分享我在Bloomberg China Show上的對(duì)談。

在訪談中，我分享了零一萬(wàn)物近期發(fā)布萬(wàn)智企業(yè)大模型一站式平臺(tái)背后的行業(yè)思考。如同 Windows 內(nèi)核需要操作系統(tǒng)釋放價(jià)值，DeepSeek 同樣需要中間件才能深入產(chǎn)業(yè)。零一萬(wàn)物所要做的就是填補(bǔ) AI 2.0 時(shí)代 “Windows 操作系統(tǒng)”的空白。

另一方面，中美超大模型的預(yù)訓(xùn)練正在逐漸寡頭化。未來(lái)中國(guó)大模型有很大概率會(huì)收攏到 DeepSeek、阿里巴巴、字節(jié)跳動(dòng)三家，其中以 DeepSeek 勢(shì)頭最猛。這也是萬(wàn)智選擇接入 DeepSeek 的原因。

以大模型為代表的 AI 2.0 是有史以來(lái)最偉大的科技革命與平臺(tái)革命。各地政府都在加緊探索以人工智能打造新質(zhì)生產(chǎn)力，零一萬(wàn)物也會(huì)積極與地方政府及產(chǎn)業(yè)展開(kāi)合作、共同探索。

本文摘編翻譯自彭博社Bloomberg China Show：

3 月 20 日，零一萬(wàn)物 CEO、創(chuàng)新工場(chǎng)董事長(zhǎng)李開(kāi)復(fù)博士接受了彭博社的專訪。

在訪談中他表示，DeepSeek 熱潮之后，中國(guó)經(jīng)歷了自己的“DeepSeek”時(shí)刻。零一萬(wàn)物看到了 DeepSeek 在基座模型領(lǐng)域的技術(shù)突破，也洞察到在基座模型與市場(chǎng)需求之間存在著缺位——由于缺少連接企業(yè)知識(shí)庫(kù)、搭建應(yīng)用的中間件，DeepSeek 優(yōu)質(zhì)模型難以轉(zhuǎn)變?yōu)槠髽I(yè)生產(chǎn)力。

李開(kāi)復(fù)以 Windows 內(nèi)核類比：DeepSeek 是 Windows 內(nèi)核，而零一萬(wàn)物提供的平臺(tái)則是 Windows 操作系統(tǒng)、應(yīng)用程序、UI 界面。沒(méi)有 Windows 操作系統(tǒng)、應(yīng)用程序、UI 界面，Windows 內(nèi)核就無(wú)法發(fā)揮真正的價(jià)值。

李開(kāi)復(fù)強(qiáng)調(diào)，AI 需要市場(chǎng)，市場(chǎng)也需要 AI。要真正發(fā)揮大模型的價(jià)值所在，AI 領(lǐng)域內(nèi)仍然沒(méi)有出現(xiàn)自己的“ Windows 操作系統(tǒng)”，零一萬(wàn)物想要做的正是填補(bǔ)這一空缺。

因此，零一萬(wàn)物選擇全面擁抱 DeepSeek，推出萬(wàn)智企業(yè)大模型一站式平臺(tái)，不僅提供安全模型部署方案、模型微調(diào)方案，還封裝了聯(lián)網(wǎng)搜索、知識(shí)庫(kù)RAG、Agent 搭建等工具，助力 DeepSeek 等優(yōu)質(zhì)基座模型賦能企業(yè)場(chǎng)景。

李開(kāi)復(fù)博士（右一）在訪談現(xiàn)場(chǎng)

談及大模型領(lǐng)域未來(lái)的競(jìng)爭(zhēng)格局，李開(kāi)復(fù)認(rèn)為，在中美，超大模型的預(yù)訓(xùn)練正在逐漸寡頭化，并且寡頭化的程度在不斷加大。他預(yù)測(cè)中國(guó)大模型領(lǐng)域未來(lái)將會(huì)收斂至 DeepSeek、阿里巴巴、字節(jié)跳動(dòng)三家，其中以 DeepSeek 勢(shì)頭最猛；美國(guó)則以 xAI 的發(fā)展勢(shì)頭最為迅速，但 OpenAI、Anthropic、Google 也都會(huì)做出令人敬佩的研發(fā)工作。

