广告
加载中

梁文锋招到CFO

冯雨晨 2026-09-16 07:01
冯雨晨 2026/09/16 07:01

邦小白快读

EN
全文速览

本文核心披露了国内头部大模型独角兽DeepSeek的核心人事变动、经营数据与资本化进展,是普通读者了解国内AI行业头部企业动态的实用参考信息。

1.核心人事信息:DeepSeek目前拟聘任的CFO人选为高瓴创投90后合伙人严文韬,其毕业于复旦大学,曾在腾讯投资、H Capital任职,主导参与过MiniMax、智谱AI、字节跳动等科技企业的投资,2025年入选投资界F40中国青年投资人,此前还有多位头部VC的年轻合伙人参与该岗位竞聘。

2.核心经营数据:今年前七个月DeepSeek营收约4.75亿元,较2025年全年增长近10倍,综合毛利率44.6%,其中API业务毛利率高达82.9%,净亏损收窄至7.15亿元;同期AI基础设施投入约110亿元,较2025年全年暴涨近10倍。

3.资本化进展:目前公司二轮融资接近尾声,估值约5000亿元,已启动科创板IPO筹备,被市场给予万亿市值预期。

本文披露的头部大模型企业DeepSeek的文化建设、业务结构与行业发展动向,可为科技领域品牌的建设与布局提供务实参考。

1.品牌人才建设参考:DeepSeek成立初期靠“怀抱雄心、保持真诚”的理想主义价值主张吸引大量年轻人才,首轮融资后即宣布全部门规模扩编一倍,通过清晰的价值主张与快速成长空间搭建团队,支撑业务快速扩张。

2.品牌价值锚定参考:DeepSeek聚焦高毛利技术服务赛道,当前API业务毛利率高达82.9%,整体综合毛利率达44.6%,高毛利的核心业务是其支撑高估值、构建长期品牌竞争力的核心基础,这一思路可作为科技品牌业务布局的参考。

3.行业趋势参考:当前上交所已出台政策支持优质未盈利大模型企业登陆科创板,行业资本化窗口已经打开,头部大模型品牌正密集冲刺IPO抢占用户与市场心智,科技品牌可把握政策窗口期推进资本化布局。

本文释放的大模型赛道政策导向、市场需求变化与企业布局信号,可为赛道相关从业者识别机会、规避风险提供实用参考。

1.政策与市场机会:上交所已明确支持尚未形成稳定收入规模的优质大模型企业登陆科创板,当前DeepSeek、智谱、MiniMax等头部企业均在推进IPO进程,行业正处于快速扩张期,AI算力配套、模型API服务等相关领域的市场需求持续上涨。

2.合作机会提示:DeepSeek二轮融资引入了石溪资本、中芯聚源等产业投资机构,产业链上下游的卖家可通过产业生态对接、资本合作等方式,切入头部大模型企业的供应链体系,比如对接其百亿级规模的AI基建采购需求。

3.风险提示:大模型行业存在较高的资金与技术门槛,即便头部如DeepSeek前七个月仍有7.15亿元净亏损,相关从业者需结合自身资源禀赋评估入局风险,避免盲目投入。

本文披露的国内大模型行业投入动向、头部企业经营思路,可为生产制造类工厂挖掘新商业机会、推进数字化智能化转型提供方向参考。

1.明确的新增商业机会:当前大模型行业正大幅加码AI基础设施建设,仅DeepSeek今年前七个月就投入110亿元用于租用AI芯片服务器、采购计算设备,阿里、智谱等企业也均以百亿级规模投入算力建设,AI服务器、算力相关配套硬件的生产采购需求正快速爆发。

2.产品升级思路参考:DeepSeek创始人早年即确立了聚焦高毛利业务的发展思路,当前其高毛利的API业务已成为重要营收支撑,工厂在调整产品结构时,可优先布局技术含量高、毛利空间充足的赛道,提升自身盈利水平。

3.转型参考:头部科技企业的发展离不开核心人才支撑与长期技术投入,工厂推进数字化、智能化改造时,也需匹配专业团队与持续的资源投入,才能真正释放技术升级的价值。

本文披露的大模型行业发展阶段、头部企业经营动作与核心需求,可为To B类科技服务商判断赛道趋势、打磨服务方案提供参考。

1.行业发展趋势:国内大模型行业已从早期技术探索阶段进入融资扩张、冲刺IPO的资本化关键期,上交所针对优质未盈利大模型企业开通科创板上市通道,头部企业估值普遍达到数千亿级别,全行业都在加大算力与研发投入,相关专业服务需求快速上涨。

