AI 商业化 · 行业观察

River AI成立数月融11亿
英伟达AMD同时押注开放AI栈

River AI Raises $1.1B in Months
NVIDIA & AMD Both Bet on Open AI Stack

xAI联合创始人出走创立River AI,成立数月就融了11亿美元,英伟达和AMD罕见同时战略投资。当巨头们同时扶持一个'开放'玩家,AI基础设施的故事正在被重写。

xAI co-founder leaves to launch River AI, raising $1.1B within months — NVIDIA and AMD make a rare joint strategic investment. When giants both back an 'open' player, the AI infrastructure narrative is being rewritten.

No.043 2026.08.26 约 5 分钟阅读 ~5 min read

11亿美元。一家AI公司,成立才几个月,产品还在开发中,没有公开的demo,没有营收——就融到了11亿美元。

这不是段子,是8月11日River AI官宣的融资消息。这轮融资由General Catalyst与AMP PBC联合领投,英伟达和AMD罕见同时以战略投资者身份入局,Y Combinator、淡马锡等跟投。公司创始人Igor Babuschkin是xAI的联合创始人,之前还在DeepMind做过AlphaCode。

明星团队+明星投资人+巨额融资=?答案是:AI基础设施创业的泡沫,已经吹到了一个新高度。

英伟达和AMD为什么同时投资?

这件事最有意思的点,不是11亿美元的金额,也不是成立几个月就融资的速度,而是英伟达和AMD同时投了

英伟达和AMD是什么关系?是GPU市场上厮杀了几十年的老对手。它们在技术路线、市场份额、生态建设上处处针锋相对。你很少会看到它们同时投资同一家初创公司——除非这家公司的存在,对两家都有战略价值。

River AI的定位是构建"开放AI栈"(Open AI Stack),主打"AI归用户所有",而不是掌握在实验室巨头手中。这个定位很微妙——它既不是闭源的Anthropic/OpenAI路线,也不是纯开源的Meta/Llama路线,而是走一条中间道路:开放的、可定制的、用户拥有所有权的AI基础设施。

对英伟达来说,投资River AI是对冲OpenAI/Anthropic的议价能力。如果AI市场被一两家闭源模型公司垄断,那它们就有能力跟英伟达砍价——"你不给我便宜的GPU,我就找AMD/台积电自研芯片"。多扶持几个玩家,让市场保持竞争,对英伟达是有利的。

对AMD来说,意义更不一样。AMD在AI加速卡市场的份额远不如英伟达,它需要更多的合作伙伴来生态绑定。如果River AI的开放栈能做起来,并且原生支持AMD的MI系列芯片,那AMD就多了一个撬动市场的支点。

所以你看,两家竞争对手同时投一家公司,不是因为这家公司有多厉害,而是因为它的存在符合两家巨头的战略利益。这就像冷战时期,两个超级大国同时给一个小国援助——不是因为这个小国有多重要,而是因为它在棋盘上的位置有用。

AI创业的估值逻辑变了吗?

River AI的融资让人不禁想问:AI创业的估值逻辑,是不是已经完全脱离了基本面?

放在几年前,一家成立几个月、没有产品、没有营收的公司,能融到几千万美元就已经很厉害了。现在呢?11亿美元起步。这个数字,比很多上市公司的市值都高。

支撑这种估值的逻辑是什么?团队叙事+赛道想象+巨头背书。创始人是xAI联创+DeepMind老兵——OK,团队顶级;赛道是AI基础设施——OK,市场足够大;英伟达AMD都投了——OK,巨头背书,跟着投准没错。

这种逻辑对不对?短期来看可能是对的——在一个快速增长的市场里,抢位置比算ROI重要。你今天不投,明天可能就被竞争对手投了,后天这个公司可能就成了下一个OpenAI。FOMO(害怕错过)的情绪,在AI投资圈里比任何行业都要强烈。

但长期来看,泡沫总是要破的。历史上每一次技术革命,都会经历"非理性繁荣→泡沫破裂→理性增长"的周期。互联网泡沫、移动互联网泡沫、区块链泡沫——都是这么过来的。AI会不会不一样?可能不会。

不过话说回来,泡沫不一定是坏事。没有互联网泡沫,就没有后来的Google、Amazon;没有移动互联网泡沫,就没有今天的移动生态。泡沫期大量的资金涌入,虽然浪费了很多,但也加速了技术的进步和生态的建立。问题只是:谁能在泡沫破裂后活下来,成为下一个时代的巨头?

River AI能不能活下来?没人知道。但11亿美元的弹药,足够它撑好一阵了。至于这11亿美元最终会变成下一个OpenAI的种子,还是变成又一个烧掉的泡沫——只有时间能给出答案。

明天见。

$1.1 billion. An AI company, founded only months ago, product still in development, no public demo, no revenue — and it raises $1.1 billion.