至于中美之間的 AI 競(jìng)賽，李開(kāi)復(fù)認(rèn)為 GPU 資源稀缺的情況反而倒逼著 DeepSeek 做出了驚艷業(yè)界的技術(shù)創(chuàng)新，實(shí)現(xiàn)訓(xùn)推成本與模型性能的突破。在印證了“需求是創(chuàng)新之母”的同時(shí)，DeepSeek 的突破也使得開(kāi)源模式呈現(xiàn)出壓倒性的優(yōu)勢(shì)，堅(jiān)持閉源策略的 OpenAI、Anthropic 等公司，都會(huì)面臨更大的壓力。

“OpenAI 在 2024 年的運(yùn)營(yíng)成本為 70 億美元，而 DeepSeek 的運(yùn)營(yíng)成本可能只有OpenAI 的 2%。目前各家的模型都很優(yōu)秀，問(wèn)題不在于哪家模型性能高出 1%，而是 OpenAI 成本高昂的模型路線是否具有可持續(xù)性?！崩铋_(kāi)復(fù)強(qiáng)調(diào)，“OpenAI 每年花費(fèi) 70 億美元，面臨著巨額虧損?，F(xiàn)在出現(xiàn)了一個(gè)競(jìng)爭(zhēng)對(duì)手，將成本低廉?dāng)?shù)倍的開(kāi)源模型免費(fèi)開(kāi)放給市場(chǎng)，且這個(gè)競(jìng)爭(zhēng)對(duì)手資源充沛，并已經(jīng)有效地將計(jì)算成本降低了五到十倍。

“有了這樣一個(gè)強(qiáng)大的競(jìng)爭(zhēng)對(duì)手，我認(rèn)為 OpenAI 的薩姆·奧爾特曼可能夜里輾轉(zhuǎn)難眠?！?/strong>李開(kāi)復(fù)開(kāi)玩笑說(shuō)。

以下為對(duì)話全文：

彭博社：今年可謂是人工智能真正滲透各行業(yè)的元年，此次我們特邀零一萬(wàn)物CEO、創(chuàng)新工場(chǎng)董事長(zhǎng)李開(kāi)復(fù)博士來(lái)深入探討這一話題。今天也是一個(gè)絕佳的時(shí)機(jī)來(lái)和您一起討論，因?yàn)榱阋蝗f(wàn)物剛剛對(duì)外發(fā)布了“萬(wàn)智企業(yè)大模型一站式平臺(tái)”，首發(fā)能幫助企業(yè)更廣泛地部署、應(yīng)用 DeepSeek 模型的解決方案。能不能請(qǐng)您詳細(xì)介紹一下這項(xiàng)業(yè)務(wù)？

李開(kāi)復(fù)：當(dāng) DeepSeek 問(wèn)世的時(shí)候，中國(guó)迎來(lái)了自己的“ChatGPT 時(shí)刻”，我們可以稱之為“DeepSeek 時(shí)刻”。在中國(guó)春節(jié)假期期間，幾乎每個(gè)人都在談?wù)?DeepSeek，許多企業(yè)的 CEO 們節(jié)后返工馬上要求在自己的公司里用上 DeepSeek。他們發(fā)現(xiàn)，DeepSeek 是一個(gè)非常出色的模型， AI 的表現(xiàn)令人驚嘆，但它不能直接賦能公司里面的人力資源、財(cái)務(wù)和客服等場(chǎng)景，因?yàn)殚_(kāi)源模型還缺少連接企業(yè)數(shù)據(jù)庫(kù)、應(yīng)用程序等必要的中間件。

零一萬(wàn)物認(rèn)同 DeepSeek 所取得的巨大進(jìn)步，我們決定全力支持 DeepSeek ，讓它更好用。我們補(bǔ)充了那些缺失的中間件，開(kāi)發(fā)了易用的使用界面，使 DeepSeek 在企業(yè)中變得可用。這就是零一萬(wàn)物于本周一發(fā)布的產(chǎn)品——萬(wàn)智企業(yè)大模型一站式平臺(tái)，它在中國(guó)大陸和香港地區(qū)的初期反響相當(dāng)不錯(cuò)。

