Chutes (SN64) on Bittensor — Decentralized AI Inference With Hardware-Private Compute

Executive Summary

Chutes (Subnet 64 on the Bittensor network) is a decentralized, serverless AI compute platform that differentiates itself through hardware-enforced privacy (TEE) and post-quantum end-to-end encryption (E2EE). As of April 6, 2026, the project demonstrates strong on-chain traction and developer velocity, processing significant inference volume. However, it carries notable risks typical of decentralized collectives: a pseudonymous core team, a lack of published tokenomics (distribution/unlock schedules), and the absence of a verified third-party security audit for its specific subnet architecture.

1. General Description

Project Overview and Problem Solved

Chutes is a decentralized compute provider designed for deploying, scaling, and running open-source AI models in production [1]. It solves three primary problems in the AI infrastructure space:

1. Cost and Scale: By leveraging a decentralized network of GPU providers via the Bittensor network, it offers AI compute at significantly lower costs than traditional centralized cloud providers like AWS [2].
2. Privacy and Security: It addresses data privacy concerns by offering hardware-enforced Trusted Execution Environments (TEEs) and post-quantum End-to-End Encryption (E2EE), ensuring that neither the GPU provider nor Chutes can read user prompts or responses [3] [4].
3. Uptime Resilience: Decentralized routing ensures that if one miner or data center goes offline, traffic is automatically routed to another without API downtime [5].

Target Audience

The platform targets AI developers, startups, and enterprise teams building AI-powered applications who require scalable, cost-effective, and highly private inference for open-source models [6].

2. Team

Corporate Structure and Anonymity

Chutes operates under "Chutes Global Corp," an International Business Corporation registered in Nevis (Registration C 61974) [7]. The project is built by a "global, decentralized collective" of engineers and researchers, and explicitly states that it has no CEO or central decision-maker [8].

Professional Presence

• LinkedIn: A company page exists showing 2-10 employees, with a few named individuals (e.g., Timon Agar, Head of Sales; Chris Kouris, Head of Ecosystem) [6].
• Socials: The team is highly active on X (Twitter) under the handle @chutes_ai, frequently posting technical updates, responding to community queries, and verifying integrations [5].
• Pros/Cons: While the active development and communication are positive, the pseudonymous nature of the core technical contributors and the offshore corporate structure introduce governance and accountability risks.

3. Concept/Documentation

Uniqueness and Technical Details

Chutes operates as Subnet 64 on the Bittensor network. Its primary differentiator is its privacy stack: it utilizes Intel TDX and NVIDIA Protected PCIe for hardware-level encryption, combined with ML-KEM-768 (Kyber) for post-quantum E2EE [3] [4].

• Consensus: Relies on Bittensor's consensus. Miners provide GPU resources, and validators rank them based on compute units (55%), invocations (25%), unique chutes (15%), and bounties (5%) [7].
• Validation: Uses custom "cord" functions for adversarial validation, cross-verifying outputs across multiple miners to prevent cheating [2].

Demand, Revenue, and Partnerships

• Demand: The project claims to process "trillions of tokens per month" and serves millions of AI inference requests daily [1] [2].
• Revenue Model: Chutes monetizes via a Pay-As-You-Go (PAYG) model and monthly fiat subscriptions ($3 Base, $10 Plus, $20 Pro) processed via Stripe, alongside native TAO token payments [9] [3].
• Partnerships: Chutes is a verified provider on OpenRouter [10]. It also claims to power inference for GenLayer, though this relies on self-reported social media posts [5].

4. Coin/Tokenomics

Supply and Distribution Metrics

The native token for the subnet is SN64. Data as of April 6, 2026, shows the following metrics:

Metric	Value	Source
Current Price	~$27.91	[11]
Circulating Supply	4,742,184 SN64	[11]
Max Supply	21,000,000 SN64	[11]
Fully Diluted Valuation (FDV)	~$132.4M	[11]
24h Trading Volume	~$1.19M	[11]
Total Holders	~13,644	[12]

Note: CoinMarketCap data for this token shows conflicting/zero supply metrics [13]; CoinGecko and GeckoTerminal provide more reliable on-chain data for Bittensor subnets.

Tokenomics Risks

• Missing Documentation: There is no official whitepaper or tokenomics documentation on the Chutes website detailing team allocations, investor distributions, or specific unlock schedules.
• Decentralization: The largest single wallet holds approximately 41,222 SN64 (less than 1% of the circulating supply), indicating a relatively decentralized holder base among the top wallets [12].
• Liquidity: Trading is heavily concentrated on decentralized exchanges (Subnet Tokens DEX), with centralized exchange (MEXC) volume accounting for less than 0.2% of activity [11].

