GPU gives voting power (subnet governance / support for PoW subnets).
  • Evidence: Tier‑1 — https://docs.gpu.net/usdgpu-token-launch
  • Why sufficient / why not (one line): the governance role is described in the token launch docs.
  • Conflict note (if any): no link to a specific onchain governance module.

TK_4 — token is replaceable (1/0)

  • TK_4 (decision): 0
  • Claim (one line): the token is built into compute payment and the staking/governance roles of the network.
  • Evidence: Tier‑1 — https://docs.gpu.net/usdgpu-token-launch
  • Why sufficient / why not (one line): the claimed economy relies on the token.
  • Conflict note (if any): none.

Evidence (table)

Tier Source What it proves
Tier‑1 https://docs.gpu.net/usdgpu-token-launch Fixed supply 200M (claimed), allocations 80/10/10, token utility (rent GPUs, staking/governance, burn gas).
Tier‑1 https://docs.gpu.net/investor-docs TGE 2025‑06‑06, network/product metrics and additional claims (including AI enhancements).
Tier‑1 https://www.gpu.net/gan-chain/ Description of node roles and consensus mechanics (PoA/PoAvailability/PoCompute), staking/slashing scenarios, and the role of $GPU.
Tier‑1 https://etherscan.io/token/0x79D464248516Bc6977cA2069ba15d8D1044479D8 $GPU contract (ERC‑20, proxy) and “Max Total Supply” (conflict with the docs/CMC).
Tier‑2 https://coinmarketcap.com/currencies/gpunet/ Aggregator: reference profile and navigation links (not Tier‑1).
Tier‑1 https://github.com/brahmGAN Official repositories (GAN chain and contracts).
Tier‑1 https://github.com/brahmGAN/ganchain-contracts/security/advisories Published security advisories are absent (as a fact).
Tier‑1 https://gpu.net/ Home page (Astra / AI agents & chat) — context.

No comments to display

Back to top