人工智能模型初創公司 Reflection AI 融資 20 億美元

Recently the New York Times reported the following:

Reflection AI, an A.I. Model Start-Up, Raises $2 Billion

The big fund-raising round was the latest sign of investor fervor for artificial intelligence companies despite concerns that the boom is overheated.

The NYT - By Michael J. de la Merced

Updated Oct. 10, 2025, 12:01 p.m. ET

For much of the past decade, American technology companies like OpenAI and Google have dominated the artificial intelligence industry. Then came DeepSeek, the Chinese start-up whose A.I. model was nearly as powerful as competing Western ones — and it was free to use.

Now, Reflection AI, a year-old American start-up, is betting that it can create a Western equivalent of DeepSeek. And it has raised billions to do so.

Reflection AI announced on Thursday that it had raised $2 billion in new funding, in a round that included Nvidia, the chipmaker at the heart of the A.I. boom. The deal values Reflection AI at $8 billion, including the new money, significantly more than the roughly $545 million the company was valued at in March, according to the data provider PitchBook.

The fund-raising is yet another sign of the investor fervor for all things A.I., as companies across the industry continue to collect eye-popping sums. Companies that build A.I. foundation models have raised $71.9 billion worldwide so far this year, according to PitchBook; by comparison, the sector collected $34.9 billion last year.

But the move also shows that investors are willing to invest not only in the proprietary software sold by OpenAI and Google, but also in open-source models, which are free to share and modify. Skeptics of open-source A.I. say the approach could cause significant harm.

Yet Reflection AI argues that the United States needs an equivalent to DeepSeek — an open-source model developer whose software can compete with top closed-source models — to maintain technological superiority around the world.

Existing Western open-source A.I. models are underperforming DeepSeek and other Chinese rivals, which may lead to greater adoption of Chinese-created models, according to Misha Laskin, Reflection AI’s co-founder and chief executive. (Among the Western companies making open-source models are Meta, Mistral AI of France and even OpenAI, though that company is primarily focused on proprietary software.)

“There’s a DeepSeek-shaped hole in the U.S., which I think is what makes it critical for a lab like ours to exist,” Mr. Laskin said in an interview, comparing the situation to the Cold War space race.

But developing any A.I. model, open-source or not, requires ever-higher amounts of computing power and research talent — in short, money. That is why Reflection AI raised $2 billion just seven months after raising $130 million in March.

In recent months, Reflection AI held talks with potential investors, and closed the round within a matter of weeks, Mr. Laskin said. Nvidia wrote the biggest check — $800 million — and had several engineers work with Reflection AI to optimize its most recent generation of A.I. chips, according to a person with knowledge of the matter, who spoke on the condition of anonymity because they were not authorized to discuss the deal’s details.

Other participants in the round included Disruptive Technology Advisers; Lightspeed Venture Partners; Sequoia Capital;  DST; the former Google chief executive Eric Schmidt; and 1789 Capital, the investment firm that counts Donald Trump Jr. as a partner.

Mr. Laskin conceded that Reflection AI would need more money to keep competing, especially as rivals are also rushing to raise money: OpenAI alone secured an investment commitment of up to $100 billion last month from Nvidia.

But he argued that demand for open-source models — including from big companies and governments that want greater control over their A.I. applications — will eventually lead to a sustainable business.

The hope, he added, is that smaller A.I. companies eventually end up being the Amazons and Metas of the future: “I anticipate us to be potentially a bigger company than the hyperscalers of today,” he said.

Translation

人工智能模型初創公司 Reflection AI 融資 20 億美元

儘管人們擔心人工智能熱潮過熱，但此次大規模融資再次顯示投資者對人工智能公司依然熱情高漲。

在過去十年的大部分時間裡，OpenAI 和谷歌等美國科技公司一直主導著人工智能產業。後來，中國新創公司 DeepSeek 應運而生，其人工智能模型幾乎與西方競爭對手的模型一樣強大，而且免費使用。

如今，一間成立一年的美國初創公司 Reflection AI 正押注於打造西方版的 DeepSeek。為此，它已經籌集了數十億美元。

Reflection AI 週四宣佈，它已完成 20 億美元的新一輪融資，融資參與者包括人工智能熱潮核心晶片製造商英偉達在內。根據數據提供商 PitchBook ，這筆交易對 Reflection AI 的估值為 80 億美元（包括新資金），遠高於該公司 3 月約 5.45 億美元的估值。

此次融資再次顯示投資者對人工智能領域充滿熱情，各行各業的公司都在持續籌集令人瞠目結舌的資金。據 PitchBook 稱，今年迄今為止，建立人工智能基礎模型的公司已在全球籌集了 719 億美元；相比之下，該行業去年的融資額僅為 349 億美元。

但此舉也表明，投資人不僅願意投資 OpenAI 和Google出售的專有軟體，還願意投資可以免費分享和修改的開源模型。開源人工智能的懷疑者表示，這種做法可能造成重大損害。

然而，Reflection AI 認為，美國需要一個類似 DeepSeek 的公司 - 一個開源模型開發商，其軟體可以與頂級閉源模型相媲美 - 以保持全球技術優勢。

Reflection AI 聯合創辦人兼執行長 Misha Laskin 表示，現有的西方開源人工智能模式表現遜於 DeepSeek 和其他中國競爭對手，這可能導致中國自主研發的模式被更廣泛地採用。 （開發開源模型的西方公司包括 Meta、法國的 Mistral AI，甚至是 OpenAI，儘管後者主要專注於專有軟件。）

Laskin 先生在一次採訪中說道: 「美國存在一個由DeepSeek所帶出來缺口，我認為這就是我們這樣的實驗室存在的關鍵」，他將這種情況比作冷戰時期的太空競賽。

但是，開發任何人工智能模型，無論是否開源，都需要越來越強大的運算能力和研究人才 - 簡而言之，就是資金。這就是為什麼 Reflection AI 在 3 月籌集 1.3 億美元後，光是七個月就籌集了 20 億美元。

Laskin 先生表示，最近幾個月，Reflection AI 與潛在投資者進行了洽談，並在幾週內完成了本輪融資。據一位知情人士透露，英偉達簽下了金額最大的一筆融資 - 8億美元，並安排多名工程師與 Reflection AI 合作優化其最新一代的人工智能晶片。由於未獲授權討論交易細節，該人士要求匿名。

參與此輪融資的其他公司包括 Disruptive Technology Adviser, Lightspeed Venture Partners, Sequoia Capita, DST, 前谷歌執行長 Eric Schmidt，以及小特朗普為其合夥人的投資公司 1789 Capital。

Laskin 先生承認，Reflection AI 需要更多資金才能保持競爭力，尤其是在競爭對手也在爭相融資的情況下：僅 OpenAI 一家就於上個月從英偉達獲得了高達 1,000 億美元的投資承諾。

但他認為，對開源模型的需求 - 包括來自希望對其人工智能應用擁有更大控制權的大公司和政府的需求 - 最終將帶來可持續發展的業務。

他補充說，願望是小型人工智能公司最終能夠成為未來的亞馬遜和 Meta, 他說：「我預計我們可能會成為比今天的超大規模運算公司更大的公司」。

So, with the coming of DeepSeek in China now comes the Reflection AI, a year-old American start-up that is betting that it can create a Western equivalent of DeepSeek. It has raised billions to do so. Apparently, the AI boom is getting hotter and hotter.