The AI many Malaysian developers, students and small businesses quietly adopted because it was free, open and cheap is no longer a scrappy newcomer. DeepSeek, the Chinese model maker, is now reportedly worth around 71 billion US dollars and getting ready to sell shares to the public. That shift from low-cost upstart to heavily funded giant changes the calculation for anyone in Malaysia who has come to depend on it.
Editor
Kai T chevron_right
Table of Contents
From free upstart to a $71 billion company
DeepSeek is in talks to raise about 1.5 billion US dollars in fresh funding at a valuation of roughly 71 billion US dollars, and is preparing for a stock market debut in 2027 that could come as early as the end of this year, according to a Bloomberg report cited by TechCrunch. The talks come barely a month after the company closed its first ever outside funding round, taking in about 7 billion US dollars at a valuation of around 50 billion. Founded in 2023, DeepSeek counts Tencent and a Beijing state-backed artificial intelligence fund among its investors, and its cloud service runs on chips made by Huawei rather than the Nvidia hardware most Western labs use.
The rise has been fast. DeepSeek drew global attention in early 2025 with models that matched much of what American labs offered at a fraction of the cost, and it has kept closing the gap despite United States export controls on advanced chips. By June this year it accounted for nearly 23 percent of the tokens processed through Vercel, an enterprise AI gateway, not far behind Anthropic on 32 percent. A Microsoft study reported by Euronews found the model gaining ground fastest in developing economies, where open weights and low prices matter most.
Why this matters for Malaysia
Malaysia sits squarely inside that pattern. Local developers, startups and SMEs have leaned on DeepSeek precisely because it is open and inexpensive next to ChatGPT or Claude, and the country went further than most: its sovereign AI stack became the first national deployment of DeepSeek outside China, as we covered in Malaysia Runs DeepSeek, Now It Is Building Its Own Chip, before the trade ministry distanced the government from the Huawei-powered project. For a lot of Malaysian builders, DeepSeek is not a curiosity. It is the default cheap engine behind a chatbot, a coding assistant or an internal tool.
That is exactly why its move toward an IPO is worth watching. A company raising billions from Tencent and a Chinese state fund, and heading for a public listing, answers to shareholders in a way a free research release does not. The open, low-cost model Malaysian teams built on now carries commercial pressure to earn a return, questions about how long pricing stays this low, and a dependence on Huawei silicon that keeps it exposed to the United States and China trade fight. None of that is a reason to abandon the tool. It is a reason to plan for a world where the terms can change.
What to watch next
Three things will tell Malaysian users where this is going. First, whether DeepSeek keeps releasing open weights once it has public shareholders, or starts holding its best models back. Second, where and when it actually lists, since a mainland or Hong Kong listing carries different disclosure and access rules than a Western one. Third, whether local outfits treat DeepSeek as one option among several rather than a single point of failure, especially for anything touching sensitive data.
DeepSeek got popular in Malaysia by being the affordable, open alternative. The safest way to keep benefiting from it is to remember that a 71 billion US dollar company on its way to an IPO does not stay an underdog for long.
Images courtesy of Zulfugar Karimov and Danial Igdery on Unsplash.