This isn't a joke. It's River AI's funding announcement from August 11. The round was co-led by General Catalyst and AMP PBC, with NVIDIA and AMD making a rare joint strategic investment, and participation from Y Combinator, Temasek, and others. Founder Igor Babuschkin is a co-founder of xAI and previously worked on AlphaCode at DeepMind.

Star team + star investors + huge funding = ? Answer: the AI infrastructure startup bubble has reached a new height.

Why Are NVIDIA and AMD Both Investing?

The most interesting part of this story isn't the $1.1 billion, or the speed of funding a few months after founding — it's that NVIDIA and AMD both invested.

What's the relationship between NVIDIA and AMD? They've been bitter rivals in the GPU market for decades. They compete head-to-head on technology roadmap, market share, ecosystem building. You rarely see them investing in the same startup — unless that company's existence holds strategic value for both.

River AI positions itself as building an "Open AI Stack", with the vision that "AI belongs to the users" rather than being controlled by lab giants. It's a delicate positioning — it's not the closed-source Anthropic/OpenAI path, nor the pure open-source Meta/Llama path, but a middle road: open, customizable, user-owned AI infrastructure.

For NVIDIA, investing in River AI is about hedging against the pricing power of OpenAI/Anthropic. If the AI market gets monopolized by one or two closed-source model companies, they gain leverage to negotiate with NVIDIA — "give me cheaper GPUs or we'll go to AMD / build our own chips with TSMC." Supporting more players to keep the market competitive benefits NVIDIA.

For AMD, the meaning is different. AMD's share of the AI accelerator market is much smaller than NVIDIA's — it needs more partners for ecosystem binding. If River AI's open stack takes off and natively supports AMD's MI-series chips, AMD gains a new lever to pry open the market.

So you see: two competitors investing in the same company isn't because the company is that great — it's because its existence serves both giants' strategic interests. It's like the Cold War, where two superpowers both aid a small country — not because the country matters that much, but because its position on the board is useful.

Has AI Startup Valuation Logic Changed?

River AI's funding makes you wonder: has AI startup valuation logic completely detached from fundamentals?

A few years ago, a company a few months old with no product and no revenue would be doing great to raise tens of millions. Now? $1.1 billion is the starting line. That number is higher than the market cap of many public companies.

What's the logic supporting this valuation? Team narrative +赛道 imagination + giant endorsement. Founder is xAI co-founder + DeepMind veteran — OK, top-tier team. The赛道 is AI infrastructure — OK, market is big enough. NVIDIA and AMD both invested — OK, giant endorsement, must be a safe bet to follow.

Is this logic correct? In the short term, maybe — in a fast-growing market, grabbing position matters more than calculating ROI. If you don't invest today, a competitor might tomorrow, and the day after that company could become the next OpenAI. FOMO (fear of missing out) is stronger in AI investing than any other industry.

But long-term, bubbles always pop. Every tech revolution in history goes through the cycle of "irrational exuberance → bubble burst → rational growth." The dot-com bubble, the mobile internet bubble, the blockchain bubble — all followed this path. Will AI be different? Probably not.

That said, bubbles aren't necessarily bad. Without the dot-com bubble, there would be no Google or Amazon. Without the mobile internet bubble, there'd be no mobile ecosystem today. The flood of capital during bubble periods wastes a lot, but also accelerates technological progress and ecosystem building. The question is just: who survives the burst and becomes the next era's giant?

Will River AI survive? Nobody knows. But $1.1 billion in ammo is enough to keep it going for quite a while. Whether that $1.1 billion eventually becomes the seed of the next OpenAI, or just another burned bubble — only time will tell.

See you tomorrow.

两个竞争对手同时投资一家公司,不是因为这家公司有多厉害——而是因为它在棋盘上的位置,对两家都有用。

—— Dawn Vision编辑部

Two rivals investing in the same company isn't because it's great — it's because its position on the board benefits both sides.

— The Dawn Vision Editorial Desk
River AI · 11亿美元融资 · Igor Babuschkin · xAI · 开放AI栈 · 英伟达AMD · AI基础设施 · 创业泡沫
River AI · $1.1B funding · Igor Babuschkin · xAI · open AI stack · NVIDIA AMD · AI infrastructure · startup bubble
Sources · 信源 Sources

本文基于 Dawn Vision 认知引擎处理的 8 个源信号生成,经编辑部人工审核。素材来源:Business Wire、SiliconANGLE、AIPressRoom、全球独角兽500强、WealthEngine。

This article was generated by the Dawn Vision cognitive engine processing 8 source signals, with human editorial review. Sources: Business Wire, SiliconANGLE, AIPressRoom, Global Unicorn 500, WealthEngine.