Janction

About Janction (JCT)

What Is Janction (JCT)?

Janction is a decentralized GPU computing network for AI workloads, developed by JasmyLab within the wider Jasmy ecosystem. The idea is straightforward: pool idle graphics processors from independent operators, normalize them into standardized units, and sell that capacity to people training models, running inference or rendering, at prices set by an auction rather than a fixed rate card. A blockchain layer records who contributed what and distributes rewards accordingly.

JCT is the network's utility and governance token. It is not JasmyCoin (JASMY), a separate token from the same ecosystem, and the ticker JCT is used by unrelated projects elsewhere, so the contract address is the only safe identifier. This page describes two deployments: the ERC-20 on Ethereum at 0xC477B6dfd26EC2460b3b92de18837Fd476Ea7549 and the BEP-20 on BNB Chain at 0xeA37A8DE1de2d9D10772EEB569e28Bfa5Cb17707. Both read on-chain as name "JANCTION", symbol JCT, 18 decimals, and their reported supplies add up to exactly 50,000,000,000 tokens, consistent with a bridge that moves supply between the chains rather than two independent issuances.

Official sources:

• Janction
• Janction documentation
• JCT on Etherscan
• JCT on BscScan
• Janction on GitHub

Who Founded Janction? (History)

Janction is operated by JasmyLab, the development arm associated with the Jasmy ecosystem, a Japanese project best known for its work on decentralized data and device identity. JasmyLab announced Janction's Layer 2 testnet as new Web3-era infrastructure, and in March 2025 launched a decentralized GPU cloud service under the Janction name offering data-center-class NVIDIA hardware at published hourly rates. Jasmy has publicly described Janction as a key partner as it advances its own Layer 2 work built on its decentralized identity technology.

That lineage explains the shape of the project: Janction is closer to an operator-led compute business with a token attached than to a purely permissionless protocol. Janction's own technical documentation does not describe the Jasmy relationship, so the association comes from JasmyLab's announcements rather than from the protocol docs.

How Janction Works

Janction's architecture has three parts:

1. Blockchain layer A Layer 2 chain acts as the settlement and data availability layer for the compute system, recording transactions, storing contribution data and distributing rewards. Janction's documentation describes a testnet built with Optimism's OP Stack, using chain ID 679 and settling to Ethereum's Sepolia test network, with published node versions, a genesis file, a rollup configuration and a bridge application. For data availability on a production network, the documentation puts Ethereum first while evaluating Celestia, Avail and EigenDA on security, maturity, scalability and cost.

2. Distributed resource pooling Heterogeneous GPUs contributed by different operators are divided into virtual GPUs, each carrying equal compute and memory, then containerized as microservices and connected across providers using VXLAN. When a customer requests capacity, scheduling and routing algorithms allocate virtual GPUs from the pool, and containers talk to each other over REST APIs behind a service API.

3. GPU marketplace The marketplace is the commercial front end, covering account and node management, service monitoring and reporting, plus a reputation system that grades the resources an operator supplies and deters bad behavior. Rental is priced through a PVCG auction mechanism rather than a flat rate, intended to find a price that serves both sides of a distributed market instead of defaulting to the provider's ask.

Key Ecosystem Components

The pieces that make up Janction are the Layer 2 chain, documented as an OP Stack rollup handling settlement and data availability; the GPU pool, where contributed hardware is normalized into virtual GPUs; the GPU marketplace with its reputation system and monitoring tools; node software with desktop installers for Windows and macOS and a peer-to-peer join flow; the JCT token contracts on Ethereum and BNB Chain; Proof of Contribution, the reward attribution mechanism; and a browser extension for consented data contribution that feeds the same reward system.

JCT Token Utility

Janction documents four roles for JCT:

• Utility, as the unit of account for network activity
• Marketplace settlement, so AI workloads are paid for in the token
• Staking, which providers use to supply GPU capacity and to take part in bidding
• Governance, letting holders vote on protocol updates and ecosystem decisions

The total supply is 50,000,000,000 JCT, with the intended deployment split documented as 90% on Ethereum and 10% on BNB Chain, though the live split moves as tokens bridge between the two. Allocations are documented as 34.29% ecosystem, 21.34% team, 18% foundation, 10% investors, 5.7% airdrop, 4% liquidity, 3.67% advisors and 3% community incentives, with roughly 23% unlocked at the token generation event. Liquidity, community incentives and airdrop released in full at that event, while team tokens carry an eighteen-month cliff followed by three-year vesting.

