惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
T
Troy Hunt's Blog
Scott Helme
Scott Helme
T
Threat Research - Cisco Blogs
T
Tenable Blog
L
LINUX DO - 热门话题
V
Visual Studio Blog
I
Intezer
Blog — PlanetScale
Blog — PlanetScale
Cisco Talos Blog
Cisco Talos Blog
A
Arctic Wolf
C
Cyber Attacks, Cyber Crime and Cyber Security
F
Fortinet All Blogs
aimingoo的专栏
aimingoo的专栏
Know Your Adversary
Know Your Adversary
CTFtime.org: upcoming CTF events
CTFtime.org: upcoming CTF events
N
Netflix TechBlog - Medium
SecWiki News
SecWiki News
I
InfoQ
Microsoft Security Blog
Microsoft Security Blog
Project Zero
Project Zero
W
WeLiveSecurity
Microsoft Azure Blog
Microsoft Azure Blog
A
About on SuperTechFans
Recorded Future
Recorded Future
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
Vercel News
Vercel News
S
Securelist
Spread Privacy
Spread Privacy
L
LangChain Blog
云风的 BLOG
云风的 BLOG
G
Google Developers Blog
MongoDB | Blog
MongoDB | Blog
Google DeepMind News
Google DeepMind News
Recent Commits to openclaw:main
Recent Commits to openclaw:main
D
Darknet – Hacking Tools, Hacker News & Cyber Security
C
CERT Recently Published Vulnerability Notes
罗磊的独立博客
K
KPMG report finds enterprise disconnect between AI and its ROI | CIO
The Last Watchdog
The Last Watchdog
Attack and Defense Labs
Attack and Defense Labs
博客园 - 司徒正美
Help Net Security
Help Net Security
L
Lohrmann on Cybersecurity
人人都是产品经理
人人都是产品经理
Forbes - Security
Forbes - Security
Hacker News - Newest:
Hacker News - Newest: "LLM"
PCI Perspectives
PCI Perspectives
博客园 - 【当耐特】
T
Tor Project blog