彭博社：這款“萬(wàn)智”新平臺(tái)背后的考量是怎樣的？

李開(kāi)復(fù)：許多人現(xiàn)在都下載了 DeepSeek，也很喜歡使用 DeepSeek。我有一個(gè) CEO 朋友問(wèn)他的員工：你們都用 DeepSeek 做什么，得到的答案是算命。這是很有意思的用法，你們或許都可以試一試，但目前的 DeepSeek 還無(wú)法直接深入到企業(yè)級(jí)應(yīng)用里去。

每個(gè)企業(yè)都有 ERP 和 CRM 數(shù)據(jù)庫(kù)，有員工數(shù)據(jù)等各種分門(mén)別類的企業(yè)內(nèi)部信息，企業(yè)需要的是具備行業(yè)縱深的知識(shí)型模型。比如彭博社會(huì)希望有一個(gè)金融知識(shí)淵博的模型，平安保險(xiǎn)集團(tuán)會(huì)希望有一個(gè)保險(xiǎn)專業(yè)知識(shí)的模型。零一萬(wàn)物發(fā)布的平臺(tái)，就是為了滿足這類各行各業(yè)的企業(yè)級(jí)需求。

這就像是如果我交給你 Windows 內(nèi)核，你可能不知道如何把它用起來(lái)，你會(huì)需要 Windows 軟件操作系統(tǒng)和應(yīng)用程序界面，還有配套的 Windows 系列應(yīng)用，Windows 內(nèi)核才能變得有用。零一萬(wàn)物推出新的萬(wàn)智企業(yè)大模型一站式平臺(tái)，就像是在為 Windows 內(nèi)核構(gòu)建可視化界面和操作中臺(tái)，讓 DeepSeek 的基座模型高效便捷地轉(zhuǎn)化為有效的生產(chǎn)力工具。

彭博社：選擇 DeepSeek 作為底層架構(gòu)而非美國(guó)模型，是否折射出中美AI競(jìng)賽格局變化？

李開(kāi)復(fù)：在中美，超大模型的預(yù)訓(xùn)練正在逐漸寡頭化，并且寡頭化的程度在不斷加大。雖然未必所有人都認(rèn)同，但我們的確同時(shí)看到開(kāi)源圈展現(xiàn)出壓倒性的優(yōu)勢(shì)，OpenAI 和 Anthropic 都相信自己還能訓(xùn)練出遠(yuǎn)超其他玩家的閉源模型。當(dāng)他們看到一個(gè)性能相近的開(kāi)源模型時(shí)，我想他們可能受到了沉重的打擊。OpenAI 在 2024 年的運(yùn)營(yíng)成本為 70 億美元，而 DeepSeek 的運(yùn)營(yíng)成本可能只有OpenAI 的 2%。

各家的模型都很優(yōu)秀，問(wèn)題不在于哪家模型性能高出 1%，而是 OpenAI 成本高昂的模型路線是否具有可持續(xù)性。OpenAI 每年花費(fèi) 70 億美元，面臨著巨額虧損。但現(xiàn)在出現(xiàn)了一個(gè)競(jìng)爭(zhēng)對(duì)手，將成本低廉?dāng)?shù)倍的開(kāi)源模型免費(fèi)開(kāi)放給市場(chǎng)，且這個(gè)競(jìng)爭(zhēng)對(duì)手資源充沛，目前看來(lái) DeepSeek 有足夠的資金儲(chǔ)備持續(xù)投入模型研發(fā)，并已經(jīng)有效地將計(jì)算成本降低了五到十倍。有了這樣一個(gè)強(qiáng)大的競(jìng)爭(zhēng)對(duì)手，我認(rèn)為 OpenAI 的薩姆·奧爾特曼可能夜里輾轉(zhuǎn)難眠。

彭博社：那么，你認(rèn)為 DeepSeek 真的挑戰(zhàn)了現(xiàn)有 AI 的商業(yè)邏輯嗎？另外，無(wú)論是在美國(guó)還是在中國(guó)，AI 的商業(yè)落地已經(jīng)相當(dāng)規(guī)?；?。在開(kāi)源模式占據(jù)優(yōu)勢(shì)的當(dāng)下來(lái)看，你認(rèn)為中國(guó)市場(chǎng)未來(lái)會(huì)跑出多少個(gè) AI 大模型？

