广告
加载中

张一鸣砸46亿开公司 准备大干一场

李响 2026-07-28 10:35
李响 2026/07/28 10:35

邦小白快读

EN
全文速览

本文核心是披露抖音集团砸46亿在宁夏新设两家全资科技公司,布局自有算力基础设施,为AI电商规模化落地铺路的行业新闻,核心干货信息如下

1. 当前AI已经成为电商行业竞争的核心赛道,阿里、京东、抖音三大头部玩家都下场布局,各自走出了不同路线,抖音选择将豆包大模型和自有电商场景深度结合,走沉浸式AI推荐路线。

2. 目前抖音豆包的AI购物已经落地,实现了从选品、领券到下单的全流程打通,2026年春节后豆包日使用人次超2亿,日均电商交易额约1000万元,AI推荐转化率达到3%,用户已经开始接受AI辅助购物的模式。

3. 未来电商行业的竞争重心会转移到消费决策入口,底层算力是支撑AI服务稳定的基础,提前布局底层算力的玩家会获得长期竞争优势。

本文透露出AI电商行业的最新发展趋势,能给品牌商布局新渠道、调整营销方向提供这些干货参考

1. 用户行为层面已经发生明显变化,越来越多用户愿意用自然语言描述模糊需求,接受AI给出的购买建议,目前AI导购已经达到3%的转化率,说明用户使用AI购物的习惯正在逐步养成,AI会成为新的消费决策入口。

2. 当前头部电商平台已经推出多款AI工具赋能商家,包括AI生成商品图、智能直播间脚本、实时场控、自动剪辑等,品牌商家可以直接使用这些工具,缩短内容制作流程,降低运营成本。

3. 未来提前适配AI导购逻辑,完善商品信息体系,接入头部平台的AI导购入口,能帮助品牌提前抢占新的流量入口,抢先获得用户心智,抓住AI电商带来的增长红利。

本文梳理了AI电商的发展现状与趋势,能给电商卖家提供这些机会参考与风险提示

1. AI电商是当前明确的高增长赛道,头部平台都在大规模投入资源布局,用户使用AI做购物决策的习惯已经初步养成,新的流量分发入口正在形成,卖家可以抓住行业变革的红利,提前布局适配AI渠道。

2. 目前各大平台已经开放AI运营工具给卖家使用,AI生成商品图可以省去传统布景拍摄修图的繁复流程,智能直播工具可以降低直播间运营成本,卖家可以借助这些工具提效降本,获取更多流量。

3. 风险提示:未来AI电商的竞争核心离不开底层算力支撑,平台如果没有充足的自有算力,AI服务的稳定性和成本都会处于劣势,卖家选择布局平台时,可优先关注技术布局到位、算力储备充足的平台,规避服务不稳定带来的流量损失。

本文关于AI电商发展的内容,能给工厂做数字化转型、挖掘商业机会提供这些干货启示

1. AI电商的发展让用户需求捕捉更加精准,用户的模糊购物需求可以通过AI拆解为具体的商品偏好、品类特征,工厂可以依托AI汇总的需求数据,优化产品的生产和设计,让产品更贴合市场真实需求,降低滞销库存的风险。

2. 当前互联网头部平台都在宁夏、内蒙等地区布局算力基地,围绕算力基地会产生大量配套设施、运营服务的需求,位于这些区域周边的工厂,可以结合自身能力挖掘相关配套业务的商业机会,拓展新的增长曲线。

3. AI电商是行业明确的发展方向,工厂在做数字化转型、布局自有直销渠道的时候,需要跟上平台的AI化节奏,提前接入AI工具和AI流量入口,才能抓住新的用户增长机会。

本文披露了AI电商行业的最新发展动向,能给电商相关服务商提供这些干货参考

1. 行业发展趋势方面,当前AI电商已经从概念创新落地进入基础设施竞争阶段,头部平台都在加码自有算力建设,对算力基础设施配套、算力调度服务的需求大幅增长,算力相关服务商可以迎来明确的业务增长机会。

2. 当前行业已经浮现明确的客户痛点,不管是平台还是商家,都面临AI应用落地带来的算力缺口问题,外部采购算力存在成本高、响应延迟高、稳定性差的问题,无法满足AI电商实时推理的需求,存在明确的服务创新缺口。

3. 中小商家有大量细分的AI运营需求,除了现有的AI作图、智能剪辑,还需要AI选品、AI客满、AI供应链规划等细分服务,服务商可以围绕这些痛点开发针对性解决方案,抢占市场。

