广告
加载中

马云出手 阿里云拿下全国第一

千帆 2026-06-22 10:12
千帆 2026/06/22 10:12

邦小白快读

EN
全文速览

本文核心内容是阿里云经过十多年持续投入,拿下中国大模型训推公有云市场全国第一的发展历程和最新成果,核心干货如下

1. 核心成果数据:据IDC 2025H2报告,中国大模型训推公有云市场规模79.38亿元,阿里云以42.2%的市场份额位列第一,同比增长116%,最新旗舰模型Qwen3.7-Max位居国产模型第一,综合能力接近海外顶级GPT、Claude等模型。

2. 发展历程:早在2008年阿里就因自身业务需求启动自研云计算系统“飞天”,顶着业内“不现实”“为时过早”的质疑,坚持每年投入10亿元,2013年突破5k集群技术壁垒,后续逐步发展为全球第三大公有云服务商。

3. 最新布局:阿里云已经完成全栈Agent化升级,推出适配智能体时代的核心产品千问云,阿里明确将全栈AI作为核心战略,持续加码投入。

本文梳理了AI云产业的发展趋势和头部玩家的路径,对品牌商布局数字化、把握行业方向有较多参考干货,内容如下

1. 产业与消费技术趋势:当前AI行业竞争已经从单点模型能力比拼,转向底层算力、数据调度与云平台的体系化综合能力比拼,未来云服务的主要调用者会变成AI智能体,交互逻辑从点击操作转向指令调用,品牌需要提前适配这一变化。

2. 产品研发与技术布局参考:阿里云的发展证明,长期坚定投入底层核心技术才能建立长期壁垒,阿里依托自身电商、零售等多元业务场景形成技术研发-落地的迭代闭环,品牌也可参考这一路径,依托自身业务场景打磨AI能力。

3. 商业化现状:当前AI云商业化已经进入落地阶段,阿里云AI相关产品收入已经占外部收入30%以上,MaaS年化收入超80亿元,年底有望突破300亿,说明AI赋能商业已经进入收获期。

本文梳理了AI云行业的最新变化,能给卖家挖掘新增长机会、选择技术服务商提供参考,核心干货如下

1. 新机会提示:AI已经进入智能体落地阶段,云服务入口逻辑重构,成熟的AI云服务能帮助卖家降本增效,卖家可依托阿里云这类头部平台的标准化AI能力,快速开发适配自身业务的运营、设计、客服类AI应用,不需要过高的技术投入就能享受AI红利。

2. 行业变化:当前底层算力服务的稳定性、成本都在持续优化,已经能支撑卖家开展各类AI相关的业务创新,竞争焦点转向体系化能力,对中小卖家来说,依托成熟平台比自行研发性价比更高。

3. 风险提示:当前AI赛道技术更新快,头部玩家竞争加剧,卖家选择技术服务商时要优先选择技术积累深、迭代节奏快、稳定性经过市场验证的平台,阿里云深耕行业17年,大模型保持三个月连更三个版本的节奏,技术稳定性经过多年双11等大流量验证,可靠性更高。

本文的内容能给工厂推进数字化转型、抓住AI时代的商业机会提供参考,核心干货如下

1. 数字化转型启示:早年国内互联网企业普遍使用海外IOE技术架构,阿里云从零开始自研底层云计算技术,最终实现对海外方案的替代,说明走自研路线、依托国内技术服务也能完成数字化升级,还能降低长期使用成本。

2. 商业与效率提升机会:AI大模型已经进入落地阶段,阿里云推出的千问云等产品,已经封装了标准化的AI能力,工厂可以借助这些能力快速开发产品设计、生产流程优化、供应链管理等场景的AI应用,有效提升生产和设计效率,不需要搭建完整的自研团队。

3. 转型路径参考:阿里云是从解决自身业务痛点出发,逐步打磨技术再对外输出,工厂推进数字化也可以参考这一思路,从自身最痛的生产、管理痛点切入,分步落地,不需要盲目追求一步到位完成全链路数字化,降低转型失败风险。