李開(kāi)復(fù)：我認(rèn)為很大概率會(huì)收攏到三家，DeepSeek、阿里巴巴和字節(jié)跳動(dòng)。他們的模型會(huì)隨著時(shí)間不斷迭代進(jìn)化。

彭博社：這份名單上少了一些公司的名字。

李開(kāi)復(fù)：當(dāng)然，還會(huì)有許多其他公司希望構(gòu)建他們自己的模型，我們也會(huì)拭目以待看是否會(huì)有第四、第五個(gè)玩家。但即使在這三個(gè)中，我看好 DeepSeek 目前最具勢(shì)頭，這也是我們基于它推出新產(chǎn)品的原因。

但我也想指出，關(guān)于你問(wèn)到 AI 模型商品化的問(wèn)題，我不認(rèn)為今天的 AI 已經(jīng)商品化了。我認(rèn)為，吸納大量資金訓(xùn)練的基座模型開(kāi)始趨近同質(zhì)化，它們成本高昂、越來(lái)越難和開(kāi)源技術(shù)競(jìng)爭(zhēng)，但人工智能技術(shù)并沒(méi)有商品化。

就像 Windows 內(nèi)核、Linux 內(nèi)核這些已經(jīng)比較商品化，但 Windows 軟件系統(tǒng)并沒(méi)有商品化，微軟多年還靠此賺了很多錢(qián)。零一萬(wàn)物想要做的是填補(bǔ) AI 圈還不存在的 Windows，并使用 DeepSeek 作為我們類 Windows 的內(nèi)核之一。

彭博社：那么，你認(rèn)為現(xiàn)在的 AI 增加了什么價(jià)值？對(duì)于投資人來(lái)說(shuō)，在思考這個(gè)行業(yè)的價(jià)值增長(zhǎng)點(diǎn)時(shí)，具體有哪些主要的經(jīng)濟(jì)因素？而為什么你認(rèn)為只有三個(gè)玩家的空間？

李開(kāi)復(fù)：的確可能還有第四、第五個(gè)，但在多數(shù)市場(chǎng)中，大部分的終局最后將以寡頭局面收?qǐng)?。第一名賺了很多錢(qián)，第二名收支平衡，后頭幾名多半不能算是成功。在大模型的賽道中，由于底層模型很難直接轉(zhuǎn)化為商業(yè)模式，所以情況更加復(fù)雜。但一旦你是最終贏家，就有許多方法可以變現(xiàn)。近期我們可以看到的變化是，中美兩地的投資人幾乎不再下注更多更貴的底層模型公司。他們更愿意投資 AI 應(yīng)用、消費(fèi)者應(yīng)用、AI 基礎(chǔ)設(shè)施類型的創(chuàng)新企業(yè)，還有企業(yè)應(yīng)用的廣大市場(chǎng)需求，未來(lái)每個(gè)應(yīng)用、千行百業(yè)都會(huì)被 AI 重塑，這些應(yīng)用能為企業(yè)提供真正的價(jià)值。所以，AI 相關(guān)的資本正轉(zhuǎn)往這些方向，而不再投入基座模型。

彭博社：半導(dǎo)體技術(shù)封鎖是否會(huì)成為制約中國(guó) AI 發(fā)展的一個(gè)限制因素？

李開(kāi)復(fù)：我們可以回顧過(guò)去來(lái)理解這個(gè)問(wèn)題。過(guò)去，中國(guó)總是能夠在底層技術(shù)方面迎頭趕上，并構(gòu)建出偉大的工程產(chǎn)品，而美國(guó)更擅長(zhǎng)取得前沿突破。我曾經(jīng)在《AI·未來(lái)》這本書(shū)中深入論述這點(diǎn)，這種情況目前來(lái)看仍在繼續(xù)。

如今我們看到 DeepSeek 展現(xiàn)出了極強(qiáng)的適應(yīng)性和競(jìng)爭(zhēng)力，再次證明了“需求是創(chuàng)新之母”這一觀點(diǎn)。因?yàn)樗麄儞碛斜让绹?guó)大廠少很多的 GPU，所以不得不設(shè)法把已有 GPU 的效率提高 10 倍。雖然現(xiàn)在很難去預(yù)測(cè)未來(lái)的走勢(shì)，但我會(huì)說(shuō)，正因?yàn)?DeepSeek 擁有計(jì)算資源比美國(guó)大廠少了幾個(gè)數(shù)量級(jí)的，反而被迫更加精益求精，倒逼出的技術(shù)創(chuàng)新使得他們的模型訓(xùn)練和推理成本比 OpenAI 和其他公司低五到十倍，讓全球刮目相看。

