Dell remained about 19% above its pre-earnings price while Credo was down more than 22%, revealing a widening divide between durable earnings winners and stocks whose initial losses continue to deepen.Recent earnings reactions have turned more defensive. Many stocks that initially fell after reporting have struggled to recover, while several suffered further losses during the following session.However, the weakness is not universal. Most of the recent earnings winners in investingLive’s sample remained above their pre-report prices, even after pulling back. That distinction matters for investors: a rebound does not necessarily erase an earnings disappointment, and a pullback does not necessarily destroy an earnings winner.Key takeaways from the latest earnings reactionsRecent results were defensive: Of 24 companies with usable completed reactions on September 10, seven rose and 17 fell.Downside moves were unusually large: Six companies fell beyond their options-implied ranges, compared with two that exceeded expectations to the upside.Most initial losses persisted: Fourteen of 15 stocks that initially fell on September 9 remained below their pre-event prices one session later.Most winners preserved their gains: Twelve of 14 initial winners were still above their pre-event prices, despite widespread next-session pullbacks.Oracle offered encouragement after hours, but its recovery attempt still requires confirmation during regular trading.Recent earnings reactions have become more defensiveThe September 10 completed-reaction group contained 24 companies with usable price reactions. Only seven rose, while 17 declined.Among the 18 companies for which options-implied expected moves were available, six fell beyond the ranges implied before earnings. Only two exceeded those expectations to the upside.The market-cap-weighted reaction within that subset was approximately -7.6%. This figure is a calculation based on the companies in the earnings sample. It is not a return for the broader stock market or a major index.Several company-level declines were substantial:Navan: down 21.75%Cooper Companies: down 14.67%American Eagle: down 13.97%The group was not dominated by the largest U.S. mega-cap companies, so it would be an overstatement to call this a market-wide earnings collapse. Still, the balance of reactions clearly became more defensive.What happened after the first earnings move?The more useful finding comes from following the stocks beyond their initial reporting-day reactions.Among 30 companies whose first regular-session earnings reactions occurred on September 9, 14 of the 15 initial losers remained below their pre-event prices by the September 10 close. Eight fell even further during the next session.That suggests the market was not quickly forgiving many of the disappointments.The picture for winners was more resilient. Twelve of the 14 initial winners remained above their pre-event prices, although 11 pulled back during the next session.This is an important distinction. A stock can decline the day after an earnings rally without invalidating the original positive repricing. Likewise, a stock can rebound after a large earnings decline while remaining substantially damaged.Previous earnings reporters Dell, Palo Alto, Broadcom and Credo show four different pathsInvestingLive reviewed historical closing prices through September 10 for four closely watched technology stocks.Dell: A durable earnings winnerDell initially rose 15.81% and remained 19.20% above its pre-earnings close through September 10.Its recent pullback may have looked negative when viewed as a single daily move, but the broader earnings story remained constructive. The stock had preserved and extended its original gain.For investors, the next question is whether Dell can stabilize while continuing to defend that positive repricing.Palo Alto Networks: A real but incomplete recoveryPalo Alto initially fell 9.28%. By September 10, it was still 6.52% below its pre-earnings close.The improvement was genuine, but the stock had recovered only about 30% of its initial dollar loss. Describing Palo Alto as fully recovered would therefore overstate the repair.Broadcom: A recovery that failed to holdBroadcom’s first reaction was a 2.74% decline. It subsequently closed above its pre-earnings baseline on September 8, but fell below that level again by September 10.It remained approximately 1.75% below its pre-event close at the end of the review period.Broadcom demonstrates why the route a stock takes matters. Briefly reclaiming a key reference price is not the same as holding above it.Credo: Persistent post-earnings weaknessCredo initially fell 20.04% and was approximately 22.42% below its pre-earnings close by September 10.An intervening rebound did not produce a lasting recovery. The original loss ultimately expanded, making Credo the clearest example of persistent downside pressure in this four-stock review.There was one counterweight to an excessively bearish interpretation. Dell, Palo Alto and Broadcom all outperformed the Nasdaq-100-tracking QQQ after their respective initial reaction-day closes through September 10. Credo underperformed.That is a targeted comparison of four stocks, not evidence of a broader technology-sector recovery.Why five-session performance can be misleadingThe wider September screen contained 154 companies with identifiable completed reaction sessions. Only 36.4% had positive rolling five-session returns, and the median return was approximately -2.9%.However, this sample was heavily concentrated. Technology represented approximately 87% of its market value, while Broadcom, Dell and Palo Alto together accounted for roughly 73%.There is also an important timing issue. A rolling five-session return may include trading that occurred before the earnings announcement. Depending on the reporting date, it can also omit some or all of the initial reaction.That makes five-session performance useful as a broad momentum screen, but not as definitive evidence that a particular earnings gain or loss survived. Historical prices measured directly from the pre-earnings close provide a cleaner company-level comparison.Oracle offers encouragement, but not confirmationOracle rose approximately 4.13% after the September 10 close, while Adobe fell about 2.35%.Oracle’s positive reaction challenged the recent defensive pattern, but the move was considerably smaller than the approximately 12.33% change implied by options before the event.A smaller-than-expected move describes the magnitude of the reaction. It does not, by itself, prove that Oracle’s financial results disappointed investors.The wider after-hours sample showed an approximately 3.2% market-cap-weighted gain, but Oracle represented roughly three-quarters of the group’s capitalization. Excluding Oracle, the remaining companies were only modestly positive.The evidence cutoff for this analysis is September 10. Oracle’s after-hours rise had not yet been tested during the September 11 regular session.What stock traders and investors can watch nextDell’s pullback: If Dell stabilizes while remaining well above its pre-earnings price, that would strengthen the argument that its latest weakness is consolidation within a durable positive repricing. A return toward the original baseline would weaken that interpretation.Partial repair versus continued deterioration: Palo Alto and Credo provide a useful comparison. Palo Alto has repaired part of its initial loss, while Credo’s damage expanded. Traders should still avoid assuming that a large existing decline automatically creates an attractive short entry.Oracle’s regular-session test: If Oracle retains much of its after-hours gain and other reporters also attract buyers, the recovery signal would become more persuasive. If Oracle fades or remains the only major source of strength, the broader defensive interpretation would remain intact.The practical takeaway is to focus on stock selection and follow-through. Recent earnings reactions should not be treated as one uniformly bullish or bearish trade.The earnings backdrop remains defensive because much of the downside damage is sticking. However, durable winners such as Dell and selective recoveries such as Palo Alto argue against calling it a universal selloff. The next useful signal is whether the latest rebounds last, not simply whether they happen.This analysis is based on investingLive’s review of earnings-calendar and historical share-price data through the September 10, 2026 regular-session close. September 10 after-hours reactions are identified separately. This article was written by Itai Levitan at investinglive.com.