📰 BloombergBloombergDoubao · 实盘不做把握 7/10不做原因:风险收益比不佳
北京时间·🕒 发布 18:32:26·⟳ 抓取 19:02:07 +29m41s·⚡ 决策 10-21 19:04:10 +2m03s·6.0 天前
Silicon Data携手CME计划推出基于AI token支出指数的可交易衍生品，尚待CFTC批准
消息主体Silicon Data未上市；唯一上市相关方CME为超大市值交易所运营商，本次只是既有合作下新增一个尚待CFTC批准、数月后才上线的早期衍生品品类，无任何收入或交易量指引，财务量级不足以驱动股价；盘前CME价格零变动（session -0.002%），市场本身也确认无edge。该产品不改变AI/GPU产业链实际供需，没有可坐实的二阶映射标的。
📄 原文Bloomberg↗ 原文
Silicon Data Inc., a startup known for its artificial intelligence indexes, said it is planning to launch products base…
不做 CME Group Inc.（CME.US）
待CFTC批准、数月后上线的早期衍生品品类扩展，无收入量级，相对CME超大市值materiality不足；合作框架5月已公开，盘前价格零反应，无预期差；直接主体Silicon Data未上市，亦无可坐实的产业链二阶标的
展开详情(成交明细 / 研究档案 / 原文全文 / 决策记录链)
研究档案
结论摘要
私营初创Silicon Data与CME计划推出AI token支出指数衍生品，但需CFTC批准、数月后才发布，且是5月以来合作的延续；对千亿美元市值的CME无可量化增量，对AI产业链无实际供需影响，无值得交易的标的。
消息传播链
- Bloomberg首发并配有CEO Carmen Li电视采访，同日公司发布新闻稿，宣布扩展与CME的token指数产品合作
关键理由
- 无做多edge：CME盘前报价295.30美元、几乎零变动，市场已以脚投票认为该早期、待批准产品不具materiality
主要风险
- 若CFTC快速批准且产品获得超预期交易量，CME可能获得主题性关注，但当前无任何数据支持该预期，不宜提前博弈
反向证据
- 做多前查反向 · 来源 web · 日期 2026-10-21 · 影响 weakens
无收入/交易量指引，产品待批准且数月后才上线；CME盘前零反应，未发现支持短期重定价的证据
检索:CME token index futures revenue materiality CFTC approval 原文 ↗
先例如何影响了这次决策
对本次决策的影响:先例缺失叠加materiality不足，支持不做而非凭主题想象开仓
方向:未找到与本次事件同机制（交易所新增待批准的AI成本基准衍生品）且可干净定锚的上市标的价格先例
涨幅量级:无高相似先例可校准幅度
形态:无足够样本判断见光死模式
工具执行结果 · 成功 3 · 降级 0 · 失败 0 · 取消 0
- search_processed_events 成功 stock_tools HTTP · 3,453 ms
{"result": "search_processed_events:agent query=`cme group ai token expenditure index futures cftc | silicon data gpu rental futures august partnership | silicon data llm token index tradable products cme`；结构化索引覆盖完整 72 小时；旧记录兼容扫描近 72.0/72 小时、22083 条本臂决策(完整)，返回 0 条候选。未命中只表示这些 agent query 表达式没有召回候选，不能 - web_search 成功 doubao_search · 1,325 ms
豆包搜索结果（原始查询:Silicon Data Inc CME LLM token index products announcement；API SearchType:web） - stock_quote 成功 stock_tools HTTP · 762 ms
{
原文全文
展开原文 · 1,779 字
Silicon Data Inc., a startup known for its artificial intelligence indexes, said it is planning to launch products based on AI token expenditures as part of an ongoing partnership with CME Group Inc.
The new tradable products will be based on the company’s LLM token indexes, which track token prices across various AI models weighted by usage, Silicon Data said in a press release on Wednesday. Tokens are units of data used by models like ChatGPT and Claude, often tracked by traders as a gauge of what customers are paying to use the technology.
The expansion follows plans announced by the two firms in August to launch futures based on the cost of renting GPUs.
Silicon Data is among several companies betting on a new category of derivatives that will enable businesses to hedge and track the cost of using AI. A number of market makers have assembled teams to trade these novel contracts — all while the nascent industry awaits approval by the Commodity Futures Trading Commission.
Read More: AI Trade Confronts Complex World of Cheaper Tokens, Rising Costs
“People have really good understanding of how the input of the GPU compute industry works,” Carmen Li, chief executive officer of Silicon Data, said in an interview on Bloomberg Television on Wednesday. “What’s interesting is, as a society, we lack understanding of the other component of AI economy, which is the consumption layer.”
The products are planned for release in the coming months, Li said, pending regulatory approval.
After sliding over summer, the Silicon Data LLM Token Expenditure Index has hovered near the lowest since it started publishing in December. Since it measures the price per million tokens, that could reflect users shifting from premium models to cheaper alternatives, Li said.
The new tradable products will be based on the company’s LLM token indexes, which track token prices across various AI models weighted by usage, Silicon Data said in a press release on Wednesday. Tokens are units of data used by models like ChatGPT and Claude, often tracked by traders as a gauge of what customers are paying to use the technology.
The expansion follows plans announced by the two firms in August to launch futures based on the cost of renting GPUs.
Silicon Data is among several companies betting on a new category of derivatives that will enable businesses to hedge and track the cost of using AI. A number of market makers have assembled teams to trade these novel contracts — all while the nascent industry awaits approval by the Commodity Futures Trading Commission.
Read More: AI Trade Confronts Complex World of Cheaper Tokens, Rising Costs
“People have really good understanding of how the input of the GPU compute industry works,” Carmen Li, chief executive officer of Silicon Data, said in an interview on Bloomberg Television on Wednesday. “What’s interesting is, as a society, we lack understanding of the other component of AI economy, which is the consumption layer.”
The products are planned for release in the coming months, Li said, pending regulatory approval.
After sliding over summer, the Silicon Data LLM Token Expenditure Index has hovered near the lowest since it started publishing in December. Since it measures the price per million tokens, that could reflect users shifting from premium models to cheaper alternatives, Li said.
决策记录链 · 只追加
每次领取产生一条记录,重新决策追加一条并以 supersedes 指向被替代的记录(ADR 0014)。
73ef813223· 2026-10-21 19:04 · doubao / doubao-seed-evolving-latest-version · 策略包658c0c1e2ba5当前