高盛预计阿里未来数年资本开支将达4600亿元，远高于该公司此前3800亿元的目标。报告认为，AI推理需求激增将推动阿里、字节等巨头持续加码投资，其中阿里押注企业级AI 云，字节则发力消费级应用。尽管技术效率不断提升，但AI需求的爆发令资本开支继续扩张。 硬·AI 作者 | 张雅绮 编辑 | 硬 AI 高盛认为，需求的爆炸式增长将继续推动云服务厂商的资本开支（Capex）上行。中国互联网巨头在AI领域的战略路径正日益分化，阿里巴巴凭借其全栈能力押 注企业级AI云市场，而字节跳动则在消费级（To-C）应用上全力冲刺。 高盛在23日的报告中，上调了对中国头部云厂商的资本开支预测， 预计阿里巴巴在2026至2028财年的合计资本开支将达到4600亿元人民币，远高于该公司此前 3800亿元的目标。 该行认为，AI推理需求的激增是支撑这一判断的核心逻辑，更高的计算效率反而可能提升资本开支向云收入的转化率，从而加速收入增长。 与此同时，巨头们的战略差异愈发明显。高盛指出， 阿里巴巴在外部AI云收入规模和企业级（To-B）服务方面处于领先地位， 其商业化路径更为清晰。 而字节 跳动则凭借旗下聊天机器人"豆包"在To-C领域和 ...

Core Insights - Qwen3 Max has emerged as the leading model in the AI trading competition, surpassing DeepSeek and achieving significant profitability [1][32] - The competition, Alpha Arena, showcases the capabilities of various AI models in real market conditions, emphasizing the financial market as a training ground for AI [30][32] Performance Summary - Qwen3 Max achieved a return of +44.38%, with an account value of $14,438 and total profit of $4,438 [11] - DeepSeek V3.1 follows with a return of +20.92%, account value of $12,092, and total profit of $2,092 [11] - Other models, such as Claude 4.5 Sonnet, Grok 4, Gemini 2.5 Pro, and GPT-5, reported negative returns, with GPT-5 showing the largest loss at -71.48% [10][11] Competition Dynamics - The competition began on October 18 and has seen Qwen3 Max steadily improve its position, particularly after a significant drop in all models on October 22 [22][24] - Qwen3 Max's strategy has been characterized as "quick and precise," allowing it to capitalize on market opportunities effectively [8][32] - The competition has highlighted the contrasting performance of models, with Qwen3 Max and DeepSeek being the only two models consistently performing well [22][24] Market Implications - The success of Qwen3 Max indicates the growing competitiveness of Chinese AI models in the global market, particularly in high-risk financial environments [33] - The Alpha Arena competition serves as a demonstration of how AI can adapt and thrive in real-world financial scenarios, reinforcing the notion that financial markets are ideal for AI training [30][32]

来源：21世纪商业评论 本文字数：1777｜预计3分钟读完 各显神通，胜负未分。 作者丨子川 AI圈和投资圈，最近都被一场炒股大赛刷屏！ 一家名叫Nof1的初创公司，组织了一场AI炒股大赛——Alpha Arena。 而收益率最低的居然是用户量最多的GPT-5，目前账户还剩5414美金，亏损高达45.81%，属实是亏麻了！ 纵观整个收益曲线，GPT-5只有在比赛开始的时候小有盈利，其他时间都在亏，特别是19号以后，收益曲线一降再降，根本就止不住。 收益率仅次于DeepSeek的是Grok 4，目前收益是11.7%，此前最高的收益率是38.15%。 和GPT-5相反的是，除了在第一天亏了点外，其余时间都在盈利，曾经有段时间收益率一度接近DeepSeek。 这并非是模拟盘，而是给每个AI工具1万美元，让它们自己炒股，并对收益进行24小时监控。 此次比赛从美东时间10月18号开始，为期2周，在11月3号结束，目前是第四天。 1 顶级阵型 参赛的AI选手阵容堪称豪华，汇集了国内外最顶级的AI模型。 海外组： OpenAI的GPT-5、谷歌的Gemini 2.5 Pro、马斯克的Grok 4，以及Anthropic的C ...

AI Model Performance in Trading Competition - Major AI models like Claude Sonnet 4.5, Gemini 2.5 Pro, and GPT 5 experienced significant losses within a week since October 18 [1] - Gemini 2.5 Pro and GPT 5 declined by over 50% [1] - DeepSeek Chat V3.1, Qwen3 Max, and Grok 4 remained profitable but with sharp drawdowns [1]

