The national team of China is personally stepping in to fund DeepSeek.
Just as Kimi hits a valuation of 20 billion, DeepSeek is rumored to be valued at 45 billion.
And I checked last night,
DeepSeek's lead investor isn't any VC, but the national team of China.
A state-level fund worth 344 billion RMB, specifically backing Chinese chips.
So from this moment on, DeepSeek is more than just an AI company.
I previously thought Liang Wenfeng would never go for funding.
Because he holds 84% of the shares,
The parent company, Huanfang Quantitative, is projected to net 5 billion in 2025,
Numerous investors like Wuyuan Capital have tried to pitch him but were turned away,
The reason being: different paths don’t make for a partnership.
Now the reason for opening up is probably quite simple.
DeepSeek has reached a level where,
funding, and the choice of chips,
are no longer just Liang Wenfeng's call.
Plus, V4 has been consistently delayed,
The reason being a migration from Nvidia to Huawei Ascend,
Almost all the underlying code needs to be rewritten;
Core researchers have also been poached by Xiaomi and ByteDance;
After R1 broke out, the daily active users reached 30 million, and the inference costs are unsustainable.
Three issues colliding at once.
OpenAI's backing comes from Microsoft + Nvidia + SoftBank.
DeepSeek's ally is the state itself.
After all, the color of money is different, and the paths diverge.
Just as Kimi hits a valuation of 20 billion, DeepSeek is rumored to be valued at 45 billion.
And I checked last night,
DeepSeek's lead investor isn't any VC, but the national team of China.
A state-level fund worth 344 billion RMB, specifically backing Chinese chips.
So from this moment on, DeepSeek is more than just an AI company.
I previously thought Liang Wenfeng would never go for funding.
Because he holds 84% of the shares,
The parent company, Huanfang Quantitative, is projected to net 5 billion in 2025,
Numerous investors like Wuyuan Capital have tried to pitch him but were turned away,
The reason being: different paths don’t make for a partnership.
Now the reason for opening up is probably quite simple.
DeepSeek has reached a level where,
funding, and the choice of chips,
are no longer just Liang Wenfeng's call.
Plus, V4 has been consistently delayed,
The reason being a migration from Nvidia to Huawei Ascend,
Almost all the underlying code needs to be rewritten;
Core researchers have also been poached by Xiaomi and ByteDance;
After R1 broke out, the daily active users reached 30 million, and the inference costs are unsustainable.
Three issues colliding at once.
OpenAI's backing comes from Microsoft + Nvidia + SoftBank.
DeepSeek's ally is the state itself.
After all, the color of money is different, and the paths diverge.