就業數據轉負,美股創新高——企業利潤已不靠雇人賺錢了

發佈日期: 作者:
就業數據轉負,美股創新高——企業利潤已不靠雇人賺錢了

7月非農就業數據出來:-2.3萬人。

市場預期是+15萬。這個差距不是小偏差,是方向完全反了。照教科書邏輯,就業走弱意味消費動能衰退,股市應該要跌。但8月7日,標普500收在7,758點歷史新高,費城半導體當週漲了2.56%。

這個反應本身就是一條線索。

市場在押注一件事

就業走弱,Fed升息的理由就少了一個。利率頂了,高估值資產就有空間繼續往上撐。這個邏輯不複雜——但它只解釋了「為什麼沒跌」,沒有解釋「為什麼還能創新高」。

要回答這個,得看企業財報。

S&P 500裡,已有85.2%的公司Q2 EPS優於市場預期,整體利潤年增達29.3%。這個數字不是小幅超標,是大幅加速。JPMorgan這季(Q2 2026)單季營收82.46億美元、EPS $7.59,比去年同期的數字明顯跳升。

就業市場在降溫。企業獲利在加速。這兩件事同時為真。

脫鉤發生在哪裡

過去的邏輯是:企業要成長,就得多雇人;多雇人,就業數字就漂亮;就業漂亮,消費力撐住,整個循環成立。這套機制在AI資本支出浪潮裡已經斷裂了一部分。

雲端算力、模型訓練、推論基礎設施——這些投資產生的產出,不需要按比例增加人頭。一家銀行導入AI處理授信流程,手續費收入可以成長,招募的放款專員不用跟著等比增加。廣告公司導入AI創意工具,提案產能可以翻倍,但團隊人數不用翻倍。

這就是為什麼企業利潤年增29.3%,就業數字卻在走弱。兩條線已經不在同一條方程式裡了。

這個訊號對市場的意義

傳統的「景氣衰退預警指標」大多錨定在就業市場:失業率上升、非農轉負,市場就應該開始定價衰退風險。但如果企業利潤的驅動力已經從「雇更多人→產出更多」移向「算力投資→效率提升→利潤擴張」,那非農數字作為景氣代理指標的可信度就在下降。

就業數據不是從此不重要——消費端的影響還在,家庭收入減少最終會反映在零售數字上。但在這波生成式AI的資本支出浪潮裡,企業財報和就業市場的連動已經鬆動了。

市場現在定價的,是「利潤加速＋利率頂了」的組合。這個組合能撐多久,取決於下一季財報能不能繼續交出85%以上的EPS超標率——以及AI資本支出的回報,什麼時候會被投資人要求兌現成自由現金流。

目前兩個條件都還沒破。所以高點繼續往上。

— 胡凱翊(Kevin)

本文為個人研究筆記,不構成任何投資建議。

Jobs Went Negative, Stocks Hit All-Time Highs — Here’s the Structural Break

The July nonfarm payrolls number landed at -23,000. The market consensus was +150,000. That’s not a miss — that’s the direction flipping entirely.

Conventional logic says weakening employment equals weakening consumer spending equals a risk-off market. But on August 7th, the S&P 500 closed at 7,758 — a new all-time high. The Philadelphia Semiconductor Index gained 2.56% on the week.

The market’s reaction is the clue worth following.

What the Market Is Actually Pricing

Weaker jobs = fewer reasons for the Fed to hike = rates have peaked = stretched valuations can hold. That logic explains why stocks didn’t fall. It doesn’t explain why they hit a record.

For that, earnings need examining.

Among S&P 500 companies, 85.2% beat Q2 EPS estimates. Aggregate profit growth came in at 29.3% year-over-year. JPMorgan posted Q2 2026 revenue of $8.246 billion with EPS of $7.59 — a sharp acceleration from the same period last year. These aren’t incremental beats. These are big numbers moving fast.

Employment is cooling. Corporate profits are accelerating. Both statements are true at the same time.

Where the Decoupling Happened

The old model was simple: growth requires headcount. More headcount means better payroll numbers. Better payroll means consumer spending holds up. The whole cycle connected.

That chain is fraying under the generative AI capex wave. Cloud compute, model training, inference infrastructure — these investments generate output without proportional headcount growth. A bank deploying AI in credit processing can grow fee income without adding loan officers at the same rate. A creative firm deploying AI tools can double proposal throughput without doubling the team.

That’s the structural break. Corporate profit up 29.3%. Nonfarm payrolls negative. Two numbers, two different equations now.

What This Changes for Market Signals

Most recession early-warning frameworks are anchored to labor: nonfarm turns negative, unemployment rises, the market starts pricing contraction risk. But if corporate profit growth is now driven more by capital efficiency than by headcount expansion, then payroll data loses some of its value as a leading economic proxy.

Employment still matters — households losing income eventually show up in retail numbers, and that feedback loop hasn’t disappeared. But right now, earnings and employment are no longer moving in lockstep the way they used to.

The market is currently pricing a very specific combination: profit acceleration plus a rate ceiling. How long that combination holds depends on whether the next earnings season can sustain an 85%-plus beat rate — and when investors start demanding that AI capex translate into free cash flow, not just EPS beats.

Neither condition has broken yet. So the highs keep coming.

— 胡凱翊(Kevin)

This article is a personal research note and does not constitute investment advice.

延伸閱讀

短網址分享
https://justfly.idv.tw/s/PFZ95zn