表现
RD 熱度指數
RD 增長指數
| 1h | 24h | 7d | 30d | 60d | 90d |
|---|---|---|---|---|---|
| 0.42% | 4.15% | 3.94% | 26.77% | 24.47% | -- |
收入和銷毀
協議收入
$73.61K回購資金
--銷毀資金
--回購率
--協議資金趨勢
月度金額與比值
| 週期 | 協議收入 | 回購資金 | 銷毀資金 | 回購率 |
|---|---|---|---|---|
| 2026-09 | $73.61K | -- | -- | -- |
| 2026-08 | $388.11K | -- | -- | -- |
| 2026-07 | $380.52K | -- | -- | -- |
| 2026-06 | $375.75K | -- | -- | -- |
| 2026-05 | $494.54K | -- | -- | -- |
| 2026-04 | $495.24K | -- | -- | -- |
比較
代幣經濟學
解鎖進度
數據源: CoinGecko
Chutes (SN64) is the Bittensor Dynamic TAO subnet alpha token, used for the AI inference subnet's staking and liquidity. Total supply is 21M, adjusted via mint/burn.
跟隨更新
跟隨列表
人物
關於Chutes
Chutes 是一個基於 Bittensor 子網 64 建構的去中心化 AI 推理平台，串聯全球 GPU 礦工與需要執行開源 AI 模型的使用者，累計處理超過 34 兆個 token，每日峰值突破 1200 億個 token。 核心價值主張：透過匯集個人愛好者與資料中心的閒置算力，Chutes 提供每百萬 token 僅需個位數美分的推理服務——僅為集中式供應商成本的零頭。安全性方面採用可信執行環境（TEE）與 GraVal 硬體認證機制，確保資料隱私，即使 GPU 擁有者也無法窺探。該平台透過 API 或聊天介面提供 DeepSeek、Llama、Qwen 與 Mistral 等模型。 市場痛點：集中式 AI 供應商對本質上屬於大宗商品的算力收取溢價。Chutes 透過建立競爭性市場解決此問題，礦工根據速度、可靠性與成本效益進行評分，在維持品質的同時壓低價格。 近期發展（2026 年 3 月至 9 月）：網路擴展至超過 1,170 個活躍 GPU 節點，包含 H200 與 B200 顯示卡。2026 年 8 月，Chutes 強制執行 TEE 驗證，淘汰所有非機密 GPU 路徑。在此之前，團隊進行了策略性縮減——將 GPU 機群從 1,700 張 H200 減至 700 張，取消免費方案，並移除無利潤的模型。Token 交易量下降 45%，但按用量付費的自主使用量成長 20%，每張 GPU 日收益躍升 44.9% 至 5.89 美元。90 天內營收約達 150 萬美元。 已知限制：去中心化網路面臨可靠性風險——礦工在推理中途離線是與雲端故障不同的失效模式。該平台主要服務開源模型；OpenAI 或 Anthropic 的專有前沿模型並未提供。
Listing and progress: SN64 went live for trading on Kraken on July 7. On July 8 Chutes announced it had completed non-blocking decentralized training of a recurrent model, landing within 0.6% of centralized training quality; its Parallax framework was first outlined around June 10.
Revenue and token mechanics: SN64 is the alpha token of Bittensor subnet 64, with no revenue share, dividend or equity, so its value tracks Bittensor emissions and TAO. The platform offers 14 TEE-attested models priced per million tokens and self-reports trillions of tokens processed monthly.
Chutes achieved two key milestones in 2026. On April 30, 2026, Nasdaq-listed AIxCrypto (AIXC) signed a non-binding collaboration framework with Chutes AI to explore using Chutes' decentralized AI inference infrastructure for AIxC's real-time, multi-agent applications (covering inference latency, concurrency management and system responsiveness)
Future strategy and roadmap: On July 8, as the decentralized compute provider on Bittensor subnet 64, Chutes announced non-blocking decentralized training of recurrent AI models with only a 0.6% quality gap versus centralized training, calling it a world first; its Parallax framework lets geographically distributed, heterogeneous GPUs train sparse mixture-of-experts models without synchronization stalls. Chutes continues to deliver SOTA open-source models (e.g. Kimi K2.6, GLM 5.2, Gemma 4) on a server-grade decentralized compute layer, often within minutes of release.
Jon Durbin — Founder; launched Chutes in 2024 and has long worked in the Bittensor ecosystem and open-source inference. The organization has no CEO and is run by core contributors.
Core contributors include Cxmplex and Florian S on backend, Bonoliver on frontend, and Veight on marketing and sales.