本文梳理了AI云行业的最新发展趋势和技术动向,对科技服务商把握行业方向有较高参考价值,核心干货如下

1. 行业发展趋势:当前AI行业竞争已经从比拼模型参数、对话能力,转向比拼底层算力、云平台的整套体系化能力,行业对稳定、高效、低成本支撑复杂智能任务的系统需求越来越高;未来云服务的调用主体会从人转向AI智能体,入口逻辑从点击操作转向指令调用,这是服务商未来核心的研发方向。

2. 最新技术动向:头部玩家已经开始布局下一代云服务,阿里云已经完成全栈Agent化升级,推出全新核心产品千问云,打造了Agent可读指令体系,封装了标准化技能库和工具集,支持AI智能体自主调用能力,最新Qwen3.7-Max大模型已经位居国产第一,更新节奏明显加快。

3. 客户痛点与解决方案方向:当前客户已经不满足于单点模型能力,需要能落地的整套解决方案,服务商需要参考阿里路径,搭建从底层基础设施到上层应用的完整体系,结合落地场景形成迭代闭环,才能满足客户需求。

本文梳理了头部云平台的发展路径和最新战略,对平台商把握行业方向、规避风险有参考价值,核心干货如下

1. 当前市场对平台的核心需求:AI大模型发展进入新阶段,市场对平台的需求已经从有没有算力,转变为能不能提供好用、低成本、能支撑复杂AI任务落地的整套体系化能力,单点技术领先已经无法满足市场需求。

2. 头部平台的可借鉴做法:阿里云坚持十多年持续投入底层技术,最早完成系统级重构,为AI时代转型打下基础;最新推出适配智能体时代的千问云,重构了云服务的交互调用逻辑;同时依托自身多元生态场景形成技术迭代闭环,持续加码自研芯片和基础设施投入,建立了长期竞争优势。

3. 风向与风险提示:当前赛道竞争已经进入密集正面交锋阶段,头部厂商都在加码布局,平台需要规避只拼单点模型能力的误区,要注重整套技术体系的搭建,坚持长期投入,不能追求短期速度,才能在长期竞争中占据优势。

本文梳理了中国云计算产业的发展历程和AI时代的最新动向,对产业研究有较高的资料价值,核心干货如下

1. 产业最新动向:当前中国AI基础设施竞争已经进入新阶段,2025年中国大模型训推公有云市场规模已经达到79.38亿元,阿里云以42.2%的市场份额位居第一,同比增长116%;竞争焦点从单点模型能力转向底层云平台体系化能力比拼,商业化落地加速,阿里云AI相关收入已经占外部收入30%以上,MaaS年化收入超80亿。

2. 产业发展新特征:AI大模型已经进入智能体发展阶段,云服务入口逻辑发生重构,未来云服务的主要调用主体会从人转变为AI智能体,交互方式从点击转向指令;早年中国互联网普遍采用的IOE架构已经被自研分布式云计算架构替代,成为行业共识。

3. 重点研究方向:早年行业头部人物对云计算的判断差异,长期投入对竞争格局的影响,阿里“底层技术+业务场景”形成迭代闭环的商业模式,以及智能体时代云服务架构的重构方向,都是值得深入研究的新课题。

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 covers Alibaba Cloud's 10-plus years of sustained investment, its journey to the top of China's public cloud market for large model training and inference, and its latest milestones. Key takeaways are as follows:

1. Core results and data: According to IDC's 2025 H2 report, China's public cloud market for large model training and inference reached 7.938 billion yuan. Alibaba Cloud ranks first with a 42.2% market share, representing a 116% year-on-year increase. Its latest flagship model Qwen3.7-Max tops all domestic Chinese models, with overall capabilities approaching top international models such as GPT and Claude.

2. Development history: Alibaba began developing its self-developed cloud computing system "Feitian" in 2008 to meet its own business needs. Despite widespread industry skepticism that the project was "unrealistic" and "premature", Alibaba sustained an annual investment of 1 billion yuan. It broke through the technical barrier of supporting 5,000-node clusters in 2013, and eventually grew into the world's third-largest public cloud service provider.