这几天，各大AI社群被一场"投资直播"刷屏。网友们实时追踪六大AI模型的交易表现，讨论的热情程度甚至超过研究自己炒股，这是一场用真金白银进行的 AI投资对决。 这场由初创公司Nof1发起的"Alpha Arena"基准测试，并非模拟交易。主办方为了衡量AI投资能力，给每个模型账户发放了一万美元的启动资金，让它们在 真实市场自主交易数字货币。Alpha Arena将直播整个过程，价格会实时波动，并对实时收益进行排名，还可以看到每个模型的交易思路。 按目前盈利能力排名，参与这次比赛的有DeepSeek chat v3.1、Claude Sonnet 4.5、Grok 4、Qwen3 Max、Gemini 2.5 pro、GPT 5六个AI模型，有三家海外头部 模型和国内两家模型。这一投资交易竞赛开始于美东时间10月18日，将持续两周，于11月3日结束。 真实市场交易有趣的地方在于，市场永远有波动，是不可预测的，即便最先进的AI也无法保持稳定的收益。正如官方所说的那样，"市场是智力的终极考 验"。 目前过去了4天，已经历了一些波动。前三天，排名第一的DeepSeek收益率还一度接近40%，盈利超过4000美元，但1 ...

Core Viewpoint - The article discusses the ongoing AI investment competition called "Alpha Arena," initiated by the startup Nof1, where various AI models are trading real cryptocurrencies with a starting capital of $10,000 each, showcasing their investment capabilities in a live environment [3][5]. Group 1: Competition Overview - The competition began on October 18 and will last for two weeks, ending on November 3, featuring six AI models: DeepSeek chat v3.1, Claude Sonnet 4.5, Grok 4, Qwen3 Max, Gemini 2.5 pro, and GPT 5 [5][9]. - As of October 21, DeepSeek was leading with a return of approximately 10%, having previously reached nearly 40% [5][7]. - Other models like Grok 4 and Claude have shown varying performance, with Grok 4 initially close to DeepSeek but later fluctuating around the breakeven point [7][9]. Group 2: Performance Analysis - DeepSeek's stable performance is attributed to its professional background in quantitative trading, employing a straightforward strategy without frequent trading [9][11]. - In contrast, Gemini 2.5 has been criticized for its erratic trading style, leading to significant losses, with a decline exceeding 30% at one point [11][13]. - Grok 4 is noted for its aggressive trading approach, while Claude's analytical skills are hampered by indecision, resulting in frequent trading mistakes [13][14]. Group 3: Insights on AI Trading - The competition highlights the distinct "personalities" of the AI models, akin to human traders, with each model exhibiting unique trading strategies and risk profiles [9][11]. - Despite the potential benefits of AI in providing logical investment strategies, industry experts caution that AI lacks the ability to predict future market movements and does not understand individual investor circumstances [13][14]. - The article emphasizes that while AI can help mitigate emotional biases in trading, the combination of rational tools and human insight may yield the best investment outcomes [14].

海内外AI投资大战打响，胜负未定。 这几天，各大AI社群被一场"投资直播"刷屏。网友们实时追踪六大AI模型的交易表现，讨论的热情程度甚至超过研究自己炒股，这是一场用真金白银进行的 AI投资对决。 这场由初创公司Nof1发起的"Alpha Arena"基准测试，并非模拟交易。主办方为了衡量AI投资能力，给每个模型账户发放了一万美元的启动资金，让它们在 真实市场自主交易数字货币。Alpha Arena将直播整个过程，价格会实时波动，并对实时收益进行排名，还可以看到每个模型的交易思路。 按目前盈利能力排名，参与这次比赛的有DeepSeek chat v3.1、Claude Sonnet 4.5、Grok 4、Qwen3 Max、Gemini 2.5 pro、GPT 5六个AI模型，有三家海外头部 模型和国内两家模型。这一投资交易竞赛开始于美东时间10月18日，将持续两周，于11月3日结束。 真实市场交易有趣的地方在于，市场永远有波动，是不可预测的，即便最先进的AI也无法保持稳定的收益。正如官方所说的那样，"市场是智力的终极考 验"。 目前过去了4天，已经历了一些波动。前三天，排名第一的DeepSeek收益率还一度接 ...

Core Insights - The article discusses a competition called Alpha Arena, where six leading AI models are tested in a real trading environment with an initial capital of $10,000 each to determine which model performs best in stock trading [4][5][7]. Group 1: Competition Overview - The competition features top AI models including OpenAI's GPT-5, Google's Gemini 2.5 Pro, Anthropic's Claude 4.5 Sonnet, xAI's Grok 4, Alibaba's Qwen3 Max, and DeepSeek V3.1 Chat [5][6]. - Each model receives identical market data and trading instructions, simulating a level playing field for performance comparison [7][11]. Group 2: Performance Metrics - As of the latest updates, DeepSeek V3.1 leads with an account value of $13,677, achieving a return of +36.77% and a total profit of $3,677 [9]. - Grok 4 follows with an account value of $13,168 and a return of +31.68%, while Claude Sonnet 4.5 has an account value of $11,861 and a return of +18.61% [9]. - In contrast, GPT-5 and Gemini 2.5 Pro are at the bottom, with account values of $7,491 and $6,787, reflecting returns of -25.09% and -32.13% respectively [9]. Group 3: Trading Strategies and Decisions - The models are required to make trading decisions based on real-time data, including price indicators and account information, determining whether to hold, buy, or sell [11]. - DeepSeek's trading strategy has been noted for its effectiveness, attributed to its quantitative trading background [12]. Group 4: Market Dynamics and Model Adaptation - The performance of the models fluctuates significantly, with DeepSeek and Grok initially experiencing losses before rebounding, while GPT-5 and Gemini 2.5 Pro show a contrasting trend of initial gains followed by declines [28][33]. - The competition highlights the rapid changes in financial markets and the necessity for models to adapt quickly to evolving conditions [10][44]. Group 5: Implications for AI Development - The article posits that financial markets serve as an ideal training ground for AI, as they present complex, real-world challenges that require models to interpret volatility and manage risks effectively [49][50]. - The competition is framed as a new type of Turing test, assessing whether AI can survive in uncertain environments rather than merely demonstrating cognitive abilities [54].

