注：本文原發(fā)于跨資產(chǎn)星球北京時間4月13日的內(nèi)容。部分內(nèi)容已刪減，完整版可在文末獲取。

過去的一段時間是科技板塊過去 50 年最弱的一段相對收益期之一，這一定價結果主要源于兩個新問題：1）hyperscaler 的 ROI 持續(xù)被市場質(zhì)疑，2）AI disruption risk 正在重寫很多科技子行業(yè)的終局假設。

Over the recent period, tech has gone through one of its weakest stretches of relative performance in the past 50 years. This pricing outcome is mainly driven by two new questions: 1) the market keeps questioning hyperscaler ROI, and 2) AI disruption risk is rewriting terminal-state assumptions across many tech sub-sectors.

不止如此，華爾街近期最火熱的討論還包括了對‘Magnificent 7 能否永遠跑贏市場’的系統(tǒng)性懷疑，資方的恐懼集中聚集在擔憂買到 AI 時代的 Kodak、IBM、Nokia、Blackberry 這種被新技術浪潮直接碾碎商業(yè)模式的舊巨頭。軟件行業(yè)尤其如此，過去被長期增長信仰和低貼現(xiàn)率支撐的終值價值，如今第一次被認真打折。

More than that, one of Wall Street's hottest recent debates has been the systemic doubt over whether the Magnificent 7 can outperform the market forever. Capital's fear is increasingly concentrated on buying the AI-era equivalents of Kodak, IBM, Nokia, or Blackberry, old giants whose business models were crushed outright by a new technology wave. Software is especially exposed: the terminal value that used to be supported by a long-duration growth belief and low discount rates is now, for the first time, being seriously marked down.

結果就是大科技內(nèi)部相關性下降、分化上升：

The result is that correlations inside big tech have fallen while dispersion has risen:

圖：不同主題交易籃子的戰(zhàn)中/?；?年初至今表現(xiàn)

上圖是一個分籃子的表現(xiàn)統(tǒng)計。光網(wǎng)絡、數(shù)據(jù)中心，存儲與內(nèi)存這些主題，不僅 YTD 表現(xiàn)最好，戰(zhàn)爭期間回撤也相對有限，停火后反彈又最快（很明顯是當前市場最愿意在回調(diào)中繼續(xù)加倉的主線），對應地，軟件、IT 服務、以及廣義的 AI at risk 方向，在沖突期間被明顯加空，?；鸷笊踔晾^續(xù)遭遇更強的做空。也就是說，科技內(nèi)部的一個存在至今的趨勢是鏟子們的確定性和實際應用上的不確定性，可以預期的是，這種情況在可見視野范圍內(nèi)仍會持續(xù)。

The chart above is a basket-by-basket performance breakdown. Themes such as optical networking, data centers, storage, and memory not only have the best YTD performance, but also saw relatively limited drawdowns during the war and then rebounded the fastest after the ceasefire. Clearly, these are the market's preferred core positions to keep adding on pullbacks. By contrast, software, IT services, and the broader AI-at-risk complex were heavily shorted during the conflict and kept facing even stronger selling after the ceasefire. In other words, a persistent trend inside tech has been certainty in the picks-and-shovels segment versus uncertainty around end applications. This dynamic is likely to remain in place for the visible future.

但與此同時，被做空的一批科技，和科技之外更廣義成長股里有沒有潛在錯殺對象？

At the same time, are there potential names being indiscriminately sold off among the shorted parts of tech and the broader growth universe outside tech?

可能也的確是有的...

There probably are...

以 GS 的 Rule of 10 籃子為例，其中，在剔除最被做空的軟件之后，那些非軟件成長股的中位數(shù)當前 P/E 29x，相對標普中位數(shù)溢價約 53%，同樣已經(jīng)接近過去十年區(qū)間底部；而市場一致預期 2027 年銷售增速仍是標普中位數(shù)的 3 倍。

Take GS's Rule of 10 basket as an example. After excluding the software names that have been hit the hardest on the short side, the median P/E of those non-software growth stocks is now 29x, about a 53% premium to the S&P median, which is likewise close to the bottom of the past decade's range. Yet the market still expects their 2027 sales growth to be roughly 3 times the S&P median.

那些真正高質(zhì)量的非軟件成長股，估值已經(jīng)降了不少，但成長性并沒有塌掉：

Those truly high-quality non-software growth names have already seen their valuations come down a lot, but their growth has not collapsed:

圖：體現(xiàn)出的是一個潛在錯殺的廣義成長擊球區(qū)域

如果未來只是溫和增長放緩而非深度衰退的宏觀背景成立，那么 secular growth 中尤其是非軟件成長股的部分將開始重新具備配置吸引力，這部分做的，是從沖擊敘事切到有真實增長、 有供給約束、且沒被AI直接替代的事情。

If the future macro backdrop turns out to be one of only a mild growth slowdown rather than a deep recession, then the secular-growth complex, especially the non-software growth segment, will start to regain allocation appeal. What this part of the market is doing is shifting away from pure shock narratives toward businesses with real growth, supply constraints, and no direct AI substitution risk.

部分已經(jīng)驗證增長，但估值明顯降溫的資產(chǎn)，

Some assets have already validated their growth, yet their valuations have clearly cooled,

在當前時間點其預期回報已經(jīng)要好于繼續(xù)追高少數(shù)核心高位對象：

and at the current point in time their expected returns already look better than continuing to chase a small number of crowded core winners at elevated levels:

（注意進退有據(jù)，不要搶跑抄底）！

(Stay disciplined on entries and exits; do not rush to catch the bottom.)

圖：AXON（日線圖，截至 260413）

圖：CVNA（日線圖，截至 260413）

圖：SMCI（日線圖，截至 260413）

圖：DASH（日線圖，截至 260413）

圖：ANET（日線圖，截至 260413）

完整內(nèi)容訂閱鏈接：https://t.zsxq.com/i90xY

本文所涉及一切內(nèi)容僅供分享參考，接收人不因閱讀本文而被視為我方客戶，也不構成向任何人發(fā)出出售或購買證券或其他投資標的的邀請，更不構成交易依據(jù)。市場有風險，投資需謹慎，本文所有信息及觀點均不構成任何投資建議；過往業(yè)績不代表未來表現(xiàn)，投資者應依據(jù)自身風險承受能力進行審慎評估、獨立決策并自行承擔全部責任。任何情況下，我方均不對因使用本文內(nèi)容而產(chǎn)生的任何損失承擔責任。相關數(shù)據(jù)可能存在滯后或誤差，我方不對信息的完整性及準確性承擔法律責任。