3. Latest strategic layout: Alibaba Cloud has completed a full-stack Agent-oriented upgrade and launched Qwen Cloud, its core product built for the agent era. Alibaba has explicitly positioned full-stack AI as its core strategy and will continue to ramp up investment.

This article outlines the development trends of the AI cloud industry and the growth path of its leading player, offering actionable insights for brands planning their digital transformation and aligning with industry direction. Key takeaways are as follows:

1. Industrial and consumer technology trends: Competition in the AI industry has shifted from competing on standalone model capabilities to competing on comprehensive system capabilities spanning underlying computing power, data scheduling and cloud platform infrastructure. In the future, the primary users of cloud services will be AI agents, and interaction logic will shift from click-based operations to instruction-based invocation. Brands need to prepare for this shift in advance.

2. Insights for product R&D and technology layout: Alibaba Cloud's growth proves that long-term, committed investment in core underlying technologies is the only way to build sustainable competitive moats. Leveraging its diverse business scenarios such as e-commerce and retail, Alibaba has built a closed R&D and commercial iteration loop. Brands can follow this example to refine their AI capabilities based on their own business scenarios.

3. Current commercialization status: AI cloud commercialization has now entered the phase of mass implementation. AI-related products already account for over 30% of Alibaba Cloud's external revenue, and its annualized Model-as-a-Service (MaaS) revenue exceeds 8 billion yuan, on track to exceed 30 billion yuan by the end of the year. This confirms that AI-enabled business has entered a period of tangible returns.

This article summarizes the latest developments in the AI cloud industry, providing reference for sellers to identify new growth opportunities and select technology service providers. Key takeaways are as follows:

1. New opportunity outlook: AI has now entered the phase of agent implementation, which is reshaping the access logic of cloud services. Mature AI cloud services help sellers cut costs and improve efficiency. By leveraging the standardized AI capabilities of leading platforms like Alibaba Cloud, sellers can quickly build custom AI applications for operations, design, and customer service to capture AI-driven growth without heavy upfront technology investment.

2. Industry shifts: The stability and cost of underlying computing power services have improved continuously, and now can support all types of AI-enabled business innovation for sellers. Competition has shifted to comprehensive system capabilities, and partnering with a mature platform delivers far better cost efficiency than independent R&D for small and medium-sized sellers.

3. Risk warning: The AI sector is seeing rapid technological iteration and intensifying competition among leading players. When choosing a technology service provider, sellers should prioritize platforms with deep technical accumulation, fast iteration speeds, and market-proven stability. With 17 years of industry experience, Alibaba Cloud has maintained a pace of launching three new large model versions within three months. Its technical stability has been validated by years of high-traffic events such as Singles' Day, making it a more reliable option.

This article offers insights for factories advancing digital transformation and capturing business opportunities in the AI era. Key takeaways are as follows:

1. Lessons for digital transformation: Early Chinese internet companies universally relied on the overseas IBM-Oracle-EMC (IOE) technology architecture. Alibaba Cloud built its underlying cloud computing technology from scratch and eventually replaced foreign solutions. This shows that factories can complete digital upgrades through self-developed routes and domestic technology services, while also reducing long-term operating costs.

2. Opportunities for business and efficiency improvement: Large AI models have now entered the implementation phase. Products like Alibaba Cloud's Qwen Cloud come with pre-packaged standardized AI capabilities. Factories can use these capabilities to quickly build AI applications for scenarios such as product design, production process optimization, and supply chain management, effectively boosting production and design efficiency without building a full in-house R&D team.

3. Recommended transformation roadmap: Alibaba Cloud started by solving its own business pain points, refined its technology gradually, and then expanded its services to external clients. Factories can follow this approach for digital transformation: start with your most pressing production and management pain points, roll out changes step by step, and avoid blindly pursuing full end-to-end digital transformation in one go, which reduces the risk of transformation failure.

