广告
加载中

与 OpenAI 分手后 微软 AI 等待灵魂人物

赵卫卫 2026-06-05 17:53
赵卫卫 2026/06/05 17:53

邦小白快读

EN
全文速览

本文梳理了微软和OpenAI结束独家合作后,微软AI的发展现状,以及核心负责人苏莱曼的处境和未来目标,核心重点信息整理如下

1.对比了当前全球四大AI企业核心负责人的社交媒体影响力:OpenAI奥尔特曼粉丝超500万,一条帖浏览量达数千万;谷歌哈萨比斯粉丝超100万,发帖保持低调;Anthropic阿莫代伊仅37.5万粉丝,却带领公司成为全球估值最高的AI初创;微软苏莱曼活跃但传播效果一般,恰好对应微软当前在AI赛道的位置。

2.明确了苏莱曼的身份和目标:他是DeepMind联合创始人,是微软去OpenAI化的重要推手,目前转岗聚焦MAI超级智能团队自研大模型,目标是让微软AI实验室进入全球前四。

3.披露了核心产品数据:当前微软Copilot日活仅600万,远低于ChatGPT的4.4亿和Gemini的8200万,新自研的MAI-Thinking-1推理模型还在内测,距离大规模商用还有较长周期。

本文对AI赛道的品牌战略调整、产品布局有很多参考干货,核心内容整理如下

1.品牌定位思路可参考:微软AI在和OpenAI分手后,调整品牌定位,主打“人文主义超级智能”标签,契合当下AI舆论存在抵制声音的环境,可有效消解公众对AI的抵触与恐惧,同时锚定人工智能助手、医疗超级智能、充足清洁能源三大长远应用领域,定位清晰切中用户情绪痛点。

2.产品研发方向调整:微软结束和OpenAI的独家合作后,转向从零开始自研大模型,推出针对企业部署优化的旗舰推理模型,同时发力Anthropic占据的企业级市场,避开消费端的红海竞争,推出适配自身办公生态的企业智能助手。

3.可吸取的教训:微软Copilot此前的升级多是补齐对手已经推出的功能,同质化严重,无法打动用户,导致用户增长长期停滞,可见AI品牌必须打造差异化的产品优势,才能拉动用户增长。

本文分析了当前全球大模型赛道的竞争格局变化,给AI相关卖家带来很多机会、风险参考,干货整理如下

1.增长市场机会:当前大模型消费端竞争已经进入白热化阶段,头部效应极强,OpenAI ChatGPT日活达到4.4亿,谷歌Gemini也有8200万日活,新进入者的机会空间很小;而企业级开发者市场还存在增长空间,头部玩家微软都在发力这一赛道,是值得卖家切入的增量市场。

2.风险提示:当前公众对AI的隐私泄露、强制植入产品问题敏感度极高,此前苏莱曼的不当发言就引发了大量用户的反对,暴露了产品理念和用户诉求的脱节,卖家推出AI产品一定要尊重用户选择权,保护用户隐私,避免引发舆论危机。

3.细分机会:当前AI自研浪潮开启,细分垂直领域还有很多机会,比如专注图像、语音、编码等垂直方向的模型,以及适配企业办公、网络安全等场景的AI产品,都有明确的市场需求。

本文围绕微软AI的战略调整展开分析,对传统工厂推进数字化转型、布局AI相关业务有不少启示,干货整理如下

1.数字化转型参考:当前AI技术已经深度渗透到To B服务领域,微软已经形成了云基础设施+办公全家桶+AI智能体的成熟商业化体系,能够覆盖工厂日常流程管理、研发辅助、生产数据处理等多类场景需求,工厂推进数字化升级可以选用这类成熟方案降低转型成本。

2.商业机会挖掘:AI大模型自研浪潮开启,整个AI产业对上下游配套产品的需求持续上涨,比如AI硬件配套、算力基础设施周边生产服务等领域,都有新的市场空间,有能力的工厂可以切入这些配套赛道,分享AI产业发展的红利。

3.可吸取的教训:微软此前推出消费级AI产品,功能多是跟随头部玩家,同质化严重,导致用户增长长期停滞,这提醒工厂不管是推进数字化转型还是拓展AI相关新业务,都要打造差异化的核心优势,才能在竞争中占据有利位置。

