DeepSeek, the Chinese AI startup that triggered a worldwide "DeepSeek shock" with its low-cost, high-performance artificial intelligence model, has abruptly halted its second round of fundraising, according to reports. The leak of a recording from a private investor meeting held by founder Liang Wenfeng, known as a "reclusive genius," is cited as the background.
Angered by Leaked Private Meeting Remarks? Sudden Fundraising Halt
Bloomberg reported on the 25th, citing multiple sources, that DeepSeek verbally informed some investors set to participate in its second funding round that it would not sign investment contracts for the time being. It is unclear whether DeepSeek conveyed this intention to all companies seeking to participate in the second funding round.
DeepSeek had originally planned to secure more than 10 billion yuan (about 2.1 trillion won) through the second round of fundraising. The final amount raised could increase depending on the number of participating firms. The first funding round, completed in June, was worth 7 billion dollars.
Industry sources cited the recent leak of a recording from Liang's investor meeting as the reason the smoothly progressing second funding process was suddenly halted. Bloomberg noted that Liang, who has been extremely reluctant to appear in public, was displeased over this, and that this was the background to the fundraising halt. However, it did not specify what exactly he was displeased about.
Earlier, on the 24th, China Business News and Securities Times reported details of Liang's remarks, saying they had obtained a recording of a meeting he held with investors in May that lasted about three hours and 44 minutes. It was the first time he had specifically expressed his views on management goals and direction, and the future of the AI industry.
"Nvidia's Monopoly Crumbling... U.S.-China Gap Is Only Computing Resources"
At the meeting, Liang acknowledged that the only gap in the U.S.-China AI industry was computing resources such as Nvidia's graphics processing units (GPUs), saying that "technological progress has paused for a moment." But at the same time, he stressed that "there is no problem at all with the Chinese chip ecosystem," adding, "In the past, many viewed them as difficult to use, but within the next year we will see them gradually replaced with Chinese-made products." He said the monopoly power of Nvidia's "CUDA" was rapidly crumbling as it became possible to build a new ecosystem using AI, and expressed hope that this was a historic opportunity to replace them with Chinese AI chips.
He also stated that DeepSeek's long-term goal is not generative AI but the realization of artificial general intelligence (AGI, an AI that learns, reasons and solves problems on its own like a human). "Early last year, we gained many users as our generative AI model became hugely popular, but we don't think about how to retain them or monetize," he said. "Because the opportunity in AGI is much greater, these factors are a small part." He added, "DeepSeek has no intention of building a super app, nor does it hope to become like ByteDance or Tencent," saying, "We plan to focus on realizing AGI rather than competing for market share." Liang also revealed that he chose investors who support DeepSeek's AGI development direction and long-term goals over those offering a higher price.
Liang particularly emphasized the importance of talent. "As long as we can maintain the stability of the team, AGI can definitely be built," he said, calling the retention of core talent the company's top priority.
DeepSeek has now begun the initial public offering (IPO) process. It is preparing to submit an IPO application within this year, aiming for a listing on the Chinese mainland stock market in 2027.
"Liang Wenfeng, No. 1 in Wealth Among AI Founders Worldwide"... Also Preparing for IPO
DeepSeek launched its low-cost, high-performance open-source AI model in January last year, shaking up the AI industry, which had once been considered the exclusive domain of capital-rich U.S. big tech.
According to Bloomberg, DeepSeek was valued at about 7.4 billion dollars (about 10.8 trillion won) in its first funding round. This was the highest-ever record for a first fundraising by a Chinese tech startup. Through this second round, DeepSeek aims to secure more than 10 billion yuan in additional funds and to be valued at more than 480 billion yuan (about 103.7 trillion won), more than 10 times higher than the first round.
Since the first funding round, Liang's net worth has more than doubled, making him the world's richest founder among global AI model founders, Bloomberg reported. According to the Bloomberg Billionaires Index, his wealth has grown from about 16.7 billion dollars previously to 36 billion dollars (about 52.7 trillion won) now. This is far ahead of Anthropic co-founder Dario Amodei and OpenAI co-founder and President Greg Brockman.
