DeepSeek Shelves $70B Fundraising Round After CEO's Leaked Investor Call Goes Viral
Summary
- • DeepSeek paused its next fundraising round, which was targeting a $70 billion valuation.
- • CEO Liang Wenfeng's private investor call leaked publicly; he admitted DeepSeek still trails US AI labs mainly due to lack of Nvidia compute.
- • Wenfeng stated DeepSeek would prioritize open research over near-term monetization — signaling delayed revenue timelines to investors.
- • The funding pause is driven by frustration over the leak itself, not any change in DeepSeek's underlying strategy or roadmap.
Details
$70B Valuation Target Paused
DeepSeek's fundraising round, targeting a $70 billion valuation, has been shelved — one of the largest valuations sought by a Chinese AI company.
Pause Triggered by Leaked Investor Call
CEO frustration over the viral spread of private investor call comments caused the pause; no underlying strategy shift has occurred according to reporting.
CEO Admits Nvidia Compute Gap
Liang Wenfeng acknowledged on the investor call that DeepSeek still trails US frontier labs primarily due to lack of access to Nvidia GPUs — a rare on-record admission.
US Export Controls Confirmed Effective
The CEO's admission directly links US chip export restrictions to a meaningful competitive disadvantage for China's leading AI lab.
Open Research Prioritized Over Monetization
Wenfeng stated DeepSeek will prioritize open research and model releases over near-term revenue generation, potentially lengthening investor return timelines.
Leak-Driven Reputational Risk
The viral spread of confidential investor communications raises questions about information security in high-stakes Chinese AI investment processes.
Source: AI Daily Brief (newsletter)
What This Means
DeepSeek's paused fundraising round reveals the real-world costs of information leaks in high-stakes AI investment. More importantly, CEO Liang Wenfeng's candid admission that DeepSeek still lags US frontier labs because of Nvidia compute access is a rare on-record data point confirming that US export controls are materially constraining China's best AI labs — even one that publicly positioned itself as thriving despite those restrictions. The stated priority of open research over monetization may also reduce DeepSeek's appeal to traditional venture investors, creating long-term uncertainty about how the company will sustain its R&D investment as the frontier compute gap persists.
Sentiment
Analytical and focused on US export controls' impact, with notes on DeepSeek's open-source priorities
“DeepSeek pauses its ~$74B second round after leaked transcript shows founder Liang Wenfeng acknowledging the US compute gap. The $7.4B first round closed just weeks ago, per Bloomberg. Signal: even the efficiency king can't outrun the hardware arms race.”
“DeepSeek founder Liang Wenfeng held a 4-hour investor meeting: AGI over profit, open-source as strategy, and compute is the real US-China gap — not talent. Then DeepSeek suspended its $7.4B round at a $74B valuation. Not desperate for capital. Possibly avoiding over-dilution before an IPO on the STAR Market.”
“According to the DeepSeek CEO leaked memo they are both hooked on US chips and limited by export controls. Seems like the best strategy is giving them juuuust enough to get hooked but not enough to lead the frontier. Threading the needle.”
“reiterating that US closed frontier is ahead of china open frontier because chip sanctions Are Working is certainly A Take. Isn't this what the Deepseek CEO said in the leaked investor call?”
“CEO Liang Wenfeng admitted that 🇨🇳 still trails the U.S. in AI sophistication & remains dependent on $NVDA chips. Following the leak of these remarks from an investor meeting earlier, DeepSeek abruptly “involuntarily” paused its 2nd fundraising round targeting a $71B valuation.”
Split
~60/40 split between those emphasizing US chip sanctions' effectiveness vs. those noting DeepSeek's deliberate open-source/anti-monetization stance as a separate factor