本文梳理了当前AI大模型行业的最新发展动向,给各类AI服务商、企业服务服务商提供了不少干货,整理如下

1.行业发展趋势:当前全球大模型行业已经从深度绑定头部技术的合作模式,转向自主研发核心模型的新阶段,微软和OpenAI结束独家合作就是标志性事件,未来企业对自主可控大模型、定制化场景模型的需求会持续上升,服务商可针对性布局相关业务。

2.核心客户痛点:当前企业用户既需要高性能、低成本的大模型产品,也需要适配自身业务场景的落地方案,同时C端用户和企业端都对AI的隐私安全、使用选择权有很高要求,服务商解决好这两个痛点,才能获得市场认可。

3.新技术路线参考:微软新推出的自研MAI-Thinking-1推理模型,采用MoE架构、256K上下文窗口,从零开始训练,针对数学运算、编码和企业部署优化,在同量级模型中达到顶尖水平,这个产品路线对AI技术服务商有参考价值。

本文分析了微软AI和云平台的发展现状与战略调整,给各类科技平台带来不少干货启示,整理如下

1.平台增长机会:当前AI产业发展给云平台带来明确的增长红利,OpenAI持续承诺使用微软Azure平台,2025年微软智能云利润达到505亿,利润率42%,收入同比增长34%,其中12%的营收增量由AI相关业务贡献,两项核心指标都高于亚马逊AWS和谷歌云,AI已经成为云平台新的增长引擎。

2.运营打法参考:微软已经建立了极为成熟的商业化产品网络,通过企业软件许可、Office 365订阅、Azure基础设施、GitHub工具链的组合,让AI产品轻松渗透到各个行业场景,这套捆绑打法值得各类平台借鉴。

3.风险规避提示:微软此前长期绑定OpenAI,受合作协议限制无法开展大算力自研模型,难以建立自己的AI叙事,这个教训说明平台必须掌握核心技术,才能把控长期发展的主动权,同时要重视用户对AI隐私、选择权的诉求,避免引发舆论风险。

本文记录了全球AI产业的最新变化,对AI产业研究者来说有很多值得研究的新动向、新问题,整理如下

1.产业新动向:全球AI大模型产业正在发生格局重塑,原本深度绑定的合作关系开始拆分,微软和OpenAI结束独家合作就是标志性事件,当前全球第一梯队大模型实验室为谷歌DeepMind、OpenAI、Anthropic,微软正投入资源冲击全球第四的位置,竞争格局出现新的变化。

2.产业新问题:本文提出了AI巨头企业发展中的新命题,也就是AI业务是否需要灵魂人物,微软AI当前的困境在于,核心盈利的云业务、开发者生态分别掌握在不同平级高管手中,自研模型负责人缺乏对核心战略板块的掌控,难以建立统一的AI叙事,这个新问题值得深入研究。

3.商业模式研究方向:当前AI产业已经形成了不同的成熟路线,既有OpenAI、谷歌的消费端路线,也有Anthropic、微软发力的企业级路线,微软的云+AI产品+开发者工具链的捆绑商业化模式,已经验证了盈利能力,值得研究者深入分析。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article outlines the current state of Microsoft's AI development following the end of its exclusive partnership with OpenAI, as well as the position and future goals of Mustafa Suleyman, Microsoft AI's core leader. Key takeaways are as follows:

1. A comparison of social media influence among core leaders of the world's four major AI players: OpenAI's Sam Altman has more than 5 million followers, with a single post garnering tens of millions of views; Google's Demis Hassabis has over 1 million followers and maintains a low posting profile; Anthropic's Dario Amodei has only 375,000 followers but has led the company to become the world's highest-valued AI startup; Mustafa Suleyman of Microsoft is active online but his content has limited reach, which aligns with Microsoft's current position in the AI race.

2. Suleyman's identity and goals are clarified: A co-founder of DeepMind, he is a key driver of Microsoft's "de-OpenAI" strategy. He has recently shifted roles to focus on in-house large model development at Microsoft's MAI superintelligence team, with the goal of placing Microsoft's AI lab among the top four globally.

3. Core product data is disclosed: Microsoft Copilot currently has only 6 million daily active users, far behind ChatGPT's 440 million and Gemini's 82 million. Microsoft's new in-house reasoning model MAI-Thinking-1 is still in internal testing, and it will take a long time before it reaches large-scale commercialization.

This article offers a wealth of actionable insights for brand strategy adjustment and product layout in the AI sector. Core takeaways are as follows:

1. Lessons for brand positioning: After splitting from OpenAI, Microsoft adjusted its AI brand positioning to emphasize "human-centric superintelligence." This positioning fits the current environment of growing public backlash against AI, effectively easing public resistance and fear of the technology. It also anchors Microsoft's AI development in three long-term application areas: AI assistants, medical superintelligence, and abundant clean energy, delivering a clear positioning that addresses key public sentiment pain points.