This article summarizes the latest development trends and technology movements of the AI cloud industry, offering high reference value for technology service providers to align with industry direction. Key takeaways are as follows:

1. Industry development trends: Competition in the AI industry has shifted from competing on model parameter size and conversational capabilities to competing on the full set of system capabilities spanning underlying computing power and cloud platform infrastructure. There is growing market demand for systems that can stably, efficiently, and cost-effectively support complex intelligent tasks. In the future, the primary users of cloud services will shift from humans to AI agents, and access logic will shift from click-based operations to instruction-based invocation, which will be the core R&D direction for service providers going forward.

2. Latest technology developments: Leading industry players have started laying out next-generation cloud services. Alibaba Cloud has completed a full-stack Agent-oriented upgrade and launched its new core product Qwen Cloud, which features an agent-readable instruction system, a pre-packaged standardized skill library and toolset, and supports autonomous capability invocation by AI agents. Its latest Qwen3.7-Max large model now ranks first among domestic Chinese models, and the company has significantly accelerated its model update pace.

3. Customer pain points and solution directions: Customers are no longer satisfied with standalone model capabilities, and now require complete end-to-end implementable solutions. Service providers can follow Alibaba's example to build a complete system ranging from underlying infrastructure to upper-layer applications, and form a closed iteration loop tied to implementation scenarios to meet customer demands.

This article summarizes the development path and latest strategy of the leading cloud platform, offering reference for platform operators to align with industry direction and mitigate risks. Key takeaways are as follows:

1. Current core market demand for platforms: The development of large AI models has entered a new stage, and market demand for platforms has shifted from "whether they can provide computing power" to "whether they can deliver a complete set of system capabilities that are easy to use, low-cost, and support the implementation of complex AI tasks". Standalone technological leadership no longer meets market demand.

2. replicable best practices from leading platforms: Alibaba Cloud has sustained over a decade of consistent investment in underlying technology, and completed system-level reconstruction early on, laying a solid foundation for its AI-era transformation. It recently launched Qwen Cloud, built specifically for the agent era, which redefines the interaction and invocation logic of cloud services. Leveraging its own diverse ecosystem scenarios, it has built a closed technical iteration loop, continues to ramp up investment in self-developed chips and infrastructure, and has established long-term competitive advantages.

3. Trend outlook and risk warning: Sector competition has now entered a phase of intense direct confrontation, with all leading vendors ramping up investment. Platforms should avoid the mistake of only competing on standalone model capabilities. Instead, they should focus on building a complete technology system, commit to long-term investment, and avoid pursuing short-term speed, to gain an edge in long-term competition.

This article sorts out the development history of China's cloud computing industry and the latest trends in the AI era, offering high data value for industry research. Key takeaways are as follows:

1. Latest industry developments: Competition in China's AI infrastructure market has entered a new stage. In 2025, China's public cloud market for large model training and inference reached 7.938 billion yuan. Alibaba Cloud ranks first with a 42.2% market share, representing 116% year-on-year growth. Competition has shifted from standalone model capabilities to comprehensive system capabilities of the underlying cloud platform, and commercialization is accelerating. AI-related revenue already accounts for over 30% of Alibaba Cloud's external revenue, and its annualized MaaS revenue exceeds 8 billion yuan.

2. New characteristics of industrial development: Large AI models have now entered the agent development phase, which has reshaped the access logic of cloud services. In the future, the primary users of cloud services will shift from humans to AI agents, and interaction methods will shift from click operations to instruction invocation. The IOE architecture universally adopted by China's early internet industry has now been replaced by self-developed distributed cloud computing architecture, which has become industry consensus.

3. Key research directions: The divergent judgments on cloud computing made by early industry leaders, the impact of long-term investment on competitive landscape, Alibaba's "underlying technology + business scenario" closed iteration business model, and the direction of cloud architecture reconstruction in the agent era are all new topics worthy of in-depth research.

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.