What Makes Janction Unique?

The more distinctive elements are:

• Standardization into virtual GPUs of equal compute and memory, so a heterogeneous fleet can be sold as a uniform product
• Auction-based pricing through PVCG rather than provider-set rates
• Contribution accounting that separates data providers from compute providers and scores them with different mathematics
• An operator-backed GPU cloud alongside the decentralized network, giving the marketplace supply that does not depend solely on volunteers

How Janction Differs from Similar Networks

Compared with other decentralized compute networks:

• Several competitors run coordination on an existing chain; Janction documents its own OP Stack rollup for settlement and data availability
• Where many networks price capacity by simple bid and ask, Janction specifies an auction mechanism meant to balance both sides
• Reward attribution is unusually formal on paper, using Shapley values for compute contributors and a reputation-weighted individual method for data contributors
• Janction is backed by a named corporate operator rather than being purely community-run
• Its network is at a documented testnet stage while several competitors already run production networks, the sharpest practical difference

Use Cases of Janction

Intended use cases include:

• Renting GPU capacity for AI model training and fine-tuning
• Running inference workloads for AI products
• Rendering and other compute-heavy graphics work
• Monetizing idle hardware by contributing it to the pool
• Contributing consented browsing and interaction data through the extension for rewards
• Staking JCT to supply capacity, bid for work or participate in governance

How Janction Is Secured

Security in this system means several different things:

• Chain security comes from the OP Stack rollup design and its settlement layer, which in the documented testnet is Ethereum's Sepolia network
• Token security today rests on Ethereum and BNB Chain, where the JCT contracts are deployed
• Marketplace integrity relies on the reputation system, which grades operators and deters malicious supply
• Reward integrity relies on Proof of Contribution, which attributes value rather than ordering blocks
• Economic security relies on provider staking, which puts value at risk behind a promise of capacity

Risks and Considerations

Points worth weighing:

• The network described in Janction's documentation is a testnet, while the token already trades on two production chains, so token utility runs ahead of the chain
• Endpoints listed in the documentation, including the testnet RPC, the testnet explorer, the bridge and the project's own block explorer, have not been consistently reachable, so verify current endpoints through official channels before relying on them
• Proof of Contribution is presented as a theoretical design to be implemented as circumstances allow, so treat it as a specification rather than shipped behavior
• Verifying that off-chain GPU work was genuinely performed is a hard, unsolved problem the documentation does not resolve
• Insider-facing allocations, counting team, foundation, investors and advisors, exceed half the supply, so the vesting schedule matters a great deal
• The project depends heavily on one corporate sponsor, a strength for delivery and a concentration risk for governance
• Established compute networks already have supply, demand and integrations, and competing with centralized cloud on price alone is difficult

Why Janction Matters

Demand for AI compute has outrun the supply centralized providers can allocate at reasonable prices, and pooling underused hardware is a sensible response. Janction is one of the more concrete attempts to pair that idea with a real operating business rather than a purely speculative network. What separates a project like this from success is unglamorous: verifiable work, reliable scheduling, honest pricing and endpoints that stay up.

Where Can You Buy Janction (JCT)?

JCT trades on centralized exchanges and on decentralized markets on Ethereum and BNB Chain. Availability and trading pairs vary by jurisdiction and by venue.

Major Centralized Exchanges (CEX)

JCT is listed on venues including Gate, MEXC, BitMart and WEEX, usually against USDT.

Storage Options After Purchase

After purchase, JCT can be held on an exchange account, which is custodial; in an EVM-compatible self-custody wallet on Ethereum or BNB Chain; or in a hardware wallet paired with compatible wallet software. The two deployments use different addresses, so confirm which network you are on and add the token by contract address rather than by ticker search.

Related Coins

You can compare JCT with related ecosystems on Coinrithm:

• Render (RENDER)
• Akash Network (AKT)
• io.net (IO)
• Aethir (ATH)
• JasmyCoin (JASMY)