2. Adjustment to R&D direction: After ending its exclusive partnership with OpenAI, Microsoft shifted to building large models from scratch in-house. It has launched a flagship reasoning model optimized for enterprise deployment, is targeting the enterprise market currently dominated by Anthropic, and is avoiding the saturated consumer-facing competition by launching enterprise AI assistants that integrate natively with its existing office ecosystem.

3. Key takeaway: Most of Microsoft Copilot's past upgrades only added features that competitors had already launched, resulting in severe homogenization that failed to resonate with users and led to long-term stagnant user growth. This demonstrates that AI brands must build differentiated product advantages to drive user growth.

This article analyzes changes in the competitive landscape of the global large model industry, offering valuable insights into opportunities and risks for AI-related sellers. Key takeaways are as follows:

1. Growth market opportunities: Competition in the consumer-facing large model space has become cutthroat with extremely strong headwinds for new players: OpenAI's ChatGPT has 440 million daily active users, and Google's Gemini has 82 million, leaving very little room for new entrants. By contrast, the enterprise developer market still has ample room for growth. Even leading player Microsoft is prioritizing this segment, making it a high-potential incremental market for sellers to enter.

2. Risk warning: The public is now highly sensitive to AI-related risks including privacy leakage and forced product embedding. Mustafa Suleyman's recent inappropriate remarks triggered widespread user pushback, exposing the disconnect between product philosophy and user demand. When launching AI products, sellers must respect user choice and protect user privacy to avoid public relations crises.

3. Niche opportunities: As the wave of in-house AI development gains momentum, there are still many opportunities in vertical segments. These include specialized models focused on image processing, speech recognition, and coding, as well as AI products tailored for scenarios such as enterprise office work and cybersecurity, all of which have clear market demand.

This article analyzes Microsoft's AI strategic adjustment, offering valuable insights for traditional factories advancing digital transformation and laying out AI-related businesses. Key takeaways are as follows:

1. Reference for digital transformation: AI technology has now deeply penetrated the B2B service space. Microsoft has built a mature commercial system combining cloud infrastructure, its full office suite, and AI agents, which can cover multiple scenario needs for factories including daily process management, R&D assistance, and production data processing. Factories can adopt such mature solutions to reduce digital transformation costs.

2. New business opportunity exploration: As the wave of in-house large model development accelerates, demand for upstream and downstream supporting products across the AI industry continues to rise. Sectors such as AI hardware supporting and production services around computing power infrastructure offer new market space. Capable factories can enter these supporting segments to share the dividends of AI industry growth.

3. Key lesson: Microsoft's past consumer-facing AI products mostly followed leading players' features and suffered from severe homogenization, leading to long-term stagnant user growth. This serves as a reminder that whether advancing digital transformation or expanding new AI-related businesses, factories must build differentiated core advantages to secure a favorable competitive position.

This article sorts out the latest developments in the large model industry, offering valuable insights for all types of AI and enterprise service providers. Key takeaways are as follows:

1. Industry development trends: The global large model industry has moved from a model of deep reliance on partnership with leading technology players to a new phase of independent core model development. The end of Microsoft and OpenAI's exclusive partnership is a landmark event for this shift. Going forward, enterprise demand for self-controllable large models and custom scenario-specific models will continue to rise, and service providers can layout related businesses accordingly.

2. Core customer pain points: Enterprise users today demand both high-performance, low-cost large model products and implementation solutions adapted to their own business scenarios. At the same time, both B2C and B2B users place high requirements on AI privacy security and usage choice. Service providers that address these two pain points well will earn market recognition.

3. Reference for new technology roadmaps: Microsoft's newly developed in-house reasoning model MAI-Thinking-1 adopts a MoE architecture with a 256K context window, is trained from scratch, and is optimized for mathematical computation, coding, and enterprise deployment, delivering top-tier performance among models of the same scale. This product roadmap offers valuable reference for AI technology service providers.