钛媒体:引领未来商业与生活新知

英特尔发布至强6+,芯片算力在“时间压缩”与“几何微缩”的交汇处-钛媒体官方网站 快撑不住的千亿粤派房企,开始抛售自家酒店了-钛媒体官方网站 公狗剧场≈性转版“崩老头+浪姐”÷2-钛媒体官方网站 宇树上会,机器人会成为半导体下一个超级终端吗?-钛媒体官方网站 玩家深恶痛绝的“广告”,却被厂商卖到了畅销榜TOP4-钛媒体官方网站 一个月内三破世界纪录,光伏龙头全面开启BC竞速赛-钛媒体官方网站 玻璃基板产业化进展到哪了?-钛媒体官方网站 为什么文案策划常被误解?-钛媒体官方网站 GPU抢了风头,西部数据说存储才是AI规模化的真正门槛-钛媒体官方网站 看懂十年两轮锂周期,就学会了投资这个赛道的一半-钛媒体官方网站 神药退潮,超4000亿美元肿瘤新战场谁能称王? 好品牌,开始竞争“互动率”-钛媒体官方网站 孙正义带领软银投资法国核电, 与施耐德这场合作能否突破欧洲算力的困局? 卷向流量的景区们,都在扎堆造“明星”-钛媒体官方网站 ASCO观察:全球首个+1“得福组合”重新定义大单品-钛媒体官方网站 从万播5元到7天充电15万+,AI漫剧开始重做“用户生意”?-钛媒体官方网站 营销失灵,电影们开始“碰运气”了-钛媒体官方网站 Edge AI Daily 早报(6月1日)-钛媒体官方网站 中国商业航天追赶的,从来不只是SpaceX-钛媒体官方网站 Anthropic 冲击 1 万亿:除了 Pre-IPO,还有哪些隐藏的「Claude 概念股」?-钛媒体官方网站 【钛晨报】提升全民人工智能素养,四部门最新部署;MiniMax Group Inc.:拟于科创板上市;国家统计局:5月份制造业采购经理指数(PMI)为50.0%-钛媒体官方网站 赢家亏本转让,输家如愿接盘?山高环能的蹊跷“便宜”不好捡-钛媒体官方网站 阿里"上货",字节"练功"-钛媒体官方网站 260亿美元,“全华班”撑起全球AI编程估值最高公司-钛媒体官方网站 加入“清华圈”,黄仁勋放不下中国-钛媒体官方网站 东方甄选布仓、辛巴开超市:中小玩家分层应战-钛媒体官方网站 “爸爸品牌”,正在集体自救-钛媒体官方网站 618暗战已开,除了低价,大家还在“卷”什么-钛媒体官方网站 监管加码补贴收紧,外卖大战落幕,但消费入口争夺战才刚刚开始-钛媒体官方网站 140万亿Token之后:中国正在修建“算力高铁”-钛媒体官方网站 暴增3100亿!存储巨头大普微,估值泡沫已现-钛媒体官方网站 一季度亏掉23亿,李想重回“苦日子”?-钛媒体官方网站 盈利没保住,小鹏品控又“翻车”-钛媒体官方网站 硅谷大裁员,韩企争红利,揭露了AI带来的“生死问题”-钛媒体官方网站 宁德时代花了30亿,为储能建了一个“风洞”-钛媒体官方网站 万科股东会只剩一个老面孔-钛媒体官方网站 小米AI的"免费获客-黑箱锁死"闭环:从100T Token到Credits陷阱-钛媒体官方网站 大模型的另一种活法,被MiniMax跑通了-钛媒体官方网站 AI写小说的套路被扒光了: Claude爱平铺,GPT总做梦,Gemini只会“他如何如何”-钛媒体官方网站 Edge AI Daily 早报(5月31日)-钛媒体官方网站 纯债基金深度掘金:2026Q1市场洞察与精选策略-钛媒体官方网站 价值判断:涨停板的投资机会和风险提示(5月29日)|证券市场观察-钛媒体官方网站 为什么价值创造,才是市值管理的核心-钛媒体官方网站 3只航空航天类股票已准备好乘上售后市场反弹的东风-钛媒体官方网站 我们看好的戴尔股票涨势惊人,我们会继续坚持持有-钛媒体官方网站 A股连续下跌,市值蒸发近40亿元,君实生物深陷合规漩涡?-钛媒体官方网站 霸王茶姬进韩国,排队188分钟,带来哪些启示? 宁德时代掏30亿建了全球第一的“储能擂台”,储能行业还有什么秘密? 小扎要和老黄“亲儿子”抢饭吃-钛媒体官方网站 Digital Quant 2026 量化交易大赛收官:真实资金、真实数据、真实竞争定义“AI 量化新标准” 中餐出海进阶:狂飙过后,该算账了-钛媒体官方网站 【数智周报】华为发表半导体韬定律,5年内冲刺等效1.4nm制程;MiniMax将A股上市;宇树科技冲刺科创板;Anthropic融资650亿美元,投后估值超OpenAI-钛媒体官方网站 “既要又要”时代,雅迪摩登解锁女性出行最优解-钛媒体官方网站 700亿融资赶紧到位吧,DeepSeek开始限制重生、修改次数了-钛媒体官方网站 铜价奔向10.5万背后: 新能源吃铜,铜也正在改变新能源的胃口 一只“死鸡”,能骗走你多少钱?-钛媒体官方网站 供应商变股东:存储芯片三巨头联手入股Anthropic,AI供应链的权力结构正在重组-钛媒体官方网站 市场不会永远低估腾讯-钛媒体官方网站 分析师观点汇总:英伟达仍是AI核心引擎;沃尔玛等五只股票受关注-钛媒体官方网站 2026上海SNEC前瞻:从“反内卷”到“反谍”,从旧范式到新生态-钛媒体官方网站 段永平,又给老家捐了1万股茅台-钛媒体官方网站 告别“邮政内循环”,中邮人寿迎来首位“外来”掌舵人-钛媒体官方网站 没有下一个泡泡玛特-钛媒体官方网站 AI 编程终于有全局视野了,3 万 Star 项目补齐最大短板-钛媒体官方网站 社区硬折扣超市狂飙的AB面-钛媒体官方网站 3小时卖爆2200万!铜师傅借“修仙”翻红,但离飞升还差几个本命法宝?-钛媒体官方网站 新茶饮“蛇吞象”,柠季洽购哈根达斯?-钛媒体官方网站 一手实测,Opus 4.8 Vs ChatGPT 5.5 Vs Kimi 2.6 ,谁最可用?-钛媒体官方网站 Anthropic估值万亿,但Claude 4.8没那么惊艳-钛媒体官方网站 Claude Opus 4.8:两个0%背后的商业逻辑-钛媒体官方网站 拿下世界杯,小红书里能不能长出另一个足球“社区”?-钛媒体官方网站 一个60亿美妆品牌决定去做药-钛媒体官方网站 全国30%的教室都需要它,江西这个“木匠窝”,何以托举教育事业的过去和未来?-钛媒体官方网站 动漫暑期档:年番保基本盘,新题材赌未来-钛媒体官方网站 新鲜零食融资百亿,零食版“蜜雪冰城”要来了?-钛媒体官方网站 从IP到AI,名创优品「十元店」的翻身仗-钛媒体官方网站 没有大厂总部,萧山凭什么抢下AI制高点?-钛媒体官方网站 《ENEMY》筹备中长剧,影视圈又从短视频「捞人」了-钛媒体官方网站 婴儿期的自变量上门保姆应该“0元购”-钛媒体官方网站 量子计算迎来“晶圆厂时刻”,谁先受益?-钛媒体官方网站 Edge AI Daily 早报(5月30日)-钛媒体官方网站 科创新源2.45亿控股东莞兆科,盈利压力凸显加码散热谋变 | 并购一线-钛媒体官方网站 China's Agricultural Robot Startup is Now Valued at over 500 Mln Yuan in Three Months After Inception-钛媒体官方网站 Hangzhou-based Dexterous Robotic Hand Startup Raises Nearly RMB 1 Billion in Six Months-钛媒体官方网站 减持与开庭赛跑,庄园牧场IPO承诺成“空头支票”,前老板携国资血亏上演罗生门-钛媒体官方网站 绿皮火车,怎么就成了3万亿的好生意?-钛媒体官方网站 尼泊尔的草,中国人的宝-钛媒体官方网站 160亿,深圳“四小龙”跑出首个IPO-钛媒体官方网站 435 万的法拉利Luce,撕掉超豪华遮羞布-钛媒体官方网站 浙江小县城“老头乐”,又要IPO了-钛媒体官方网站 Anthropic发布Claude Opus 4.8,重点是:“我不会骗你”-钛媒体官方网站 华为发布“韬(τ)定律”,重构后摩尔时代的中国技术路径-钛媒体官方网站 法拉利第一台电车,绕开了保时捷和仰望-钛媒体官方网站 被AI替代的人,和没被替代的人,差在哪?-钛媒体官方网站 疯狂的Anthropic-钛媒体官方网站 2026世界杯,为什么小红书买了,抖音没买?-钛媒体官方网站 炒币年赚80亿,以色列Biotech顿悟了-钛媒体官方网站 全球创新药最后一个万亿级未开垦市场-钛媒体官方网站 李开复背叛李开复-钛媒体官方网站 年赚27亿、复合增长62%,卖给美国人的阿麦斯凭什么逆势冲刺“中国糖果第一股”?-钛媒体官方网站
Seven Questions About DeepSeek
Chelsea_Sun2026.06.17 00:26 · 来自海外全文9085字 · 2026-06-17 · via 钛媒体:引领未来商业与生活新知