本文关于抖音布局AI电商算力的内容,能给电商平台的布局和运营提供这些干货参考

1. 当前商家和用户对AI电商都有明确需求,商家需要AI工具降低运营成本,用户需要AI辅助快速做出购物决策,平台需要加快AI与电商全场景的融合,才能满足各方需求,留住用户和商家。

2. 抖音的闭环布局思路值得参考,抖音同时握住了成熟AI产品豆包、完整的电商场景、持续扩张的自有算力三个核心环节,三者互相赋能,产品靠场景迭代,场景靠算力支撑体验,算力靠前两者转化为商业回报,这种模式验证了底层算力投入的长期价值。

3. 风险提示:如果平台只在应用层做AI创新,不提前布局自有算力,会陷入算力受制于人、成本高、体验不稳定的困境,很快会触碰到业务增长的天花板,需要提前布局底层算力基础设施,既可以降低长期运营成本,也能规避未来的增长风险。

本文披露了AI电商领域的最新产业动向,整理出这些值得研究的干货内容

1. 产业新动向方面,AI电商已经从概念落地阶段进入基础设施竞争阶段,行业竞争重心从传统的流量获取、商家资源、物流效率,转移到底层算力支撑和消费决策入口的争夺,抖音砸46亿布局自有算力就是这个动向的典型案例,说明底层基础设施的投入已经成为竞争核心。

2. 当前头部玩家已经摸索出三种不同的AI电商商业模式,分别是阿里大模型打通导购全链路、京东开放AI能力输出到全环节、抖音豆包结合内容场景做沉浸式AI导购,三种路线的底层逻辑差异较大,适合做进一步对比研究。

3. 目前AI导购的商业化已经初步跑通,抖音豆包已经形成免费获客+高阶付费订阅+交易佣金的多元商业化模式,公开数据显示其当前日电商交易额约1000万元,转化率3%,初步验证了AI电商的商业可行性,为行业研究提供了真实样本。

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 reports that Douyin Group has invested 4.6 billion yuan to establish two wholly-owned tech subsidiaries in Ningxia, building self-owned computing power infrastructure to pave the way for the large-scale deployment of AI-powered e-commerce. Key takeaways are as follows:

1. AI has become the core competitive battleground in China’s e-commerce industry, with top three players Alibaba, JD.com and Douyin all pursuing their own unique development paths. Douyin has chosen to deeply integrate its Doubao large language model with its in-house e-commerce scenarios, focusing on an immersive AI recommendation route.

2. Douyin’s Doubao-powered AI shopping is already live, with a full-process end-to-end experience covering product selection, coupon redemption and order placement. As of post-Spring Festival 2026, Doubao records over 200 million daily active users, generates approximately 10 million yuan in daily e-commerce transaction volume, and delivers a 3% conversion rate for AI recommendations, indicating that users have already begun to accept AI-assisted shopping.

3. Going forward, industry competition will increasingly center on consumer decision entry points. Underlying computing power is the foundation for stable AI services, so players that pre-emptively build out in-house computing infrastructure will gain a long-term competitive advantage.

This article outlines the latest development trends in AI-powered e-commerce, providing actionable insights for brands looking to expand into new channels and adjust their marketing strategies:

1. User behavior has shifted noticeably: a growing number of consumers are comfortable describing vague needs in natural language and accepting AI-generated purchase recommendations. The 3% conversion rate already achieved by AI-guided shopping proves that user habits are gradually forming, and AI will become a new core entry point for consumer decision-making.

2. Top e-commerce platforms have already launched a range of AI tools to empower merchants, including AI-generated product images, smart live broadcast scripts, real-time stream management, and automatic content editing. Brands can leverage these tools directly to shorten content production workflows and cut operational costs.

3. Brands that adapt to AI-guided shopping logic, refine their product information systems, and connect to AI entry points on leading platforms early will be able to pre-empt this new traffic channel, capture user mindshare, and capture the growth opportunities brought by AI e-commerce.

This article summarizes the current status and development trends of AI-powered e-commerce, offering opportunity analysis and risk warnings for e-commerce sellers:

1. AI e-commerce is clearly a high-growth sector today, with top platforms investing heavily in the space. As user habits for AI-driven shopping decisions have begun to take shape and a new traffic distribution entry point is emerging, sellers can capture the dividends of this industry transformation by adapting for AI channels early.