彭博社：大約幾個(gè)月前，在 2024 年的 10 月份我們?cè)?jīng)有過(guò)一次交流，當(dāng)時(shí)零一萬(wàn)物也推出了與 DeepSeek 相似的模型，并且早于 DeepSeek 發(fā)布V3 之前。您如何看待模型迭代的頻率？這種頻率是被什么因素主導(dǎo)？DeepSeek 的成功是一種趨勢(shì)使然，還是產(chǎn)品本身的魅力？今年，大模型行業(yè)又會(huì)給我們什么樣的驚喜？

李開(kāi)復(fù)：中國(guó)大模型行業(yè)的競(jìng)爭(zhēng)仍會(huì)熱火朝天好一段時(shí)間，最終可能有三個(gè)贏家，美國(guó)可能有四個(gè)贏家，但這并不意味著其他幾十家公司會(huì)放棄競(jìng)爭(zhēng)，所以我判斷那些正在快速發(fā)展壯大的公司會(huì)繼續(xù)高頻發(fā)布新的模型和產(chǎn)品。在我看來(lái)，xAI、DeepSeek 是發(fā)展速度最快的公司。同時(shí)，OpenAI、阿里巴巴、谷歌和 Anthropic 都會(huì)做出令人敬佩的研究工作。

盡管字節(jié)跳動(dòng)可能不在大多數(shù)人的視野里，但我會(huì)把它列在名單上。字節(jié)在大模型領(lǐng)域已經(jīng)投入了海量資源。在前沿模型探索上敢于投入大量資金，其中一個(gè)最合理的邏輯是掌握了大量的用戶，當(dāng)所投入的巨額成本分?jǐn)偟侥塬@利的大體量用戶平臺(tái)時(shí)，商業(yè)模式就能夠成立?；谧止?jié)跳動(dòng)擁有最多的用戶、最多元化的商業(yè)變現(xiàn)模式，我做了這個(gè)預(yù)測(cè)。

彭博社：百度是個(gè)讓我印象深刻的公司，但不在你的名單上。此外，中國(guó)政府對(duì)AI的支持，有哪些變化？

李開(kāi)復(fù)：我們都非常尊重并感謝百度在中國(guó)真正開(kāi)啟了Transformer革命。他們?cè)噲D聘請(qǐng)杰弗里·辛頓，后來(lái)延攬了吳恩達(dá)。我認(rèn)為這證明了李彥宏先生具有遠(yuǎn)見(jiàn)。

另一個(gè)關(guān)于政府支持的問(wèn)題。中國(guó)政府的一個(gè)重點(diǎn)工作方向被稱為“新質(zhì)生產(chǎn)力”。這意味著利用高科技創(chuàng)造生產(chǎn)力、降本提效，并通過(guò)高新技術(shù)的引入，將所有傳統(tǒng)行業(yè)轉(zhuǎn)變?yōu)楦哂芰透?jìng)爭(zhēng)力的行業(yè)，而 AI 正是其中的核心引擎。

彭博社：幾周前的民企座談會(huì)上，中國(guó)科技巨頭的 CEO 們與國(guó)家領(lǐng)導(dǎo)人同處一室，這種情形很少見(jiàn)。從這個(gè)角度來(lái)看，你認(rèn)為政府會(huì)有哪些具體的措施來(lái)支持這一目標(biāo)？

李開(kāi)復(fù)：我認(rèn)為目前中國(guó)地方政府都在非常積極地?fù)肀?AI，各地都在探索如何將 AI 應(yīng)用到當(dāng)?shù)氐膫鹘y(tǒng)優(yōu)勢(shì)產(chǎn)業(yè)中，用技術(shù)打造“新質(zhì)生產(chǎn)力”，進(jìn)而促進(jìn)城市 GDP 增長(zhǎng)，這是一個(gè)從上到下都在貫徹落實(shí)的方向。對(duì)于地方政府來(lái)說(shuō)，“新質(zhì)生產(chǎn)力”是令人振奮的、非常有效的新增長(zhǎng)點(diǎn)，零一萬(wàn)物和其他友商公司都開(kāi)放和地方政府及產(chǎn)業(yè)展開(kāi)緊密的合作探索。