马云:其实正因为我不懂技术,但我们尊重技术,我们公司的技术才最好。

出品 | 电商派Pro作者 | 千帆

在2014年的第五届阿里技术论坛上,马云曾说:“我坚持一点,未来世界的竞争是数据的竞争,上面是云,下面是端,我们知道移动互联网起来了,要出现App加云,这是未来的趋势。”

在当时,这样的判断在很多人看来仍然偏超前。它更像是一个技术圈里的设想,复杂、遥远,甚至带着一点“不现实”的意味。

而且围绕云计算的争论一直没有停过:算力能不能整合?系统能不能稳定?这样的架构到底有没有商业价值?

但马云给出了一个很直接的态度:“如果说我们拥有这个能解决社会的问题,那当然应该做下去。所以,想也没想,从预算、人头、资金,我们一路投,最后我们走了出来。”

没有退路,也没有摇摆。从预算到人力,再到长期投入,阿里选择了一条最难走的路,把争议留在当下,把时间押向未来。

多年之后,这条路迎来了结果验证。

近日,全球知名咨询机构IDC最新发布了《中国AI软件市场半年度追踪,2025H2》。

数据显示,在中国大模型训推公有云市场中,2025年规模达到79.38亿元,阿里云以42.2%的市场份额位列第一,同比增长116%。

直白地讲,这个数字的意义,不只是一次排名变化,更像是中国AI基础设施竞争进入新阶段的信号。竞争焦点已经从单点模型能力,转向底层算力、数据调度与云平台综合能力的体系化比拼。

换句话说,AI行业正在从“模型比谁更聪明”,转向“系统比谁更能稳定、高效、低成本地运行复杂智能任务”。

而阿里云之所以站在前排,并不是偶然发生的结果,而是源于一场关于“未来”的豪赌。

2008年,随着淘宝与支付宝业务规模快速增长,系统在流量高峰期频繁承压,传统IOE架构逐渐难以支撑业务扩张需求。

彼时,中国互联网公司普遍使用IOE作为内部系统解决方案,这三个大写字母分别代表IBM服务器、Oracle数据库与EMC存储系统。

随着业务复杂度提升,阿里开始面对一个无法回避的问题:必须重建底层计算体系。

同年,阿里正式确立云计算与大数据战略,启动自研分布式系统“飞天”,试图从底层重构计算能力。

2009年,阿里云正式成立,云计算业务开始从技术实验走向组织化推进。

在最初的几年,这条路走得异常艰难。“飞天”系统早期稳定性不足,资源调度效率低,甚至出现“原本一台服务器可以解决的问题,在云上需要几十台资源才能跑通”的情况。

与此同时,外界对云计算的价值也普遍存在疑问。

2010年的一次IT领袖峰会上,百度李彦宏直言“云计算这个东西,不客气地讲,是新瓶装旧酒”,腾讯马化腾也认为“为时过早”。

但外界的质疑并未改变阿里云的投入节奏。马云曾在内部强调:“每年投入10亿元,坚持投入十年,做不出来再说。”可以说,这种投入方式,本质上是在用时间换确定性。

所幸,功夫不负有心人。

2013年,阿里云成功突破“5K”集群技术壁垒,成为中国第一个独立研发拥有大规模通用计算平台的公司,也是世界上第一个对外提供5K云计算服务能力的公司。

所谓“5K”,指的是可以统一调度5000台服务器的能力。对于当时的阿里内部系统而言,也只有跨过这一门槛,“飞天”才真正具备承载核心业务的基础能力。

同年,余额宝全部核心系统迁移至阿里云,证明了其金融级稳定性与可靠性。

此后,阿里云进入快速扩张阶段。2014年,阿里云香港节点开服,成为国内首家提供海外云计算服务的公司;2015年,成功支撑“双11”912亿元交易峰值;2017年,年度收入突破百亿元,跻身全球第三大公有云服务商。

俗话说,三十年河东,三十年河西。市场格局也在同步发生变化。

腾讯在2010年内部成立云平台部,启动云计算立项研发;百度也逐步转变对云计算的认知并加大云与AI投入。各大互联网企业纷纷入场布局云计算,行业规模不断扩大,也从侧面印证了马云当年关于“云+端”趋势判断的前瞻性。