This article analyzes the current state and strategic adjustment of Microsoft's AI and cloud platform business, offering valuable insights for all types of technology platforms. Key takeaways are as follows:

1. Platform growth opportunities: The growth of the AI industry has brought clear growth dividends to cloud platforms. OpenAI has repeatedly committed to continued use of Microsoft Azure, and by 2025, Microsoft Intelligent Cloud is projected to reach $50.5 billion in profit with a 42% margin, and 34% year-over-year revenue growth, with 12% of incremental revenue coming from AI-related businesses. Both core metrics outperform Amazon AWS and Google Cloud, confirming that AI has become a new growth engine for cloud platforms.

2. Reference for operational strategy: Microsoft has built an extremely mature commercial product network, through the combination of enterprise software licensing, Office 365 subscriptions, Azure infrastructure, and the GitHub toolchain, allowing its AI products to easily penetrate into industry scenarios across sectors. This bundled go-to-market approach is a valuable reference for all types of platforms.

3. Risk mitigation lessons: Microsoft's long-term exclusive reliance on OpenAI left it restricted by partnership agreements from developing high-computing in-house models, making it difficult to build its own independent AI narrative. This lesson demonstrates that platforms must master core technology to control the initiative of long-term development. They must also prioritize user demands around AI privacy and usage choice to avoid public opinion risks.

This article documents the latest changes in the global AI industry, highlighting new trends and research questions for AI industry researchers. Key points are as follows:

1. New industry trends: The global large model industry is currently undergoing landscape restructuring, and previously deeply integrated partnerships are beginning to split, with the end of Microsoft and OpenAI's exclusive partnership as a landmark event. The current first tier of global large model labs consists of Google DeepMind, OpenAI, and Anthropic, and Microsoft is investing resources to claim the fourth position, bringing new changes to the competitive landscape.

2. New industry questions: This article raises a new proposition for large AI companies: whether AI businesses require a charismatic visionary leader. Microsoft AI's current dilemma stems from the fact that its core profitable cloud business and developer ecosystem are controlled by separate peer executives, leaving the head of in-house model development without control over core strategic segments, making it difficult to build a unified AI narrative. This new question merits in-depth research.

3. Directions for business model research: The AI industry has now developed multiple mature business routes, including the consumer-facing route of OpenAI and Google, and the enterprise-focused route of Anthropic and Microsoft. Microsoft's bundled commercial model of cloud + AI products + developer toolchain has already validated its profitability, making it worthy of further in-depth analysis by researchers.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

微软的AI产品是否足以让任何人成为灵魂人物?

谁是美国AI巨头中影响力最低的关键先生?答案很可能是微软 AI首席执行官穆斯塔法·苏莱曼(Mustafa Suleyman)。

以社交媒体X 为例,OpenAI的山姆·奥尔特曼无疑是粉丝量最高的,超过 500万粉丝,他把 X当作 ChatGPT和 OpenAI的免费营销平台,一条帖子的浏览量会高达数千万,「Codex活跃用户超过 100万」的帖子,能引发超过上万人分享回应。

谷歌的哈萨比斯排名第二,有超过100万用户关注他,但他一直在社交媒体上保持低调,发帖频率很低,上一次高曝光的帖子,还是 2023年 Gemini 1.0发布,人们关注和期待他,更多是因为他在 AI学术上和 DeepMind发展的权威地位。

Anthropic的达里奥·阿莫代伊也很低调,他根本不把 X当社交媒体来用,只更新过几条内容,把读者引流到自己的官方博客网站中。虽然他在社交媒体上粉丝不高,只有区区 37.5万,但这不妨碍他带领着Anthropic成为全球人工智能估值第一的初创公司。

相比之下,第四名穆斯塔法·苏莱曼像极了微软如今在 AI行业的位置。

苏莱曼在社交媒体X 上是最活跃的,但内容的传播效果和量级明显一般。两年前,作为 DeepMind的联合创始人之一,苏莱曼加入微软成为微软 AI的 CEO,全面负责微软消费者端 AI产品,直接汇报给微软CEO萨蒂亚·纳德拉。

今年4 月,微软与 OpenAI分手,正式调整为非独家合作关系。而在这一幕发生的 40天前,苏莱曼也从微软 AI的全面领导者位置抽身,聚焦内部组建 MAI超级智能团队,把更多精力花在打造独立前沿模型开发上。

可以说,苏莱曼是微软「去 OpenAI化」的重要推手,也是纳德拉光环下,未来承载微软 AI战略的灵魂人物。

「我们的目标是证明我们能够跻身世界前四,真正重要的实验室有三个:谷歌 DeepMind、OpenAI和 Anthropic。我们目前还不是其中之一,而这始终是我的目标。」苏莱曼在最近的微软 Build大会后说。

要成为行业第四,就要有对应的灵魂人物。

而成为 AI行业的灵魂人物,注定要有第一性原理作为强大信仰,要靠产品和技术成果来造就,要靠时间的锤炼和打磨来自我证明,而对苏莱曼来说,更深刻的问题是,微软的 AI产品是否足以让任何人成为灵魂人物?