2. Major platforms have already opened AI operation tools to sellers. AI-generated product imagery eliminates the complex workflow of traditional set-up, shooting and retouching, while smart live tools cut operational costs for streaming rooms. Sellers can leverage these tools to boost efficiency, cut costs, and capture more traffic.

3. Risk warning: Long-term competition in AI e-commerce relies fundamentally on underlying computing power. If a platform lacks sufficient in-house computing capacity, it will face disadvantages in both AI service stability and cost. When choosing which platforms to expand into, sellers should prioritize platforms with completed technology布局 and sufficient computing reserves to avoid traffic losses caused by unstable service.

This article offers actionable insights for factories pursuing digital transformation and new business opportunities from the development of AI-powered e-commerce:

1. AI e-commerce enables far more accurate capture of user demand: consumers’ vague shopping needs can be broken down by AI into specific product preferences and category features. Factories can use aggregated demand data from AI to optimize product design and production, align offerings with real market demand, and reduce the risk of unsold inventory.

2. Today’s leading internet platforms are building computing hubs in regions such as Ningxia and Inner Mongolia, which will generate strong demand for supporting infrastructure and operational services. Factories located near these regions can leverage their existing capabilities to pursue supporting business opportunities and unlock new growth curves.

3. AI e-commerce is a clear industry development direction. When factories pursue digital transformation and build their own direct-to-consumer sales channels, they must keep pace with platforms’ AI transformation and connect to AI tools and AI traffic entry points early to capture new user growth opportunities.

This article covers the latest industry developments in AI-powered e-commerce, providing key insights for e-commerce-related service providers:

1. In terms of industry trends, AI e-commerce has moved beyond conceptual innovation and entered a phase of infrastructure competition. Leading platforms are ramping up investment in in-house computing power, driving a sharp increase in demand for computing infrastructure supporting services and computing scheduling services. Computing-related service providers are positioned to see clear business growth opportunities.

2. Clear customer pain points have already emerged in the industry: both platforms and merchants face computing capacity gaps brought by AI deployment. Outsourced computing suffers from high costs, high latency, and poor stability that cannot meet the real-time inference requirements of AI e-commerce, creating a clear gap for service innovation.

3. Small and medium-sized merchants have a large number of segmented AI operation needs. Beyond existing AI image generation and smart editing, they require specialized services for AI product selection, AI customer service, AI supply chain planning and more. Service providers can develop targeted solutions to address these pain points and capture market share.

This article, focused on Douyin’s computing power布局 for AI e-commerce, provides key insights for the strategy and operation of e-commerce platforms:

1. Today both merchants and users have clear demand for AI e-commerce: merchants need AI tools to cut operational costs, while users need AI assistance to make faster shopping decisions. Platforms must accelerate the integration of AI across all e-commerce scenarios to meet the needs of all stakeholders and retain both users and merchants.

2. Douyin’s closed-loop strategy offers a useful reference: the company controls three core pillars simultaneously: a mature AI product (Doubao), a full-fledged e-commerce ecosystem, and continuously expanding in-house computing power. The three elements reinforce each other: the AI product iterates via real-world scenario data, scenarios rely on computing power to deliver a smooth user experience, and computing power is monetized via the product and scenario to generate commercial returns. This model validates the long-term value of investing in underlying computing infrastructure.

3. Risk warning: Platforms that only pursue AI innovation at the application layer and fail to pre-emptively build in-house computing power will fall into a predicament of relying on third-party providers, high costs, and inconsistent user experience, and will quickly hit a ceiling on business growth. Pre-emptive investment in underlying computing infrastructure can both reduce long-term operational costs and mitigate future growth risks.

This article shares the latest industry developments in AI-powered e-commerce, and compiles key insights for academic and industry research:

1. In terms of new industry trends, AI e-commerce has transitioned from the conceptual deployment phase to infrastructure competition. The core of industry competition has shifted from traditional advantages including traffic acquisition, merchant resources and logistics efficiency to competition over underlying computing power support and consumer decision entry points. Douyin’s 4.6 billion yuan investment in in-house computing power is a typical example of this trend, confirming that investment in underlying infrastructure has become the core of competition.

2. Top industry players have already developed three distinct AI e-commerce business models: Alibaba’s large language model powers an end-to-end guided shopping experience; JD.com opens up its AI capabilities across all operational links; and Douyin’s Doubao integrates content scenarios to deliver immersive AI-guided shopping. The underlying logics of these three paths differ substantially, making them well-suited for further comparative research.