彭博社：開(kāi)復(fù)博士，非常感謝您加入今天的對(duì)話，與你交談?wù)媸橇钊擞淇?，也祝賀零一萬(wàn)物萬(wàn)智企業(yè)大模型一站式平臺(tái)的發(fā)布。

以下為英文對(duì)話全文：

Q: This is the year where we start to see really AI being applied across a range of different industries. Let's get more on that now because we've got Kai-Fu Lee. He's the CEO and Founder at 01.AI and also Chairman of Sinovation Ventures. It's actually perfect timing because you've just launched an enterprise large language model that's also gonna be deploying DeepSeek more broadly. But talk us through exactly what it entails.

A: Yeah, well, I think China had it's ChatGPT moment when DeepSeek came out. We can call it "DeepSeek moment." Everyone iss aware of it over the Chinese holidays. Everyone is talking about it. And the CEOs came back to work saying "put DeepSeek to work at my company." And what they found out was DeepSeek is a fantastic model, amazing AI, but it doesn't have the middleware and the user interface that it takes to connect to corporate databases, to build applications, to make it useful for HR, Finance and Customer Service. So what 01.AI did was we saw DeepSeek has been making great momentum. And we decided to really bet on DeepSeek and build that missing middleware and UI so that DeepSeek can be made useful for corporations. That's the product we announced this Monday and we're getting fantastic reception in China and also in Hong Kong.

Q: Tell us more about that launch.

A: Yeah, so we launched on Monday. And what we talked about was many of you have DeepSeek now. You love to use it. In fact, the one CEO friend of mine asked his employees, what do you use it for? And the answer was fortune telling. By the way, that's a great thing to try for you. But it's not very deep into the industry for the companies. Every company has ERP and the CRM databases. They have employee records, they have their internal information. And they want the model to be more a generalist. They want it to be knowledgeable. Bloomberg would want a finance knowledgeable model, right? Ping An would want an insurance knowledgeable model. So our job is to really build that layer for that purpose.

It's sort of like if I gave you a Windows kernel that is the core operating system, you wouldn't know what to do with it. You need all the Windows layers and the application interfaces so that the Windows kernel can be useful. And we like to think that 01.AI provides that layer for DeepSeek, which is the underlying model and technology.

Q: And using DeepSeek as the underlying model. Obviously, you could have chosen other models instead, including from the U.S. What does that tell you about where the competition or the race for supremacy stands between China and the United States right now.

A: Well, clearly in both U.S. and China, the pre-training of a giant model has consolidated and is consolidating. And it's become clear that open-source will be the winner. There's so many that will not concede, right? OpenAI, Anthropic who build their businesses believing they can build a better closed model than everyone else. I think they got shocked when they saw a model as good, they would argue nearly as good. But either way, they're getting it for much much lower cost.

If you think about OpenAI's cost——$7 billion of operating cost in 2024. DeepSeek probably operated with 2% of the operating expense. So the issue really isn't whose model is 1% better. I think they're all very good, but the issue is whether OpenAI's model is even sustainable. I mean, you're spending 7 billion dollars a year making a massive loss. And here you have a competitor coming in with an open-source model that's for free. And that company also is infinitely lasting because this founder has enough money to fund it at a current level and has reduced the cost of computing by a factor of 5 to 10. So with that kind of formidable competitor, I think Sam Altman is probably not sleeping well.

Q: Well, you do think, I mean, obviously, DeepSeek would have really challenged some of the monetization strategies around AI. But at the same time, given the AI is so commoditized now and in China as well, with the open-source approach, how many AI models do you think that the Chinese market can actually sustain?

A: Well, I think probably, I'm guessing three: DeepSeek, Alibaba, and ByteDance. And they'll evolve over time.

Q: OK. There are a few many names missing on that list.