人文主义超级智能?

微软 Build 2026大会上,微软拿出了一批产品,包括超级应用、内部推理模型、网络安全工具以及类似 OpenClaw的人工智能助理,尤其是在模型上,推出包括一个旗舰推理模型 MAI-Thinking-1,和 6个专注于图像、语音、转录和编码的新模型。

MAI-Thinking-1是微软第一款完全自主研发的推理模型,35亿活跃参数,MoE架构,256K上下文窗口,从头训练,不含任何来自 OpenAI的数据蒸馏。

一个月之前,微软与OpenAI分手,微软放弃了对 OpenAI知识产权的独家授权,解除了向其支付的收入分成义务,换取的是一份延续至 2032年的 IP使用权,以及 OpenAI对 Azure平台的持续承诺。

如今,微软要自力更生,作为微软AI的负责人,苏莱曼在介绍微软自主研发的推理模型时特别强调,「从零开始训练的,专为严肃的数学运算、编码和实际企业部署而设计」。

在它的产品介绍中,MAI-Thinking-1在同量级模型中处于顶尖水平:在关键的软件工程基准测试中,其表现与领先模型不相上下;而在盲测对比评估中,其在人类偏好方面更是达到了与 Sonnet 4.6持平的水平。

即便微软强调这款模型在性能和价格上的优势,但微软仍然落后它的同行OpenAI和 Anthropic。有意思的是,苏莱曼承认这种差距,他公开说「在六个月之内,我们追平了一个巨大差距」。

微软 Build大会上最明显的一点,就是朝着 Anthropic占据的企业级市场冲刺。

在线办公智能体 Scout,就是一款基于开源人工智能平台 OpenClaw构建的智能助手,它可以搭配微软办公全家桶使用,帮助企业完成诸如管理日程、费用报销、撰写电子邮件等工作。

苏莱曼公开对媒体说,微软更关注的是 Anthropic主打的企业开发者市场,对谷歌、Meta、OpenAI偏消费化路线没有那么关心。

而他给微软AI品牌打造的标签,是「人文主义超级智能」,也就是以人为本,意在在当下充斥 AI抵制声音的舆论环境里,消解外界对微软 AI战略的抵触与恐惧。在这标签之下,微软 AI长远更看重的三个应用领域,分别是人工智能助手、医疗超级智能和充足的清洁能源。

但眼下的竞争更为紧迫,微软拿出的旗舰模型还处于内测阶段,在微软 Foundry上向部分企业开放,距离落地大规模商用部署仍有较长周期,这也是微软 AI尚不足以产生强叙事的根本原因。

苏莱曼这两年,微软AI难言成功

苏莱曼执掌微软AI业务的两年间,旗下对话产品 Copilot周活长期徘徊在 2000万上下,基本原地踏步。

这不是他一个人的责任。2024年苏莱曼加盟微软,微软以 6.5亿美元打包引进他和 Inflection AI的核心团队,苏莱曼负责的微软 AI,接管Copilot、Bing、Edge全线消费者产品。

但有一个约束一直都在,那就是微软与OpenAI的捆绑协定,明确禁止微软训练超过特定算力阈值的模型,这等于封锁了微软自研大模型的可能性,所以苏莱曼后来对财富杂志说,这是一个很大的限制,尤其是对于微软这种体量的公司来说。

一年后,苏莱曼发布了 Copilot的一次重大升级:记忆功能、个性化伴侣定位、实时屏幕分析、购物等功能,让 Copilot能代用户完成订票等任务。

但实际上,这些功能都已经是 ChatGPT和谷歌一年前就已经发布的技术,微软 Copilot只是在补齐对手们已经做到的功能,这些同质化的更新自然无法支撑用户增长的叙事,用户很难被打动。

