US companies are delivering one of the strongest earnings seasons in years, yet the market's response has exposed a growing disconnect for investors who still treat a quarterly beat as the main event. By early August, 86% of S&P 500 companies that had reported second-quarter results exceeded earnings estimates, according to FactSet. That is well above the five-year average of 78% and the ten-year average of 76%.
However, the reward for clearing that hurdle has often been surprisingly thin. Bloomberg Intelligence found that companies beating revenue, earnings, or both generated roughly flat one-day excess returns on average during the season, while misses were punished more heavily. Western Digital, Datadog, SanDisk, and DaVita all beat on the top and bottom lines and still sold off.
This points to something more important for traders: the published consensus is increasingly only one hurdle. The market is also judging the expectations already embedded in a share price and, above all, whether management gives investors reason to raise or cut the numbers that come next.
The contrast between Datadog and Nvidia
The contrast between Datadog and Nvidia captures the problem. Datadog beat second-quarter estimates and raised its full-year outlook in August, but its shares still sank as investors focused on a slower growth rate implied for the next quarter. Nvidia also beat, but its shares surged after management went much further, projecting revenue growth of about 70% in fiscal 2028, far above the roughly 45% analysts had expected.
One failed to clear the market's invisible hurdle, while the other changed the earnings path investors were modelling years ahead.
When almost everyone beats, the beat tells you less
The 86% figure matters because it changes what an earnings surprise actually tells investors. FactSet said the share of S&P 500 companies reporting positive earnings surprises in the second quarter was running at its highest level since the second quarter of 2021. Companies were also beating estimates by unusually large margins. The aggregate surprise was distorted by large investment-related gains at Alphabet and Amazon, but even excluding those two companies, earnings were still 10.9% above estimates.
The bar had not been unusually easy either. Analysts raised the bottom-up S&P 500 earnings estimate for the second quarter by 3.4% between the end of March and the end of June. Historically, analysts tend to cut estimates as a quarter progresses. Over the previous five years, the average reduction had been 2%. Expectations rose before companies reported, and companies still beat them at a very high rate.
The difficulty is that investors were prepared for much of that strength. “A lot of the good news has been priced in,” Bank of America strategist Jill Carey Hall told Bloomberg in August. She said investors had already positioned for strong earnings, helping explain why the reward for subsequent beats was more muted than usual. This asymmetry matters to traders, as a beat may confirm what the market already believes, but a miss can destroy that belief.
Larry Adam, chief investment officer at Raymond James, highlighted the same pressure before the reporting season accelerated, writing in a July market note that “with earnings expectations already at lofty levels, guidance will be key.” The point is not that consensus has become useless. It remains the cleanest public benchmark for measuring the quarter. But when beat rates become extremely high, the question shifts from whether a company cleared that benchmark to what was not already anticipated.
The invisible hurdle sits above consensus
Earnings are normally presented as a contest between one actual number and one consensus number. Reality is messier. Consensus is an observable average of published analyst forecasts. The price of a stock reflects something broader: the expectations of mutual funds, hedge funds, retail traders, quant strategies, options markets, and investors who may have very different assumptions about growth, margins, and valuation.
“The key issue is the expectations gap: a company can beat analysts’ published forecasts and still disappoint the market if investors had it already priced in even stronger growth, margins, guidance or cash flow,” Jauregui told Invezz. Datadog is a useful case because it shows why even a beat and raise can fail. The cloud monitoring company reported second-quarter adjusted earnings of 65 cents a share against a 58-cent consensus and revenue of $1.12 billion versus expectations of about $1.08 billion. It also raised its full-year revenue and earnings guidance. The stock nevertheless fell sharply.
The problem was in the trajectory. Datadog guided third-quarter revenue to between $1.135 billion and $1.145 billion. That was above the published consensus, but the midpoint implied year-on-year growth of roughly 29%, a deceleration from 36% in the second quarter. The shares had more than doubled during 2026 before the report, leaving investors positioned for continued acceleration. In other words, Datadog beat the analysts’ number but did not beat the expectation embedded in the stock.
“That is why stocks can fall after apparently excellent results,” Jauregui said. SanDisk offered a more conventional version of the same trade. The company reported fiscal fourth-quarter revenue of $8.97 billion and adjusted earnings of $39.25 a share, both ahead of market estimates, but its shares fell after the current quarter outlook disappointed.
This article is for informational purposes only and does not constitute financial advice.
