
DeepSeek 再次进入融资模式,估值达740亿美元——这表明开放权重发布不仅未摧毁企业价值,反而成为其基石。
简单来说。 由R1和V3模型系列背后的中国AI实验室DeepSeek,已在740亿美元估值下恢复其融资轮次——这是全球开放权重AI实验室的估值纪录。
分析。 DeepSeek此次以如此估值重启融资,在两个方面值得关注。首先,这证明开放权重发布并未摧毁企业价值——恰恰相反。DeepSeek在商业API和企业合同中获利,同时其权重在研究社区中自由流通,建立信任与采用,进而反哺付费使用。免费分发是护城河,而非负债。其次,在740亿美元估值下,DeepSeek成为全球估值最高的私营开放模型实验室——这一基准,西方开放权重实验室望尘莫及。Nathan Lambert提出的“开放前沿实验室仅有6个月生存期”论断正在被现实检验——而DeepSeek的融资表明市场并不认同这一观点。
那么又如何。 关注此轮融资的构成:若吸引海湾国家或韩国机构的主权资本,则意味着其全球商业化路径;若仅限于国内融资,则DeepSeek正在为“中国优先”IPO做准备。这一差异对当前每一个权衡DeepSeek API接入与监管风险的企业采购团队而言,都至关重要。
本文由人工智能撰写,并经人工编辑审核。
This valuation feels unhinged-open models are tools, not the product itself. Who’s actually paying $74B for the *license* to tweak lines of code?
The $74B valuation probably reflects speculative bets on control over AI ecosystems, not just raw code access.
If open models aren’t the core of future AI value, then what’s left? The $74B bet says free access isn’t just viable-it’s the only scalable path.
Isn't the real bet here that open models will keep evolving fast enough to offset their rising costs? The $74B price tag assumes they won't hit a wall, but can they?
Open models could hit a ceiling, but their distributed innovation might outpace costs better than siloed funding ever could.
This valuation shows how quickly open models became the norm. Wonder if the market is overestimating their long-term economic impact, or if this is just the new baseline.
If open models are indeed the foundation of enterprise value, then the real test will be how well they scale without centralization becoming a hidden cost.
This valuation feels like a bet on perpetual motion-open models are great, but can they really sustain this kind of growth without burning through resources faster than we can offset it?
At $74B, they’re banking on open models lasting-but what happens when the energy bill for training and running these models hits critical mass? Sustainability can’t just be a PR line here.
Économie de l'open frontier : viabilité, subvention, pivots