Why great earnings keep producing falling share prices
28th August 2026
A pattern is repeating often enough now that it is worth explaining properly, because it confuses a lot of people who are watching their portfolios and cannot square what they are seeing.
This week alone gave two clean examples. Nvidia reported a beat and raise on Wednesday, then closed the following session down. Marvell followed on Thursday with record quarterly revenue of $2.739 billion, up 37% year on year, third quarter guidance of $3.15 billion against a Wall Street consensus of $3.03 billion, and adjusted earnings guidance of $1.10 a share against a consensus of $1.07. By any conventional measure, that is a strong quarter with a confident outlook attached. The shares fell 7.6% in after hours trading, wiping out roughly $16.1 billion of market value in minutes.
Neither of these companies disappointed. Both beat what analysts expected and both told the market things were getting better, not worse. The share price reaction had almost nothing to do with the quarter that had just happened and everything to do with the quarter the market had already assumed would happen.
This week alone gave two clean examples. Nvidia reported a beat and raise on Wednesday, then closed the following session down. Marvell followed on Thursday with record quarterly revenue of $2.739 billion, up 37% year on year, third quarter guidance of $3.15 billion against a Wall Street consensus of $3.03 billion, and adjusted earnings guidance of $1.10 a share against a consensus of $1.07. By any conventional measure, that is a strong quarter with a confident outlook attached. The shares fell 7.6% in after hours trading, wiping out roughly $16.1 billion of market value in minutes.
Neither of these companies disappointed. Both beat what analysts expected and both told the market things were getting better, not worse. The share price reaction had almost nothing to do with the quarter that had just happened and everything to do with the quarter the market had already assumed would happen.
What "priced in" actually means
A share price is not a scoreboard for the last three months. It is a running bet on everything the company is expected to earn for years into the future, discounted back to today. By the time a widely covered company like Marvell reports, that bet has already absorbed months of analyst upgrades, management commentary, and investor enthusiasm. The share price on the day of the results already contains an assumption about how good the numbers will be.
When the actual result matches that assumption, nothing changes and the stock goes nowhere. When it beats the assumption by a wide enough margin, the stock goes up. But when a company merely delivers a good quarter that management had already all but promised, and the market had already priced that promise in weeks in advance, there is no fresh information left to move the stock upward. Sometimes there is a specific detail in the result, a shrinking margin, a customer concentration, a slightly softer guide on one line, that gives the market a reason to sell into the good news rather than buy it. In Marvell's case, non-GAAP gross margin guidance for the next quarter came in 90 basis points lower than the prior quarter, even as revenue guidance rose. That was enough.
A share price is not a scoreboard for the last three months. It is a running bet on everything the company is expected to earn for years into the future, discounted back to today. By the time a widely covered company like Marvell reports, that bet has already absorbed months of analyst upgrades, management commentary, and investor enthusiasm. The share price on the day of the results already contains an assumption about how good the numbers will be.
When the actual result matches that assumption, nothing changes and the stock goes nowhere. When it beats the assumption by a wide enough margin, the stock goes up. But when a company merely delivers a good quarter that management had already all but promised, and the market had already priced that promise in weeks in advance, there is no fresh information left to move the stock upward. Sometimes there is a specific detail in the result, a shrinking margin, a customer concentration, a slightly softer guide on one line, that gives the market a reason to sell into the good news rather than buy it. In Marvell's case, non-GAAP gross margin guidance for the next quarter came in 90 basis points lower than the prior quarter, even as revenue guidance rose. That was enough.
Why this is happening more often in 2026
This is not a new phenomenon, but the frequency has increased this year for a specific reason. A small number of companies now carry an outsized share of expectations for the entire AI infrastructure trade. When expectations are already stretched to the point where merely good news counts as disappointing, that is not a flaw in any one company. It is a description of what happens near the top of a valuation cycle in a narrow part of the market.
Nvidia's results this week fit the same pattern. A beat and raise followed by a lower share price on the day, in a quarter with double digit revenue growth and guidance ahead of estimates. The specifics differ from Marvell but the mechanism is identical: the bar for a positive reaction had already been set above what even a genuinely strong quarter could clear.
What this means for anyone holding these companies, directly or through a fund
None of this means the businesses are performing badly. Marvell posted 37% revenue growth and improved operating income by 58.5% year on year. Nvidia's own numbers this week showed similar strength. The share price move on the day tells you about sentiment and expectations, not about the underlying company.
It does mean that following daily share price moves as a verdict on business quality is a poor way to judge an investment. A stock falling on strong results is not evidence the company is struggling. It is evidence that the market had already decided, in advance, exactly how good the news needed to be, and the news, however good, did not clear that specific bar.
For anyone holding concentrated positions in a small number of AI-linked names, either directly or through a passive index fund where the weighting has crept upward without a decision being made, this is worth sitting with. The more a share price depends on flawless execution against an already elevated set of expectations, the more violent the reaction when reality, even good reality, falls even slightly short of that bar.
This is not a new phenomenon, but the frequency has increased this year for a specific reason. A small number of companies now carry an outsized share of expectations for the entire AI infrastructure trade. When expectations are already stretched to the point where merely good news counts as disappointing, that is not a flaw in any one company. It is a description of what happens near the top of a valuation cycle in a narrow part of the market.
Nvidia's results this week fit the same pattern. A beat and raise followed by a lower share price on the day, in a quarter with double digit revenue growth and guidance ahead of estimates. The specifics differ from Marvell but the mechanism is identical: the bar for a positive reaction had already been set above what even a genuinely strong quarter could clear.
What this means for anyone holding these companies, directly or through a fund
None of this means the businesses are performing badly. Marvell posted 37% revenue growth and improved operating income by 58.5% year on year. Nvidia's own numbers this week showed similar strength. The share price move on the day tells you about sentiment and expectations, not about the underlying company.
It does mean that following daily share price moves as a verdict on business quality is a poor way to judge an investment. A stock falling on strong results is not evidence the company is struggling. It is evidence that the market had already decided, in advance, exactly how good the news needed to be, and the news, however good, did not clear that specific bar.
For anyone holding concentrated positions in a small number of AI-linked names, either directly or through a passive index fund where the weighting has crept upward without a decision being made, this is worth sitting with. The more a share price depends on flawless execution against an already elevated set of expectations, the more violent the reaction when reality, even good reality, falls even slightly short of that bar.
The planning point
We are not going to tell anyone whether Marvell or Nvidia are good long-term holdings. That is not the purpose of this piece. The point is broader. If a portfolio's returns increasingly depend on a small number of companies clearing an ever-rising bar of expectations every single quarter, that portfolio carries a specific kind of risk that a simple diversification conversation should address before the next earnings season, not after a bad one.
You can book an initial call below.
We are not going to tell anyone whether Marvell or Nvidia are good long-term holdings. That is not the purpose of this piece. The point is broader. If a portfolio's returns increasingly depend on a small number of companies clearing an ever-rising bar of expectations every single quarter, that portfolio carries a specific kind of risk that a simple diversification conversation should address before the next earnings season, not after a bad one.
You can book an initial call below.