NextFin News -- China's artificial intelligence startupDeepSeek on Tuesday completed fundraising of more than RMB 50 billion. Along with this round, DeepSeek’s valuation has surpassed USD 50 billion, according to The Information.

It should be noted that, as of now, the above report from overseas media has not been officially confirmed by DeepSeek or the relevant investors. However, based on the developments so far, DeepSeek’s round appears to be close to being finalized.

This round began in April 2026 and has so far lasted two months—also marking the first time since DeepSeek’s founding that it has opened up to external financing. For a long period prior to this, since its establishment, DeepSeek had adhered to principles such as “no fundraising, no IPO, and no commercialization.”

Precisely because of this backdrop, DeepSeek’s round has attracted considerable attention both within China’s AI sector and across the global AI market. And with a large number of investors eager to get in, DeepSeek founder Liang Wenfeng set an extremely high bar for participation, adding another layer of intrigue to the deal.

Below are the seven key questions in this round that deserve the most attention.

1. What’s the overall picture of DeepSeek’s current funding round?

According to overseas media reports, DeepSeek raised more than RMB 50 billion in this round. Liang Wenfeng himself invested RMB 20 billion, while other investors included Tencent, CATL, JD.com, NetEase, IDG Capital, and others.

With this funding round completed, DeepSeek’s valuation has now topped US$50 billion.

Notably, DeepSeek officially launched this round of fundraising as early as April 2026, when it was valued at around US$10 billion. But over the following two months, DeepSeek’s valuation multiplied several times over, and the round subsequently became the largest financing round in China’s large-model sector.

2. What’s special about this funding round?

In terms of deal execution, DeepSeek adopted a special structure for this round.

According to reports from foreign media, in this fundraising, most investors were required to inject capital into a limited partnership controlled by DeepSeek CEO Liang Wenfeng, rather than investing directly into DeepSeek the company itself—and these investors were also subject to a five-year lock-up period.

Under this structure, these investors do not have voting rights in DeepSeek, but they can receive preferential financial information and have priority rights to participate in potential subsequent funding rounds.

Meanwhile, foreign media reports said that among all investors, the only exception was the National AI Industry Investment Fund. The fund could invest directly in DeepSeek, with an investment amount of RMB 1 billion; it has voting rights in the company and is not subject to the lock-up period.