2.客户核心痛点:当前大模型企业普遍处于高投入、亏损逐步收窄的阶段,一方面存在海量的算力基础设施搭建、运维需求,另一方面在融资对接、财务体系搭建、IPO合规辅导等方面存在迫切的专业服务缺口,比如DeepSeek目前正引入专业CFO搭建资本运作团队。

3.服务方向提示:财务咨询、资本运作服务、算力运维、产业资源对接类的服务商,可围绕大模型企业的上市筹备、算力建设两大核心需求打磨针对性解决方案,对接头部企业的服务采购需求。

本文披露的大模型赛道政策导向、企业发展动态与核心需求,可为科创服务平台、产业对接平台优化运营、调整招商方向提供参考。

1.招商方向参考:AI大模型是当前政策明确支持的核心科创赛道,上交所已支持未盈利的优质大模型企业登陆科创板,头部大模型企业估值达5000亿级别,且正处于团队扩张、基建投入的快速增长期,平台可将优质大模型企业、AI算力产业链企业作为核心招商标的,匹配对应扶持资源。

2.平台服务优化方向:当前大模型企业普遍存在融资对接、产业资源衔接、IPO合规辅导、算力资源对接的共性需求,平台可针对性搭建资本对接、产业协同、合规服务等功能模块,提升对科技类企业的服务能力与平台粘性。

3.运营风险提示:大模型行业属于高投入、长回报周期的赛道,即便头部企业仍存在数十亿规模的年度亏损,平台在引入相关企业、匹配扶持资源时,需做好企业资质与经营能力的分层评估,避免出现资源错配。

本文披露的DeepSeek核心人事安排、经营数据与资本化进展,为研究国内大模型产业发展阶段、商业模式演化、政策落地效果提供了鲜活的行业样本。

1.产业发展新动向:国内头部大模型企业已进入资本化冲刺的关键阶段,DeepSeek、MiniMax、智谱等头部企业纷纷筹备科创板上市,二轮融资普遍引入产业类投资方;行业人才流动呈现明显的“投而优则企”特征,头部VC的年轻投资人纷纷加入创业公司担任资本运作核心岗位,高瓴等机构已成为大模型企业核心财务人才的重要供给方。

2.商业模式演化特征:头部大模型企业正实现营收的十倍级快速增长,亏损逐步收窄,其中API类业务毛利率超80%,但AI基建投入同样呈十倍级增长,高毛利、高投入、快增长是当前阶段大模型企业的典型经营特征。

3.政策效果参考:上交所出台的未盈利AI企业上市支持政策已实质带动头部企业启动IPO筹备,科创板已成为国内大模型企业资本化的核心渠道,产业政策对新兴赛道发展的引导作用正在持续显现。

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 provides key updates on leading Chinese large language model (LLM) unicorn DeepSeek, covering its core leadership changes, operating metrics and capitalization progress, serving as a practical reference for general readers to track the dynamics of top players in China’s AI industry.

1. Core leadership update: DeepSeek is poised to appoint Yan Wentao, a 1990s-born partner at Hillhouse Venture Capital, as its CFO. A Fudan University graduate who previously held roles at Tencent Investments and H Capital, Yan led or participated in investments in technology companies including MiniMax, Zhipu AI and ByteDance, and was named to the 2025 F40 China Young Investors List. Several other young partners from top-tier venture capital firms also competed for the position.

2. Key operating metrics: In the first seven months of this year, DeepSeek posted revenue of approximately RMB 475 million, a nearly 10-fold increase from its full-year 2025 figure, with a combined gross margin of 44.6%. Its API business delivered an especially high gross margin of 82.9%, while net loss narrowed to RMB 715 million. Over the same period, its investment in AI infrastructure reached roughly RMB 11 billion, also surging nearly 10 times from full-year 2025 levels.

3. Capitalization progress: The company is closing in on the final stage of its Series B funding round at a valuation of around RMB 500 billion, and has launched preparations for a STAR Market IPO, with market expectations pointing to a potential RMB 1 trillion market capitalization.

This article’s findings on leading LLM firm DeepSeek’s culture building, business structure and industry development trends offer practical takeaways for technology brands seeking to refine their brand building and strategic positioning.

1. Talent and culture building lessons: In its early days, DeepSeek attracted large numbers of young talent with its idealistic value proposition of “ambition paired with sincerity”. Shortly after its first funding round, the company announced plans to double headcount across all departments, building its team through a clear value proposition and clear pathways for rapid growth to support fast business expansion.

