DeepSeek reportedly in talks to raise $1.5B, then IPO

据报道，DeepSeek 正商谈筹集 15 亿美元资金，随后计划 IPO

DeepSeek, the Chinese large language model developer, is preparing for a 2027 IPO debut, but it could come as early as the end of this year as it also looks to raise around $1.5 billion in new funds at about a $71 billion valuation, Bloomberg reports. 据彭博社报道，中国大语言模型开发商 DeepSeek 正准备在 2027 年首次公开募股（IPO），但这一进程最早可能在今年年底启动。与此同时，该公司正寻求以约 710 亿美元的估值筹集约 15 亿美元的新资金。

The news comes after the company raised $7 billion in funds just a month ago at around a $50 billion valuation in its first-ever outside funding round. 在此消息传出之前，该公司仅在一个月前完成了其首次外部融资，当时筹集了 70 亿美元，估值约为 500 亿美元。

The China-based startup, founded in 2023, made headlines early last year after releasing AI technology that was both more efficient and more cost-effective than U.S. model makers. It has since grown exponentially in popularity. 这家成立于 2023 年的中国初创公司在去年年初因发布了比美国模型制造商更高效、更具成本效益的 AI 技术而登上头条。自那以后，其受欢迎程度呈指数级增长。

In June, it accounted for nearly 23% of all the tens of trillions of tokens processed by enterprise-focused AI gateway Vercel, the company reports. This is compared to Anthropic taking 32% of tokens. 据该公司报告，6 月份，在面向企业的 AI 网关 Vercel 处理的数万亿个 Token 中，DeepSeek 占据了近 23% 的份额。相比之下，Anthropic 占据了 32% 的份额。

DeepSeek continues to show how Chinese open source models perform within a few breaths of top U.S. AI labs, despite U.S. export controls on chips. DeepSeek’s cloud service runs on chips made by the Chinese company Huawei Technologies. 尽管美国对芯片实施了出口管制，但 DeepSeek 持续展示了中国开源模型在性能上如何紧追美国顶级 AI 实验室。DeepSeek 的云服务运行在中国华为技术有限公司生产的芯片上。

Investors in DeepSeek include Tencent and Beijing’s National Artificial Intelligence Industry Investment Fund, per Bloomberg. DeepSeek could not be reached for comment. 据彭博社报道，DeepSeek 的投资者包括腾讯和北京的国家人工智能产业投资基金。DeepSeek 目前未就此事置评。