A: Well, there will be many other who aspire to build one. And we'll see if there's a fourth player. But even within these three, I think DeepSeek currently has the momentum and that's why we're embracing that. But I also want to point out that I wouldn't say AI is commoditized. I would say the underlying pioneer model, foundation model that is trained with a huge amount of money, is commoditized. It costs a lot and it's hard to really compete with open-source. But AI is not commoditized. It's like, you know, Windows kernel, Linux kernel, that's kind of commoditized. But Windows is not commoditized. Microsoft makes a ton of money. And what 01.AI wants to do is build Windows and use DeepSeek as our Windows kernel.

Q: And is that where you think the value-add comes now? In other words, for VC's thinking about where the value added is in this industry, it's no longer, for example, AI still presents that value. And I wanna understand, for example, economics and why you think, say "there's only room for three."

A: There may be room for four, but in most markets you end up with room only for two. The winner who makes a ton of money, the runner up who breaks even and everyone else who dies. And this one is exacerbated because it's hard to monetize the underlying pioneer model. But once you're the winner, there are many ways to monetize it, right? So that's the trick. For the VCs in both US and China are moving away from wanting to fund more pioneer models. They're happy to fund AI infrastructure, AI applications, consumer apps, because every app will be remade. And also enterprise apps that can provide real value add to enterprises. So AI funding is going gangbusters. It's just that nobody wants to fund pioneer models.

Q: Does China's innovation, though, come up against the headwind this year of greater restrictions on China's access to advance semiconductor technology? Is that gonna end up presenting a problem?

A: We can look at it historically. China is able to catch up and build great engineering products, while US is better at making breakthroughs. That was what I argued in AI Superpowers, and that continues to be the case.

Today, we actually see DeepSeek not only resilient, competitive, but also basically made true the statement that "necessity is the mother of innovation." Because they had so few GPUs, they had to make it 10 times more effective. I would say that DeepSeek actually benefited from having less resources and able to make training and inference at 5 to 10 times lower the expense incurred by OpenAI and others.

Q: You know, I remember so DeepSeek was, let's call it two months ago. I remember, which we had the chat and the show, I think was back in October, when you guys actually launch something similar months before DeeSeek. You guys came out, DeepSeek came out and then since DeepSeek came out that moment. Just the cadence and the frequency of all these ones coming up with their own iterations. Okay, I guess my question is, what do you make up the frequency? What is behind you think the frequency of these releases? And is DeepSeek the symptom or is it the cause of what's going on? In other words, what can we expect from the industry this year?

A: I think it's still ultra competitive. The fact that I think China will have 3 winners, US might have 4 winners at the end. It doesn't mean dozens of others will not want to compete. So I do expect frequent releases to continue, especially from the fast moving companies. And I think the fastest moving companies probably are xAI's Grok and DeepSeek. I see these are the fastest. And then I think OpenAI, Alibaba, Google, and Anthropic are, of course, all making respectable moves.

I listed ByteDance because they're not on the list by people, but we know how much they're spending. And I think one way you can justify spending a ton of money on pioneer models is because you have so many customers. When amortized over many customers, you can make the business case. And ByteDance has more customers and more ways to monetize than anybody. So I would make that speculation.

Q: Yeah, I think Baidu is one of the names that sort of stands out to me. Not on that list and maybe can share some reaction on that, but also government support from Beijing and exactly how do you see that also shaping up over the coming year?

A: We all respect and thank Baidu for really starting the Transformer revolution in China. They tried to hire Jeff Hinton and ended up hiring Andrew Ng. And I think that shows Robin has that kind of vision.

The Chinese government's new big direction is called new quality productivity. What that means is using high tech to create productivity, efficiency and really turn all the traditional industries into more profitable, more competitive by the use of high tech. And AI is at the very core of that.

Q: So our final question for you, it was rare that we saw the CEOs of China's big tech companies in the same room as the president, for example, weeks before that. And I'm wondering from tangibly, what measures do you think we could see from the government to be able to back up really what really see it as very good optics?

A: Yeah, I think right now the local governments probably are the most active, in trying to see. For instance, how would I apply AI to create this "new quality productivity" in these industries that will end up helping the growth of GDP in the city. And I think, having a single direction from the top, new quality productivity is very effective and invigorating to the cities. We're seeing a lot of interest for our company and our peer companies to work more closely with these local cities.

Q: Right. Kai-Fu, thanks so much for joining us. So it's a real pleasure to speak with you and congratulations on the release of the new platform as well.