2. Brand value anchoring reference: DeepSeek has prioritized the high-margin technology services segment: its API business currently records an 82.9% gross margin, lifting the company’s overall gross margin to 44.6%. A high-margin core business forms the foundational support for its high valuation and long-term brand competitiveness, a strategy that offers useful guidance for technology brands mapping out their business portfolios.

3. Industry trend insights: The Shanghai Stock Exchange has rolled out policies to support qualified unprofitable LLM companies in listing on the STAR Market, opening a new capitalization window for the sector. Leading LLM brands are now racing to launch IPOs to capture user mindshare and market position, and technology brands can leverage this policy window to advance their own capital market plans.

Signals in this article regarding policy direction, shifting market demand and strategic moves by leading players in the LLM sector offer practical guidance for relevant industry practitioners to identify opportunities and mitigate risks.

1. Policy and market opportunities: The Shanghai Stock Exchange has explicitly stated support for high-quality LLM companies without stable large-scale revenue to list on the STAR Market. Leading players including DeepSeek, Zhipu AI and MiniMax are all advancing IPO plans, and the sector is in a phase of rapid expansion, with rising market demand for supporting segments such as AI computing power infrastructure and model API services.

2. Partnership opportunity highlights: DeepSeek’s Series B round includes strategic investments from industry-focused institutions such as Shixi Capital and SMIC Juyuan. Sellers across the industrial value chain can access the supply chains of leading LLM firms via industry ecosystem collaboration and capital partnerships, for example by tapping into DeepSeek’s RMB 10 billion-scale AI infrastructure procurement demand.

3. Risk warnings: The LLM sector has high capital and technology barriers. Even an industry leader like DeepSeek recorded a net loss of RMB 715 million in the first seven months of the year. Practitioners should assess entry risks against their own resource endowments to avoid blind investment.

Updates in this article on investment trends in China’s LLM industry and the operational strategies of leading players provide directional reference for manufacturing factories seeking new business opportunities and advancing digital and intelligent transformation.

1. Clear emerging business opportunities: The LLM sector is scaling up investment in AI infrastructure at a rapid pace: DeepSeek alone invested RMB 11 billion in the first seven months of this year to rent AI chip servers and procure computing equipment, while companies such as Alibaba and Zhipu AI have also made RMB 10 billion-level investments in computing power buildout. Procurement demand for AI servers and related computing hardware is seeing explosive growth.

2. Product upgrade insights: DeepSeek’s founder established a strategy of focusing on high-margin businesses early in the company’s development, and its high-margin API business has now become a core revenue driver. When adjusting product portfolios, factories can prioritize segments with higher technological content and sufficient margin headroom to improve profitability.

3. Transformation takeaways: The growth of leading technology companies relies on core talent and long-term technology investment. When pursuing digital and intelligent upgrades, factories must also match these efforts with specialized teams and sustained resource investment to fully unlock the value of technology upgrades.

Insights in this article on the development stage of the LLM industry, operational moves and core demands of leading players offer reference for B2B technology service providers to assess sector trends and refine their service offerings.

1. Industry development trends: China’s LLM sector has moved beyond early-stage technology exploration into a critical capitalization phase marked by financing expansion and IPO pushes. The Shanghai Stock Exchange has opened a dedicated STAR Market listing channel for high-quality unprofitable LLM firms; leading players are generally valued at hundreds of billions of RMB, and the entire industry is scaling up investment in computing power and R&D, driving rapid growth in demand for related professional services.

2. Core client pain points: LLM companies are currently in a phase of heavy investment with gradually narrowing losses. On one hand, they have massive demand for computing infrastructure deployment and operation and maintenance; on the other, there are urgent unmet needs for professional services covering financing matchmaking, financial system building and IPO compliance advisory — as illustrated by DeepSeek’s ongoing recruitment of a professional CFO to build its capital markets team.

3. Service direction guidance: Service providers specializing in financial consulting, capital market advisory, computing power operation and maintenance, and industry resource matching can develop targeted solutions focused on the two core needs of LLM firms — IPO preparation and computing power buildout — to tap into procurement demand from leading industry players.

Updates in this article on policy direction, company developments and core demands in the LLM sector provide reference for sci-tech innovation service platforms and industry matchmaking platforms to optimize operations and adjust investment attraction priorities.

1. Investment attraction guidance: AI large models are a core sci-tech track with clear policy support: the Shanghai Stock Exchange has opened access to the STAR Market for high-quality unprofitable LLM firms, leading LLM players are valued at around RMB 500 billion, and the sector is in a rapid growth phase marked by team expansion and large-scale infrastructure investment. Platforms can prioritize high-quality LLM companies and AI computing value chain enterprises as core investment attraction targets, and match them with tailored support resources.

