Chutes (SN64)

Overview

Chutes is a platform for running artificial intelligence (AI) models on a network of computers that provide processing power. It operates as Subnet 64 of Bittensor, a blockchain network with separate subnets for different kinds of digital work. The subnet’s token is commonly called SN64. Chutes serves developers who want to use an existing AI model or deploy their own application without managing a fleet of servers. (docs.chutes.ai)

The platform’s main focus is inference: the work an AI model does when it responds to a prompt, creates an image, or processes another request. Chutes connects those requests to computing capacity supplied by network participants. Its developer tools handle tasks such as deployment and access through an application programming interface, or API. SN64 belongs to the Bittensor system that rewards and coordinates work on the Chutes subnet; it is distinct from the service’s usage charges. (chutes.ai)

History & Team

Chutes announced its launch in December 2024 as a way to make AI computing capacity available through Bittensor. Its early work centered on giving developers a simpler path to deploy models and run AI workflows. Over time, the project expanded its developer documentation, deployment tools, and miner dashboards. (chutes.ai)

Chutes describes itself as a project developed by Chutes Global Corp and a distributed team rather than one led by a single chief executive. Its published team page names contributors across backend development, product work, sales, and community support. That mix reflects the two sides of the project: building the software that serves AI requests and working with the people who supply or use the service. (chutes.ai)

In 2026, the team described a shift toward serving workloads that better support the cost of running its infrastructure. It has also published work on confidential AI computing and experiments in training models across distributed computers. Those efforts sit alongside its core inference service. (chutes.ai)

Technology & How It Works

From an AI request to a result

A developer can call a hosted model through an API or package their own code as a deployable application called a chute. A chute defines the code to run, the computing power it needs, and the ways users can send it requests. Chutes then handles deployment on available infrastructure. Developers can create an endpoint that responds to an API call or a background job that continues working after the initial request. (chutes.ai)

The computing power comes from miners, participants who connect hardware to Subnet 64. Validators help coordinate the subnet and check participating miners. Chutes’ miner documentation describes registration on Bittensor, GPU setup, and checks that confirm a miner’s advertised hardware. This arrangement lets developers use a service built on machines supplied by multiple participants. (chutes.ai)

Checking the computing environment

Running someone else’s code on a distributed machine calls for checks on the machine and its software. Chutes documents measures that include encrypted communication, checks on application files, GPU verification, and controls on what a running program can access. It also supports trusted execution environments for certain workloads. These use hardware-based checks, known as attestation, to help confirm the environment in which code is running. (chutes.ai)

For developers, much of this work stays behind the API and deployment tools. They can focus on a model and its output while the platform handles the steps needed to run it across the network. (chutes.ai)

Tokenomics & Utility

SN64 is the alpha token associated with Bittensor Subnet 64. Under Bittensor’s Dynamic TAO design, each subnet has its own alpha token alongside TAO, the network-wide token. The system uses a pool containing TAO and the subnet’s alpha token to support exchanges between them. Users can also stake into a subnet, linking their participation to that subnet’s economy. (docs.learnbittensor.org)

Bittensor issues subnet tokens through its network reward system. Part of a subnet’s issuance supports rewards for miners, validators, and the subnet’s creators, while another part enters its token pool. For Chutes, this gives SN64 a role in the incentives around supplying and evaluating computing work. An exchange’s Chutes listing identifies 21 million SN64 as its total supply. (docs.learnbittensor.org)

The distinction between network rewards and customer payments matters. Chutes’ terms describe subscriptions and pay-as-you-go charges based on usage, with payment methods that include card or bank payments through Stripe and blockchain deposits in TAO. A developer can therefore pay to use Chutes without SN64 serving as the direct payment token for every AI request. (chutes.ai)

Ecosystem & Use Cases

Chutes has two closely related groups of participants. Developers and organizations bring AI workloads; miners bring the hardware used to run them. Validators support the Bittensor side of the system by evaluating subnet activity. This gives Subnet 64 a practical task—delivering computing capacity—alongside its token incentives. (chutes.ai)