5. Code

Open Source and Development Activity

Chutes maintains a highly active, open-source presence on GitHub under the chutesai organization.

• Repositories: Core repositories like chutes (CLI/SDK), chutes-miner, and chutes-audit are public and licensed under MIT [14] [15] [16].
• Activity: The repositories show frequent and recent commits (as recent as April 2026), indicating active, ongoing development [14] [15].

Security Audits

• Internal Audits: The team provides a chutes-audit repository containing scripts to prove the fairness of request distribution between validators and miners [16].
• External Audits (WARNING): There is no verified third-party smart contract or security audit for the Chutes (SN64) Bittensor subnet. Note: A search for "N64 audit" returns a Cyberscope audit for a Solana-based token named N64 [17]. This is a completely different project and should not be conflated with Chutes.

6. Risks

• Financial Risks (High): The complete lack of published tokenomics (allocations, vesting schedules) makes it impossible to assess dilution risks. Liquidity is highly dependent on a single DEX pool [11] [12].
• Technical Risks (Medium): While the codebase is open and active, the absence of a formal third-party security audit for the subnet architecture introduces smart contract and infrastructure risks.
• Team Risks (Medium/High): The project is run by a pseudonymous collective registered in an offshore jurisdiction (Nevis) [8] [7]. Recourse in the event of a failure or exploit is highly limited.
• Market Risks (Medium): The project relies heavily on the broader Bittensor (TAO) ecosystem's emission schedule and miner stability [2].

7. Community

• Twitter/X: The @chutes_ai account has ~9,639 followers and is highly active, posting detailed technical threads and engaging with users [5].
• Discord: An active Discord community exists for developer support and discussions [18].
• On-Chain Users: The token has over 13,600 on-chain holders, indicating a robust grassroots community [12].

8. Final Assessment

Risk Level: MEDIUM-HIGH (for the token) / LOW-MEDIUM (for the product)

Key Strengths

• Exceptional Privacy Tech: The implementation of hardware TEEs and post-quantum E2EE is a massive differentiator for enterprise AI adoption [3] [4].
• Real Product-Market Fit: The platform is live, generating revenue, and processing billions of tokens with verified integrations like OpenRouter [2] [10].
• Active Development: Transparent, open-source codebase with high developer velocity [14].

Key Issues and Warnings

• Opaque Tokenomics: The absolute lack of documentation regarding token distribution, team allocations, and unlock schedules is a severe red flag for investors.
• No Third-Party Audit: Relying solely on internal audit scripts for a platform handling significant compute value is risky.
• Pseudonymous Governance: The "no CEO" collective model under an offshore entity limits accountability [8] [7].

Conclusion: Chutes is a highly capable technical product delivering real value in the decentralized AI space. However, from a Web3 investment perspective, the $SN64 token carries significant risk due to the opacity of its tokenomics and the lack of independent security audits. Users should feel comfortable using the API for private inference, but investors should exercise strict risk management regarding the token.

References

1. Chutes | Serverless AI Compute for Open-Source Models. https://chutes.ai/
2. Chutes. https://subnetalpha.ai/subnet/chutes/
3. Privacy Policy | Chutes. https://chutes.ai/privacy
4. End-to-End Encrypted AI Inference with Post-Quantum Cryptography | Chutes Blog. https://chutes.ai/news/end-to-end-encrypted-ai-inference-with-post-quantum-cryptography
5. Chutes (@chutes_ai) / Posts / X. https://x.com/chutes_ai?lang=en
6. Chutes AI. https://www.linkedin.com/company/chutesai
7. SN64 - Chutes | TAO.app | Your Gateway to Bittensor. https://www.tao.app/subnet/64
8. Who are the founders of Chutes AI?. https://chutes.ai/news/who-are-the-founders-of-chutes-ai
9. Pricing | Chutes. https://chutes.ai/pricing
10. Chutes | OpenRouter. https://openrouter.ai/provider/chutes
11. Chutes Price: SN64/USD Live Price Chart, Market Cap .... https://www.coingecko.com/en/coins/chutes
12. SN64/TAO - Chutes Price on Subnet Tokens | GeckoTerminal. https://www.geckoterminal.com/bittensor/pools/0-64
13. Chutes price today, SN64 to USD live .... https://coinmarketcap.com/currencies/chutes/
14. chutesai/chutes. https://github.com/chutesai/chutes
15. chutesai/chutes-miner. https://github.com/rayonlabs/chutes-miner
16. chutesai/chutes-audit: Auditing scripts with optional weight .... https://github.com/rayonlabs/chutes-audit
17. N64 Smart Contract Audit. https://www.cyberscope.io/audits/1-n64
18. Discord. https://discord.gg/chutes