半年后,谷歌Gemini 3.0与 ChatGPT打得难解难分,OpenAI不得不拉起红色警报。

但苏莱曼却在引发一场舆论风暴,他在社交媒体的发言引发大量质疑与反对。他说:哎,这么多愤世嫉俗的人!每次听到有人说人工智能不怎么样,我都觉得好笑。

「我小时候可是用诺基亚手机玩贪吃蛇长大的!现在我们能和超级智能的人工智能流畅地对话,还能生成任何图像/视频,居然有人对此不以为然,这真是让我匪夷所思。」

很多用户反对苏莱曼的发言,不是不认可AI技术,而是因为微软把不成熟的 AI对话技术塞入电脑系统和办公全家桶的产品中,人们反对的是无法选择退出、隐私和数据收集等问题。

当时有媒体认为,暴露出微软产品理念和用户实际诉求之间严重脱节,它暴露出苏莱曼在微软的深层困境:他相信产品,但是产品还没有达到他相信的高度。

这期间的苏莱曼,不得不为业务着急,因为两周之前,他成为了微软MAI超级智能团队的负责人,提出了「人文主义超级智能的愿景」,试图为陷入增长停滞的消费者产品线寻找破局方案。

现在回头看,那是苏莱曼角色转型的开始,他不再是一个消费者产品业务操盘手了,也给微软和OpenAI分手埋下伏笔。

今年3 月,苏莱曼的职权收窄。他专注超级智能和前沿模型研发,在未来五年内为微软交付世界级模型,而消费者和企业 Copilot团队合并,交给前 Snap高管雅各布·安德烈欧负责。一个月后,微软和 OpenAI官宣分手,结束独家授权协议。

Sensor Tower发布的数据显示,今年 2月,微软 Copilot应用拥有600万日活用户,这是苏莱曼留下的成绩单。而同期 OpenAI的 ChatGPT拥有 4.4亿日活用户,谷歌的 Gemini拥有 8200万日活用户。

等苏莱曼到2031年

微软AI目前没有真正意义上的灵魂人物,这不是苏莱曼的个人问题。

微软的AI叙事,一直是与 OpenAI进行了深刻绑定,OpenAI是模型,微软是分销渠道。如今,二者的松绑才刚刚开始,微软需要足够的产品和技术成果,去建立自己的 AI叙事节奏。

更深层的问题是,微软AI需要一个真正的灵魂人物吗?或者说微软的 AI产品是否足以让任何人成为灵魂人物?

答案可能是残酷的,在当下的整个AI产业链中,微软云是无人能及的存在,它是云厂商中最赚钱的一个。2025年微软智能云(含Azure)利润 505亿,利润率 42%,两项指标都高于对手亚马逊 AWS和谷歌云,它收入同比增长 34%,营收增量里约 12% 由 AI相关业务贡献。

云厂商连接着产业上游存储芯片公司,以及下游的大模型厂商和无数企业用户。微软已经建立了一套极为成熟的商业化机制:企业软件许可、Office 365订阅、Azure基础设施、GitHub工具链,这套成熟的产品网络,让微软 AI可以轻轻松松渗透到互联网的各个角落。

但用户选择用微软的AI产品,是因为苏莱曼做得好吗?

不是,GitHub Copilot的开发者渗透率,是微软 CTO凯文·斯科特的功劳,他长期负责底层技术战略和开发者生态;企业付费使用 AI大模型服务,是斯科特·格思里(Scott Guthrie)的职权范围,他负责微软的云业务,是目前微软 AI最赚钱最稳健的部分;负责开发者平台、智能体工厂的杰伊·帕里克(Jay Parikh),他们都是直接汇报给纳德拉。

苏莱曼名义上是微软AI负责人,但实际上最赚钱、最具战略意义的领域掌握在其他平级高管手里,这让他注定现在难以成为微软AI的灵魂人物。

未来呢?苏莱曼必须背水一战了。

因为没有了对OpenAI的依赖,微软所有 AI产品和 Agent的长期竞争力,都要靠自身的底层模型建设,模型能否实现高性能、低成本、可控落地,直接决定微软上层 AI产品与底层云基础设施的长期稳定性,这也就是苏莱曼在职权调整后的表态:模型就是产品。

苏莱曼会成为微软AI的灵魂人物吗?现在消费者业务已经证伪了这个命题,模型层的进展比以往任何时候都更关键,留下来的机会只有一个,那就是他主导的 MAI模型在未来五年内,能够抢占到世界前四的位置,与谷歌、Anthropic和 OpenAI站在一个梯队。

注:文/赵卫卫,文章来源:蓝洞商业(公众号ID:value_creation),本文为作者独立观点,不代表亿邦动力立场。

文章来源:蓝洞商业

广告
微信
朋友圈

这么好看,分享一下?

朋友圈 分享

APP内打开

赞 +1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0