I Just Returned From China. We Are Not Winning.

By Steven Rattner

NYT Feb. 10, 2026

A few weeks ago, I attended a dinner in New York where the conversation quickly turned to China. Some of the most seasoned experts were sharply divided: some backed President Trump’s hard-line, aggressive stance, while others advocated for a less confrontational, more traditional approach.

I am no expert on trade, but I have made investments in China for years and had just returned from a weeklong visit. Eventually summoning my nerve, I suggested that neither approach would work. China is just too formidable as a rival — as well as a critical manufacturing powerhouse — to be reined in by diplomacy or an aggressive shift in policy. The only real solution is to get our house in order and beat China at its own game.

The need to do so is only growing, because the commotion of Mr. Trump’s first year back in office has set America back. In addition to manufacturing, China is threatening America’s pre-eminence in a range of fast-growing sectors, including artificial intelligence and pharmaceutical drug development. For all of Mr. Trump’s tariff bluster, we are not winning this trade war. The Asian goliath powers on as the world’s largest exporter, its trade surplus having notched a record $1.2 trillion last year. That overall increase suggests that many Chinese goods are simply passing through middleman countries before reaching U.S. shores. Tariffs or no, everybody needs Chinese goods.

Consider cars. During my trip, I toured Xiaomi, a telecommunications equipment manufacturer that has moved aggressively into electric vehicles. In the lobby sat a yellow sports car that could easily be mistaken for a Porsche. But the price tag was about $40,000, not $200,000. And the performance was just as good.

China’s progress in A.I. has been stunning. It has more than twice as much generating capacity as we do, and some of its data centers pay half as much as ours for power. That has helped it develop products like Manus, an A.I. agent comparable to ChatGPT, which was sold to Meta for more than $2 billion shortly after my visit.

I visited a robotics company where what looked like plastic children’s toys scampered across the floor, demonstrating the firm’s progress toward building humanoids that could replace humans in certain tasks.

China’s pharmaceutical industry is also surging. It now licenses more drugs to other countries than the United States does. And while Mr. Trump has sought to cut spending on vital government functions like basic research, China has made these areas national priorities.

Competing against China will be difficult under the best of circumstances. Clearly we need to rethink our industrial policy — the way we can deploy our government resources to support strategically important industries, which is our version of state-directed capitalism.

We also need to invest far more in our own infrastructure, education and workforce development. And we must stop demonizing China and start respecting it as a formidable competitor.

The bottom line is this: We cannot wish China away. We cannot tariff it into submission. We cannot negotiate it into irrelevance. The only way to prevail is to outcompete it. And that starts with fixing what’s broken at home.

中文譯文

我剛從中國回來，我們沒贏

史蒂文?拉特納

NYT 2026 年 2 月 10 日

幾周前，我在紐約參加了一場晚宴，話題很快就轉(zhuǎn)向了中國。一些最資深的專家意見尖銳對立：有些人支持特朗普總統(tǒng)強硬、激進的立場，而另一些人則主張采取不那么對抗、更傳統(tǒng)的做法。

我并非貿(mào)易專家，但多年來一直在中國投資，且剛剛結(jié)束了為期一周的訪問。最終，我鼓起勇氣提出，這兩種方法都行不通。中國作為一個對手太過強大——同時也是至關(guān)重要的制造業(yè)強國——僅靠外交手段或激進的政策轉(zhuǎn)變都無法遏制它。唯一真正的解決方案是整頓我們自己的內(nèi)部事務(wù)，在其擅長的領(lǐng)域擊敗它。

這種需求正變得愈發(fā)迫切，因為特朗普重返總統(tǒng)職位第一年的種種亂象已經(jīng)讓美國倒退。除了制造業(yè)，中國還在人工智能和藥物研發(fā)等一系列快速增長的領(lǐng)域威脅著美國的領(lǐng)先地位。盡管特朗普大肆宣揚關(guān)稅政策，但我們并沒有贏得這場貿(mào)易戰(zhàn)。這個亞洲巨人繼續(xù)作為世界最大出口國穩(wěn)步前行，去年其貿(mào)易順差創(chuàng)下了 1.2 萬億美元的歷史紀錄。這一整體增長表明，許多中國商品只是通過中轉(zhuǎn)國然后抵達美國海岸。無論是否征收關(guān)稅，人人都離不開中國商品。

以汽車為例。此行中，我參觀了已大舉進軍電動汽車領(lǐng)域的電信設(shè)備制造商小米。大廳里停放著一輛黃色跑車，很容易被誤認為是保時捷。但它標價約為 4 萬美元，而非 20 萬美元，性能卻同樣出色。

中國在人工智能領(lǐng)域的進展令人震驚。其算力是我們的兩倍多，部分數(shù)據(jù)中心的電力成本僅為我們的一半。這助力其開發(fā)出了像 Manus 這樣的產(chǎn)品 —— 一款可與 ChatGPT 媲美的人工智能代理，在我訪問后不久，它就以超過 20 億美元的價格被 Meta收購了。

我還參觀了一家機器人公司，看起來像塑料兒童玩具的設(shè)備在地板上快速移動，展示了這家公司在研發(fā)可在特定任務(wù)中替代人類的人形機器人方面取得的進展。

中國的制藥行業(yè)也在迅猛發(fā)展。如今，它向其他國家授權(quán)的藥品數(shù)量超過了美國。當(dāng)特朗普試圖削減基礎(chǔ)研究等重要政府職能的開支時，中國已將這些領(lǐng)域列為國家優(yōu)先事項。

即使在最理想的情況下，與中國競爭也將十分艱難。顯然，我們需要重新思考我們的產(chǎn)業(yè)政策 —— 即如何調(diào)配政府資源以支持具有戰(zhàn)略意義的產(chǎn)業(yè)，這是美國版本的國家主導(dǎo)型資本主義。

我們還需要在自身的基礎(chǔ)設(shè)施、教育和人力資本發(fā)展上投入更多資金。并且，我們必須停止妖魔化中國，開始將其視為一個強大的競爭對手予以尊重。

歸根結(jié)底是這樣：我們無法讓中國消失，無法通過關(guān)稅使其屈服，也無法通過談判使其變得無足輕重。獲勝的唯一途徑是在競爭中超越它，而這首先要從修復(fù)國內(nèi)的問題開始。

作者" bdsfid="370">史蒂文?拉特納 現(xiàn)為紐約時報專欄作家，曾在奧巴馬政府時期任財政部顧問

" bdsfid="373">— THE END —

紫京講談

財經(jīng)領(lǐng)域創(chuàng)作者

3416篇原創(chuàng)