从结果来看,云计算不再是“是否需要”的问题,而是“必须拥有”的能力。

在这一阶段,阿里云的优势逐渐显现:它不是后来者,而是最早完成系统级重构的一批企业。也正因如此,它在AI时代的转型具备更强的延续性。

如果说过去十年阿里云解决的是“算力有无”的问题,那么如今在AI大模型时代,它正致力于解决“算力好用”与“价值落地”的问题。

大模型的演进正在从单纯的对话能力,走向能够执行任务、调用工具、完成流程的智能体(Agent)阶段。

在2026阿里云峰会上,阿里云正式宣布完成全栈Agent化升级,同时发布新一代AI产品体系,包括“千问云”平台、自研AI芯片驱动的真武M890超节点服务器,以及最新旗舰模型Qwen3.7-Max。

值得一提的是,Qwen3.7-Max作为最新旗舰大模型,在多项国际评测中表现突出。

在三方机构Arena全球大模型盲测总榜中,Qwen3.7-Max超过Kimi-K2.6、DeepSeek-v4-pro、GLM-5.1,其综合能力已接近GPT、Claude与Gemini等最强模型,并在国产模型中位列第一。

另外,在短短3个月内,千问旗舰模型已连续完成3.5、3.6、3.7三个版本迭代,更新节奏明显加快,保持着极高的发布节奏。

更重要的是,全新上线的“千问云”,是阿里云成立17年来重磅推出的全新核心产品,也是适配AI智能体时代的产品。

不同于传统人机交互官网,千问云打造出全新的Agent可读指令体系,将平台所有模型服务、核心技术能力,封装为标准化的Skills技能库与CLI工具集。

AI智能体可直接解析平台指令,自主学习官网全部能力,并根据用户实际需求,自动匹配、调用对应功能,实现全流程自主作业。

这意味着一个重要转折:未来云服务的主要使用者,可能不再是人,而是Agent。当机器成为主要调用主体,传统以人为中心设计的界面体系、操作路径与交互逻辑,都将被重新定义。

在这种结构下,一行指令可能就替代一个完整产品入口,系统调用方式正在从“点击式操作”转向“指令式理解”。这也标志着云计算入口逻辑的进一步演化:从PC时代的网页,到移动互联网的App,再到AI时代的Agent调用接口。

在近期VivaTech 2026峰会上,阿里巴巴集团主席蔡崇信进一步强调,全栈AI是公司面向未来的核心战略方向。

他提到,在基础设施层与模型层,阿里较早布局云计算与相关技术体系,并推出开源模型千问(Qwen),在全球开发者社区中具有较高活跃度。

在应用层,则依托电商、即时零售、地图、酒旅等多元生态场景,为AI提供落地空间。

这种结构的优势在于,既有底层算力与模型能力,又具备真实业务场景,可以形成持续反馈与迭代闭环。

从财报来看,截至2026年3月,阿里云AI相关产品收入已占外部收入30%以上。MaaS与应用年化经常性收入超过80亿元,并管理层预计年底将突破300亿元。

此外,蔡崇信与阿里CEO吴泳铭在致股东信中进一步明确表示将持续加码AI基础设施与自研芯片投入,推动“AI+云”成为新的增长引擎。

不过,赛道竞争也在同步加剧。火山引擎、百度、腾讯等头部厂商在MaaS服务、大模型解决方案领域持续加码,算力、模型与行业方案的边界也在不断被打通,竞争进入更密集的正面交锋阶段。

今天的AI行业,已经很难用“谁领先一步”来简单概括。表面看是模型在更新,底层其实是整套基础设施在重建。云、算力、数据、工具链正在重新组合,拼的也不再是单点能力,而是能不能形成一套真正可持续运行的系统。

正如马云所说,“阿里云就是未来方向。”

说到底,这一轮竞争,比拼的从来不是速度,而是能不能走得足够久。

注:文/千帆,文章来源:电商报(公众号ID:kandianshang),本文为作者独立观点,不代表亿邦动力立场。

文章来源:电商报

广告
微信
朋友圈

这么好看,分享一下?

朋友圈 分享

APP内打开

+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