Core Insights - The article discusses a new experiment called Alpha Arena, initiated by nof1.ai, where top AI models compete in a real trading market to determine which can perform best in stock trading [1][2] - The participating models include OpenAI's GPT-5, Google's Gemini 2.5 Pro, Anthropic's Claude 4.5 Sonnet, xAI's Grok 4, Alibaba's Qwen3 Max, and DeepSeek V3.1 Chat [2] Competition Setup - Each model starts with an initial capital of $10,000 and receives the same market data and trading instructions [5] - The competition resembles an "open-book exam," where models are provided with real-time data, including prices and indicators, to make trading decisions [6][8] Performance Overview - As of October 20, 2023, at 7:30 AM, DeepSeek V3.1 led with a profit of $2,264, followed by Grok 4 with $2,071, while Gemini 2.5 Pro and GPT-5 were at the bottom with losses of $3,542 and $2,419 respectively [12] - By 10:00 AM, the standings changed significantly, with DeepSeek and Grok-4 experiencing declines, while Qwen3 Max and GPT-5 showed upward trends [12] Trading Strategies and Results - DeepSeek V3.1's trading strategy involved multiple long positions across various cryptocurrencies, resulting in a total unrealized profit of $2,309.79 [16] - Claude Sonnet 4.5 and Qwen3 Max also reported profits, while Gemini 2.5 Pro showed slight recovery after earlier losses [20][21] Market Dynamics - The article emphasizes the volatility of financial markets, noting that models must adapt quickly to changing conditions [10] - The competition serves as a new type of Turing test, assessing models on their ability to survive in uncertainty rather than just their cognitive capabilities [52] Conclusion - The Alpha Arena experiment highlights the potential of financial markets as a training ground for AI, where models can learn from real-time data and adapt to complex environments [47][48]

Group 1: AI Trading Competition - A trading competition was launched by the AI research lab nof1.ai, featuring six top AI models each trading with $10,000 in real cryptocurrency, ensuring a fair environment with identical prompts and market conditions [1] - The six AI models participating are GPT-5, Claude Sonnet 4.5, DeepSeek Chat V3.1, Gemini 2.5 Pro, Grok 4, and Qwen3 Max, showcasing their trading capabilities [1] - As of the latest update, DeepSeek is leading the competition with a total portfolio value close to $14,000 and a return of approximately 40%, having peaked at nearly $15,000 [4] Group 2: Semiconductor Industry - The Netherlands is seeking to resolve the deadlock surrounding ASML Semiconductor, which has implications for China-Netherlands economic relations and the global automotive chip supply chain [5] - The Dutch Minister of Economic Affairs indicated that the intervention in ASML's affairs was aimed at preventing the transfer of business and intellectual property out of Europe, emphasizing the need for cooperation with China [5] - The deadlock originated from the U.S. "penetration rules" announced on September 29, leading to direct Dutch government intervention in ASML's internal matters [5] Group 3: AI Developments by Alibaba - Alibaba's Quark is secretly advancing a project known as "C Plan," which focuses on conversational AI applications, with expectations for a new product launch soon [6] - The project has been in development for a significant period and is anticipated to require long-term investment and technological breakthroughs [6] - Speculations about the name "C Plan" suggest it may relate to "Chat" or be a competitive move against ByteDance's Doubao [6] Group 4: Financial Performance of Companies - iFlytek reported a net profit of 172 million yuan for the third quarter, marking a year-on-year increase of 202.4%, with total revenue of 6.078 billion yuan, up 10.02% [9] - Apple achieved a historic market capitalization of $3.89 trillion, making it the second-largest company in the U.S. by market value, following Nvidia [8] Group 5: Innovations in Medical Technology - Science Corporation, founded by the former president of Neuralink, has developed a retinal microchip implant that restores vision for blind patients, allowing them to read and engage in activities like crossword puzzles [10] - The implant works by bypassing damaged photoreceptors in the retina, utilizing signals from a camera mounted on glasses to stimulate the retina [10]