3. The commercialization of AI-guided shopping has already been preliminarily validated. Douyin’s Doubao has built a diversified commercial model combining free user acquisition, premium paid subscriptions and transaction commissions. Public data shows it currently generates approximately 10 million yuan in daily e-commerce transaction volume with a 3% conversion rate, which preliminarily confirms the commercial viability of AI e-commerce and provides a real-world sample for industry 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.

财大气粗。

出品 | 电商派Pro作者 | 李响

AI席卷全行业的时代,电商早已成为技术厮杀的主战场。

阿里、京东、抖音等头部大厂齐齐下场,疯狂砸钱大搞AI电商,竞争格外激烈。

在一众玩家里,抖音的打法可谓是稳、准、狠,一边凭借豆包这款国民级AI产品,在行业里跑出差异化优势,一边在默默扩容算力,为下一轮爆发悄悄积蓄力量。

近日,抖音集团在宁夏中卫新设两家科技公司——中卫塞上鸣沙科技有限公司、中卫朔方星河科技有限公司,注册资本合计46亿元。

天眼查显示,这两家公司不仅在同一天成立,法定代表人均为田凯,还都由北京字跳网络技术有限公司全资持股,注册地均在中关村科创园。

要知道,此前抖音在内蒙古已注册过四家类似全资子公司,注册资本共45亿元。这四家公司的经营范围普遍包含“非居住房地产租赁”,主要服务于数据中心和算力基地的配套资产运营。

由此看来,这并非简单的工商注册,更像是抖音围绕“东数西算”工程所进行的算力储备工作。

不管怎样,抖音布局算力基础设施的节奏已然非常快。

对抖音来说,豆包大模型正在快速迭代,模型规模越大,训练所需的算力就越庞大。

过去,抖音主要依赖第三方租赁,成本高不说,资源调配的灵活性也受制于人。

这种情况下,遇到算力需求集中爆发,供给紧张几乎不可避免,而自建算力设施相当于把这块资源变成内部可调控的基础能力,稳定性和成本结构都会有明显改善。

更深一层看,这46亿并不仅仅是为了迎合工程发展,目前抖音的电商业务正在全面拥抱AI,而这恰恰需要另一种形态的算力支撑。

现在打开抖音电商,AI生成商品图的工具已经被商家广泛使用。

过去需要布景、拍摄、修图的流程,现在输入需求就能快速产出多个版本。直播间里,智能脚本、实时场控、自动剪辑成为标配,用户端的推荐系统更是如此,每一次刷新、每一次浏览,背后都有大量模型在做实时推理。

未来随着豆包深度介入购物决策,用户随口一句描述,系统就要在毫秒内完成理解、检索、比对、推荐,整个过程对算力的消耗和速度要求,会比现在高出几个数量级。

这些场景和离线训练大模型不同,它们需要的是实时、低延迟、高并发的推理能力。如果这部分算力依赖外部采购,大规模铺开时成本会成倍增长,最终限制应用的边界。

此时,自有算力一旦形成规模,单次推理成本可以持续优化,让高频次的AI交互服务变得经济可行。

从这个角度讲,算力投资正在为AI电商应用的规模化落地铺路。

放眼整个电商行业,不只是抖音在发力。

阿里巴巴将千问大模型与淘宝全面打通,走的是导购路线,让AI贯穿商品发现、比较、下单、售后的完整链路。

京东选择开放输出,把AI能力嵌入营销、客服、物流各环节,同时将京东AI智能体接入华为小艺、OPPO小布、荣耀YOYO等主流手机厂商智能体和腾讯元宝。

抖音则是另一套思路,以豆包和抖音电商的深度结合为基础,围绕内容场景做沉浸式的推荐和决策辅助。

这几条路线谁会更占优势,现在谁也说不准。但有一点挺明显的,就是单纯在应用层面做创新,已经不够看了,模型训练不能停,日常调用又那么高频,将来分发量更大,没点稳定的算力根本扛不住。