A developer can use hosted models for tasks such as chat and text generation, or build a custom chute with its own code and requirements. The SDK supports API endpoints for requests that need a prompt response. It also supports background jobs for longer tasks, including file processing and other work that needs progress tracking. Streaming tools can send parts of a result as they become ready, which is useful in interactive applications. (chutes.ai)

Chutes also provides account and authentication tools for applications built on top of its service. API keys let developers connect server-side software, while its sign-in system can let an application make requests on behalf of a Chutes user. These tools make the compute network easier to include in a product that people use every day. (chutes.ai)

Advantages & Challenges

A key advantage of Chutes is that it combines a familiar developer experience with distributed hardware. The SDK, deployment commands, and APIs reduce the amount of server setup needed to run an AI application. Support for both quick responses and long background jobs gives developers options for different kinds of work. The project also publishes much of its core software and describes how it checks the machines serving requests. (chutes.ai)

Its challenges come from the work required to operate AI infrastructure. Large models need capable GPUs, and running a reliable service involves deployment, validation, and ongoing coordination between miners and validators. Chutes’ miner guide describes substantial setup for hardware providers. The team’s 2026 business update also shows that processing more requests is only one part of operating the platform; the cost of serving those requests matters too. (chutes.ai)

Where to Buy & Wallets

SN64 is available on MEXC, which has announced an SN64/USDT spot market. Bitget offers Chutes (SN64) through its purchase and spot-trading pages. Exchange access depends on the services offered in a customer’s location. On an exchange, purchased tokens appear in the customer’s exchange account until transferred elsewhere. (mexc.com)

For participation on Bittensor, a wallet must support the network and its subnet tokens. SubWallet supports Bittensor alpha-token staking, including the process of choosing a subnet. Chutes’ own developer setup also uses Bittensor wallets for authentication, although a developer account and a wallet holding staked subnet tokens serve different purposes. (docs.subwallet.app)

Regulatory & Compliance

Rules for crypto assets depend on the jurisdiction and on how a token or related service is offered. In the United States, the Securities and Exchange Commission explains that an offer or sale may fall under securities law when the facts meet the test for an investment contract. That analysis considers the transaction and the expectations created around it, rather than applying one label to every subnet token. (sec.gov)

In the European Union, the Markets in Crypto-Assets regulation, known as MiCA, sets rules for many crypto-asset services. Firms providing covered services in the EU generally need the applicable authorization or must qualify under another permitted route. These rules concern activities such as custody and operating a trading platform; they are separate from the technical role of an AI compute subnet. (esma.europa.eu)

Shariah compliance calls for a separate assessment of how an asset is acquired, used, and rewarded. In SN64’s case, that assessment would need to consider Bittensor’s staking and token-reward mechanics as well as Chutes’ computing service. The project’s description of its AI utility does not, by itself, establish a halal classification. (docs.learnbittensor.org)

Future Outlook

Chutes’ direction is tied to a practical question: how well can a distributed network serve AI workloads that developers want to run? Its work on hardware checks and confidential computing addresses the demands of operating models on machines supplied by different participants. Its support for custom deployments and longer jobs broadens the tasks developers can bring to the platform. (chutes.ai)

The team has also reported research on training a model across distributed GPUs. That research points to an area beyond its main inference service, while its 2026 business update places attention on the efficiency of the service it already operates. For SN64, continued activity in the subnet depends on the relationship between useful computing work and Bittensor’s incentives for the people who support it. (chutes.ai)

Summary

Chutes connects developers to AI computing capacity through Bittensor Subnet 64. Its tools make it possible to call hosted models, deploy custom applications, and run longer jobs, while miners supply hardware and validators support the subnet. SN64 is the subnet’s alpha token within Bittensor’s staking and reward system. Together, the service and token show how a crypto network can organize participants around a concrete computing task. (chutes.ai)