3. Why arrange the financing this way?

Based on the current situation, the core reason for adopting this special arrangement is to safeguard Liang Wenfeng’s absolute control over DeepSeek.

After all, various pieces of information from the past suggest that Liang Wenfeng is firmly associated with a brand of technological idealism: he founded DeepSeek with the original intention of achieving artificial general intelligence and ensuring it can benefit ordinary people.

According to media reports, at a previous meeting with investors, Liang Wenfeng made it clear that the company would continue pushing forward open-source AI models, with artificial general intelligence (AGI) as its goal. DeepSeek’s top priority at present, he said, is to keep expanding the technological frontier—in other words, to prioritize breakthrough AI research over profitability and short-term commercialization.

Beyond this original vision, Liang Wenfeng set strict terms for this financing round. He established a five-year equity lock-up period to ensure that everyone coming on board is patient, long-term capital with strategic staying power—rather than investors looking for a quick cash-out.

4. Why was Tencent able to get in on this financing round?

For Tencent, investing in DeepSeek is an action that goes far beyond a purely financial investment; it is clearly a strategic move.

From Tencent’s perspective, it was among the earliest adopters of DeepSeek. When DeepSeek shot to fame in February 2025, Tencent was also one of the first to add support for DeepSeek in its Yuanbao App, and the two sides have very strong business synergies.

With this financing round, there is a strong chance that this synergy will continue as both parties further develop their AI businesses.

From another angle, Tencent’s ability to join the round is also related to its long-standing investment style. In its long-term investment practice, Tencent often takes an approach of providing multifaceted support to portfolio companies without proactively intervening, and it is generally very patient overall. This may also have been a key factor enabling Tencent to participate.

5. What is the core logic behind CATL’s investment in DeepSeek?

For DeepSeek, as it advances AI toward AGI, large-scale training compute is undoubtedly a crucial technical pillar. And at the physical-world level, compute is primarily supported by data centers—especially power-hungry AI data centers.

Data centers are precisely the major strategic and business convergence point between DeepSeek and CATL.

In particular, the energy-storage segment related to AI data centers is exactly the area CATL is currently doubling down on, and it is where CATL’s next growth curve lies.

It is worth noting that in a previous public interview discussing AI, CATL Chairman Zeng Yuqun said:

In the AI space, we view AI as a tool overall, rather than trying to compete with AI companies within the AI industry. But AI runs on energy; in the end, your AIDC (AI data center) and your tokens are essentially forms of energy.

On this basis, Zeng Yuqun emphasized that CATL hopes to contribute to energy solutions for AIDCs—which is also a core point of alignment behind CATL’s move to join this round of DeepSeek’s financing.

6. In terms of this financing round, what does it mean for the “national team” to come in?

In DeepSeek’s latest financing round, the entry of the “national team” is a particularly noteworthy point to watch.

Earlier, in May, the Financial Times reported that the National Integrated Circuit Industry Investment Fund was in talks with DeepSeek to lead its first financing round. However, in The Information’s latest report on June 16, the investing entity from the “national team” was instead the National AI Industry Investment Fund.

As a result, exactly how the “national team” has entered the deal remains murky.

What can be confirmed, however, is that in the current environment, for a company like DeepSeek, its competitiveness in AI has already gone beyond technology itself—and from the perspective of compute compatibility, eight domestic AI chipmakers, including Huawei Ascend, Cambricon, Muxi, and Moore Threads, have all completed full adaptation for DeepSeek-V4.

It is also worth noting that, according to a person familiar with the matter, this round included a special clause: Liang Wenfeng’s team required verification of the true identities of the limited partners behind all participating funds, to avoid the risk of equity ultimately ending up in the hands of unknown parties—the logic behind this is self-evident.

7. After this round of financing is completed, how will DeepSeek develop next?

On the one hand, from a product perspective, DeepSeek may have the opportunity to roll out new model versions down the line, offer enterprise users more tools, and support the industry-standard MCP protocol, including adding support for images and audio. To some extent, this is also aimed at moving more effectively toward commercialization.

On the other hand, DeepSeek will naturally need to step up investment in infrastructure, talent acquisition, and model training. Based on the latest publicly available updates so far, DeepSeek has already launched a hiring initiative for roles related to its Inner Mongolia data center (and the pay is not low), which is enough to show that its goal of building its own data center has begun to move into execution.

In the long run, DeepSeek’s goal is to achieve AGI. That is a long journey and will still require substantial投入—capital, talent, and more. For Liang Wenfeng, the unavoidable challenge will be how to maintain the company’s technical approach and its original commitment to broad accessibility while also managing investors’ expectations around DeepSeek’s commercialization and a reasonable return.

In a sense, DeepSeek’s first round of financing was merely the beginning of this journey. (Reporting by Wang Zhi and Editing by Zhou Yi)