如果算力受制于人,功能做得再好,天花板也会很快到来。

抖音在这一轮布局中的特点,是同时握住了三样东西:成熟的AI产品、完整的电商场景、持续扩张的自有算力。

这三者彼此需要。产品需要场景来验证和迭代,场景需要算力来保障体验,算力则需要前两者把投资转化为商业回报。

缺少任何一环,整个结构都会有缺口。

豆包目前在电商方面的进展,可以说明这种结构是如何运转的。

去年双十一期间,豆包开始接入抖音商城的购物链接,成为国内少数打通从推荐到成交的AI应用。

到了今年,豆包在首页导航栏新增了购物助手入口,并上线了帮你选功能。用户可以在豆包内完成选品、领券、下单、查物流,直接支付的功能也正在灰度测试。

“帮你选”和传统搜索电商功能不太一样,用户用自然语言描述需求,系统会搜索信息、提供方案,并拆解每个选项的具体特点。

比如问一句适合送给父母的健康食品,豆包会列出几种品类,分别说明成分、适用人群和注意事项。

这种交互更接近咨询,而不是简单的货架陈列。

据36氪透露,2026年春节过后,豆包每天有超过2亿人次使用,每天产生的电商交易额约1000万元,AI推荐的购物转化率做到了3%。

这个转化数字和成熟电商平台相比不算高,但它说明用户愿意在AI辅助下做购买决策,行为习惯正在形成。

此外,豆包的商业化也初步成型,免费版继续吸引用户,付费订阅提供更高阶服务,六月已经公布了三档订阅方案。

晚点LatePost也披露,豆包目前主要靠佣金获得收入,每天收入不足百万元。

看起来,体量还不大,但随着支付功能上线和与抖音商城联动加深,增长空间是可预见的。

这一切得以运转的前提,是算力足够充沛。

豆包每次做商品推荐,都要实时调用大量信息。用户每多问一句,系统就要重新理解需求、匹配信息、生成回答,当用户规模达到亿级,每天进行哪怕一次购物相关交互,对后台的并发处理能力都是严峻考验。

如果算力跟不上,响应速度下降,推荐质量变差,用户很快会流失。

所以,前面提到的分布在宁夏、内蒙的算力公司,不只是为了训练更聪明的大模型,它们更日常的使命,是保证当无数用户同时打开豆包询问购物建议时,每一秒钟都能给出准确流畅的反馈。

这种稳定和低成本,是AI购物从尝鲜走向常态化的前提。

长期来看,电商行业的竞争重心正在发生转移。过去大家比拼的是流量获取、商家资源、物流效率。

这些当然还很重要,但一个新的维度正在浮现,那就是消费决策的入口。当消费者产生一个模糊的购物需求,第一个问的是哪个应用,哪个应用就可能掌握分发的主导权。

这不是靠营销补贴就能抢下来的位置。它需要用户真正信任这个应用能帮自己做出好选择,也需要这个应用在任何时候都能及时响应,不给用户犹豫和离开的机会。

而要做到后者,算力就是最基础的条件。

抖音在宁夏中卫投下的这46亿,表面上只是成立两家科技公司,本质上是在为AI电商的长周期竞争铺路,当算力供给足够稳定和便宜,抖音就能以更低的成本服务更多电商用户,更频繁地出现在消费决策的初始环节,也更有竞争优势。

虽然这种优势不会立刻体现在某个月的交易数据里,但它会随着时间推移,逐渐拉开与追赶者的距离。

沙漠边缘的服务器集群,距离北上广的电商战场很远。

可它们日夜运转所支撑的每一次推荐、每一次对话、每一次支付,正在悄悄改变电商行业的底层逻辑。

行业的变化人人都看得见,真正决定终局的,往往是这些埋在深处的投入。

注:文/李响,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

FAQ回顾

AI电商企业为什么要布局自有算力?

第三方租赁算力成本高、资源调配灵活性受限,难以应对算力需求集中爆发的情况。自有算力可改善稳定性与成本结构,支撑大模型训练及AI电商实时低延迟高并发推理需求,降低单次推理成本,助力AI交互服务规模化落地。

抖音布局AI电商的核心优势是什么?

抖音同时拥有成熟AI产品豆包、完整电商场景、持续扩张的自有算力三类核心资产,三者互相支撑,可实现从AI推荐到成交的完整链路,适配AI电商长周期竞争的要求。

当前头部电商平台布局AI电商的路线有什么区别?

阿里将千问大模型与淘宝全面打通,走全链路AI导购路线;京东开放输出AI能力嵌入各业务环节,接入多厂商智能生态;抖音以豆包和电商深度结合为基础,围绕内容场景做沉浸式推荐和决策辅助。

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