2. Platform service optimization direction: LLM companies share common demand for financing matchmaking, industry resource connection, IPO compliance advisory and computing power resource coordination. Platforms can develop targeted functional modules covering capital matchmaking, industrial collaboration and compliance services to improve service capacity for tech enterprises and boost platform stickiness.

3. Operational risk warnings: The LLM sector requires high upfront investment and has long return cycles; even leading players record annual losses in the billions of RMB. When onboarding relevant enterprises and allocating support resources, platforms should conduct tiered assessments of enterprise qualifications and operational capacity to avoid resource misallocation.

The details in this article on DeepSeek’s core personnel arrangements, operating data and capitalization progress provide a timely industry sample for studying the development stage, business model evolution and policy implementation effects of China’s LLM industry.

1. New industry development trends: Leading domestic LLM firms have entered a critical phase of pushing for capital market entry, with players including DeepSeek, MiniMax and Zhipu AI all preparing for STAR Market listings, and most bringing in industrial strategic investors for their Series B rounds. Talent mobility in the sector is marked by a clear “investor-to-entrepreneur” trend: young partners from top VC firms are taking core capital markets roles at startups, with institutions such as Hillhouse becoming key talent sources for senior financial positions at LLM companies.

2. Business model evolution characteristics: Leading LLM firms are delivering ten-fold revenue growth with gradually narrowing losses. While API businesses report gross margins above 80%, investment in AI infrastructure is also growing at a ten-fold rate. High margins, heavy capital expenditure and rapid growth are the defining operational features of LLM companies at the current stage.

3. Policy effect observations: The Shanghai Stock Exchange’s support policy for unprofitable AI enterprise listings has effectively driven leading players to launch IPO preparations. The STAR Market has become the core capitalization channel for domestic LLM firms, demonstrating the continuing guiding role of industrial policy in the development of emerging technology tracks.

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.

DeepSeek的CFO快定了。”

早在一个月前,圈内开始流传DeepSeek新CFO来自高瓴,“大概率是来自高瓴创投90后合伙人严文韬”。据我们多方了解,该岗位目前虽还未完全到位,但圈内不少人已经默认此事。

此外,近两个月内,还有多位头部VC的年轻合伙人也参与了DeepSeek的CFO岗位竞聘。

回到三年前,幻方量化决定进军大模型,官宣海报上那句话诱人且热烈:务必要疯狂地怀抱雄心,且还要疯狂地真诚。

理想主义如此不由分说,吸引许多年轻人排队加入DeepSeek。今年6月,刚完成首轮融资的DeepSeek宣布将所有部门规模扩大至少一倍,招聘启事中就包括CFO这一重要职位。

外界笃信,DeepSeek离IPO已经不远。这家被赋予了万亿市值想象的公司,CFO何许人也,同样万众瞩目。

梁文锋选CFO,高瓴90后浮出水面

即便估值高达5000亿元,DeepSeek依然算作一家神秘的公司。

比如,当我们向创投圈人士打听“DeepSeek的CFO是否已经到位”这件事时,大抵会从对方那里捕捉到一丝“此番天上事,我等何以近身窥探”的不可及意味。

其实早在2025年2月,DeepSeek就开始公开招聘CFO。直到今年6月,DeepSeek仍在官网发布财务团队招聘,岗位包括CFO。

今天这个时间,正值DeepSeek密集融资之后——早前,DeepSeek完成首轮融资,融资金额共计约500亿元,梁文锋个人出资约200亿元;现在,DeepSeek传出二轮融资近尾声,估值约为5000亿元。

众所周知,CFO是一家公司在融资与资本运作阶段的重要角色。对比国内大模型同行,DeepSeek的这一角色有些姗姗来迟。

譬如,高瓴第一个接触MiniMax的投资人薛子钊在2023年就加入了MiniMax,担任投融资副总裁兼联席公司秘书,直接操盘这家公司资本市场事宜;

再如前金沙江创投管理合伙人张予彤,2024年4月加入月之暗面后担任总裁,负责包括融资在内的整体战略与商业化。

“投而优则创”,是创投圈一种心照不宣的现象——昔日投资人加入被投创业公司,角色定位往往为公司的资本运作操盘手。MiniMax薛子钊、蜜雪冰城张渊等均循此路径,有意思的是,这两位都来自堪称“年轻CFO黄埔军校”的高瓴。

