现在业界给他的估利润 有点不同的见解。 这个公司确实强，共识度高。 现在它分了10几亿收入给coinbase，之前按照收入 计算给的。最新的，业界认为这10几亿收入其实是利 润。因为不分这10+亿，也不会增加其它成本。 现在讨论的是什么时候可以不再分一半了？自然是等法案确定后，它的份额增大。议价能力起来。 ...

全球ai由训练走向推理了。软件应用开始冒头。 2024年发布的模型总数同比均有所下降。美国为2024年发布知名模型最多的地区，数量达40个，较2023年的61个同比下降34.43%。分机构看，2024年贡 献知名模型最多的机构分别是OpenAI（7个）、谷歌（7个）和阿里巴巴（4个）。受MoE等新技术推动，2024年模型的参数数量保持快速上升趋势，规模 扩大仍是模型性能提升的重要方式。 随着强化学习时间和推理思考时间的增长，模型性能也将得到显著提升。据前OpenAI应用研究负责人Lilian Weng数据，s1实验中，通过强制延长思维链 推理路径长度，以Token衡量的平均思维时间与下游评估准确率之间展现出明显的正相关关系。据上海交通大学研究表明，通过延长AI的推理时间，仅需 500个样本训练，就能让模型在医疗诊断准确率上提升6%-11%，达到专业医生的诊断水准 。 趋势看，tokens调用激增。 2023-2024年，开发人员采用AI工具率由44%提升至63%。Google每月处理Tokens增长50倍，Microsoft Azure AI Foundry处理Tokens增长5倍。AI模型训练成 本高+的推 ...

Core Insights - The article highlights a significant shift in the AI industry from training to inference applications, indicating a growing demand for inference chips and related technologies [5][6]. Group 1: Chip Market Dynamics - The rental prices for leading inference chips have recently increased, indicating a rising demand in the market [3]. - In contrast, the rental prices for the main training chip, H100, have seen a decline, suggesting a shift in focus towards inference capabilities [4]. - The demand for AI model invocation has surged, with popular models like Claude 3.5/3.7 and Gemini 2.5 leading the charge [4]. Group 2: Usage Metrics - Microsoft Cloud reported a fivefold increase in token invocation from January to April, surpassing 500 trillion tokens in a single month [5]. - Alibaba Cloud also experienced a significant increase, with daily token invocation exceeding 30 trillion, marking a fourfold growth [5]. Group 3: Market Performance - The article notes that Broadcom, recognized as the king of ASIC chips, has outperformed Nvidia, the leader in training chips, reaching new highs [6]. - The AI application sector in the US stock market has also reached new highs across various industries, including military and education [6].

感受一下 估计有资金当空气币干了 ...

挺热闹的。 还在现场竞价，区间很宽。估计还要1-2个小时。国内心理价开在100亿比较能接受。目测比较尴尬。 感觉全球都在盯呀，不仅 股票圈，币圈也在看。 这些都是币圈媒体。 ...

今天一些算力老面孔 配合稳定币新面孔发力。 已经休息了3个月了。 熟悉的记忆卷土重来。 ...

一、 用户、用量与资本支出三线齐飞 到 2025 年 4 月， ChatGPT 每周活跃用户达到 8 亿 ， 最近 17 个月用户增长 8 倍。 ChatGPT 的用户增长速度显著快于以往的基础技术。 ChatGPT 仅用了 5 天就达到了 100 万用户，而 iPhone （ 2007 年）则用了 74 天， TiVo （ 1999 年） 用了约 1680 天，福特 T 型车（ 1908 年）用了约 2500 天。 AI 比其他基础技术产品 推进 更快 、价格 更便宜 。 。 2023 年 7 月至 2025 年 4 月， 美国活跃用户每天在 ChatGPT 应用程序上花费的分钟数增长 202% ，会话次数增长 106% ，同期平均会话时长增长了 47% 2014-2024 年， 美国六大科技公司（苹果、 NVIDIA 、微软、 Alphabet 、亚马逊和 Meta ）的资本支出 每年增长 21% ，随着数据量的增加，需要建立更 多的超大规模数据中心、更快的网络基础设施和更多的计算能力的资本支出，全球超大规模云收入增长 37% 。资本支出目前占其收入的 15% ，高于十年 前的 8% ，年增长率为 21% ...

Core Viewpoint - The article discusses the recent developments in stablecoin legislation, particularly the GENIUS Act in the U.S., which aims to regulate stablecoins and solidify the dominance of the U.S. dollar in the global stablecoin market [1][4]. Group 1: Definition and Types of Stablecoins - Stablecoins are defined as cryptocurrencies pegged to traditional fiat currencies or physical assets, designed to maintain value stability amidst the volatility of the crypto market [1]. - There are three main types of stablecoins: fiat-backed, commodity-backed, and algorithmic stablecoins. Fiat-backed stablecoins, like USDT and USDC, are the most prevalent due to their transparency and credit backing [1][2][3]. Group 2: GENIUS Act Key Points - The GENIUS Act mandates that all stablecoins must be backed 1:1 by high-quality, low-risk liquid assets, reinforcing the role of stablecoin issuers as quasi-money market funds [4]. - The Act aims to consolidate the dominance of the U.S. dollar, as over 99% of stablecoin market value is tied to dollar-pegged stablecoins [4]. - The Act restricts technology companies from issuing stablecoins unless they meet strict financial risk and consumer data privacy requirements, reflecting concerns over monopolistic practices [5]. Group 3: Market Dynamics and Regulatory Landscape - As of March 2025, the global stablecoin market exceeded $210 billion, with Tether (USDT) and USDC holding significant amounts of U.S. Treasury securities [4]. - The article highlights the emergence of a multi-centered regulatory landscape, with Hong Kong passing its own stablecoin regulations, contrasting with the U.S. approach [5][6]. - The GENIUS Act prohibits interest payments on regulated stablecoins, potentially redirecting capital to tokenized money market funds [8]. Group 4: Political and Economic Implications - The article discusses the political implications of stablecoin legislation, particularly the concerns regarding potential conflicts of interest involving political figures and their ties to cryptocurrency projects [9]. - The USD1 stablecoin has gained significant traction, raising concerns about its use as a shadow banking tool for political financing [9]. - International coordination on stablecoin regulation is ongoing, with efforts to align regulations across jurisdictions to mitigate systemic risks [10].

端午安康 一路长虹。 今天就不点评nvidia 昨夜看业绩后的高开低走了 愿永不跳水 ...

每次出现这种现象都要重视。 9.24 那天都没有这么多批量。 ...