本次DeepSeek CFO人选,亦是来自高瓴的年轻合伙人严文韬。资料显示,严文韬出生于1991年,毕业于复旦大学,2013年至2020年间先后在腾讯投资、H Capital任职,于2020年加入高瓴创投并成为其90后投资人代表。悉数过往,严文韬曾参与了MiniMax、智谱AI、字节跳动、极兔、WeBull等一批前沿科技企业投资项目。2025年,严文韬入选投资界「F40中国青年投资人」。

“此前也有其他人也在面试这个岗位”,9月初,一位头部机构人士告诉我们。据了解,几家头部机构的年轻合伙人也参与了DeepSeek的CFO面试。

在圈内不少人看来,DeepSeek的CFO似乎是一件比机构合伙人职位更具吸引力的事情。

无论如何,一家估值如此之高的独角兽,这一重要角色难以低调,更不会简单——早年,梁文锋在谈及幻方创业时提到用人观,就强调了个人能力,“这不能说是成功的秘密,但这是幻方的文化之一。”

IPO快了,下一个万亿公司

回到2009年,梁文锋曾去浙大校友创办的公司实习,此时还未毕业的他成为新技术部经理,月薪给到一万六。离职时,那位校友给了年轻梁文锋一个建议:去找毛利高的业务。

这句话梁文锋记了很多年,并在后来的幻方和DeepSeek上得以印证。

据The Information报道,DeepSeek今年前七个月综合毛利率为44.6%,其中API(应用程序编程接口)业务的毛利率高达82.9%。与此同时,DeepSeek今年前七个月营收约4.75亿元,较2025年全年增长近10倍,其净亏损收窄至7.15亿元,明显低于2025年全年9.35亿元的亏损规模。

不过,DeepSeek花钱也更凶猛了。

今年前七个月,DeepSeek在AI基础设施上花了约110亿元,涵盖租用AI芯片服务器及采购计算设备等。而2025年全年,这一数字仅为约12亿元。一年多时间,这个方向的支出暴涨近10倍。

但不止DeepSeek,全行业的AI基建支出都在飞涨——最新露出的一笔是智谱,完成约50亿美元的融资,用于下一代GLM模型、完全自训练体系及相关算力基础设施;稍早前的8月23日,阿里巴巴公告配售7.1亿股新股募资800亿港元,净额将100%投向全栈AI能力,加强AI基础设施建设。

行至此时,无论通往AGI的路上DeepSeek会不会缺钱,IPO都是一个顺势而为的选择。

早在6月17日,上交所宣布支持尚未形成一定收入规模的优质人工智能大模型企业在科创板发行上市。机会敞开,已经于港股IPO的智谱、MiniMax均在抓紧推进科创板IPO。

9月消息正式传出,DeepSeek的IPO开始筹备,目标在科创板上市。同时在二轮融资中,DeepSeek引入以产业投资机构为主的新投资方,例如石溪资本、中芯聚源等。

中国大模型酣战已经两三年,放眼目前中国未IPO的大模型企业阵营,梁文锋创立的DeepSeek和杨植麟创立的月之暗面Kimi,是当下估值最高的两家——后者在最新一轮融资中,投前估值亦达500亿美元。

这是两位等待IPO的广东年轻人。也许很快,中国AI版图就会新增万亿市值。

注:文/冯雨晨,文章来源:投资界(公众号ID:pedaily2012),本文为作者独立观点,不代表亿邦动力立场。

文章来源:投资界

广告
微信
朋友圈

FAQ回顾

DeepSeek拟聘任的CFO是谁,有哪些从业背景?

DeepSeek拟聘任的CFO为高瓴创投90后合伙人严文韬。他1991年出生,毕业于复旦大学,曾任职于腾讯投资、H Capital,参与过MiniMax、智谱AI、字节跳动等科技企业投资,2025年入选投资界F40中国青年投资人。

DeepSeek当前的经营与融资进展如何?

DeepSeek目前估值约5000亿元,首轮融资额约500亿元,二轮融资已接近尾声,引入石溪资本、中芯聚源等产业投资方。据The Information报道,其今年前7个月营收约4.75亿元,较2025年全年增长近10倍,净亏损收窄至7.15亿元。

DeepSeek的上市计划是什么?

DeepSeek已启动IPO筹备工作,目标在科创板上市,选聘CFO是其资本运作层面的重要布局。当前上交所已出台政策,支持尚未形成一定收入规模的优质大模型企业登陆科创板。

创投圈“投而优则创”指的是什么现象?

“投而优则创”是创投圈常见现象,指投资人离职加入被投创业公司担任资本运作操盘手。高瓴因输出多位年轻资本人才被称为“年轻CFO黄埔军校”,代表案例包括MiniMax薛子钊、蜜雪冰城张渊等。

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