Why CRM, NOW, TEAM, and MNDY Keep Falling While the S&P 500 and Nasdaq Hit Record Highs
The headline numbers describe a bull market. The S&P 500 trades near 7,400, the Nasdaq Composite sits above 25,000, and the Dow closed at a fresh record above 51,500 earlier this month. By the only measure most people check, 2026 has been a good year to own stocks.
Now look inside the index. Salesforce is down roughly 31% on the year. ServiceNow is off about a third. Atlassian has lost close to 28%, and Monday.com has been cut nearly in half. Four enterprise software franchises, all growing revenue at double-digit rates, are bleeding while the averages that contain them print all-time highs. This is not noise. It is the defining feature of the 2026 tape, and it has a cause.
The Index Is Not the Market
The first thing to understand is that the record highs are narrow. This is a two-speed market wearing a single jersey. Strip the index apart and the divergence is violent: the information technology sector is actually down on the year despite the broader gains, dragged by a collapse in software and services of more than 20% — the worst start that group has ever recorded. Semiconductors and hardware, meanwhile, are up. So is energy. So are materials, industrials, and the rest of the long-neglected cyclical complex.
The index rises because a handful of AI infrastructure winners and a broad rotation into non-tech cyclicals are doing the heavy lifting. The average obscures the dispersion beneath it. When a commentator says “the market is up,” what they mean is that the dollar-weighted basket is up. The median software stock is having one of the worst years of the past decade. CRM, NOW, TEAM, and MNDY are not falling against the market. They are the part of the market that is falling.
Same Catalyst, Opposite Sign
Here is the mechanism that ties it together, and it is almost elegant. The force lifting the index and the force sinking these four stocks are the same force. Artificial intelligence is both.
The index is up because AI demand is real and it flows first to the people selling compute. Chipmakers, hardware vendors, networking, power, and the hyperscalers building the capacity have absorbed a wave of capital expenditure measured in the hundreds of billions. That spending is a revenue line for the picks-and-shovels names, and the market has rewarded them.
The same AI is a threat to application software. The bear thesis against seat-based SaaS is that autonomous agents reduce the headcount these companies bill against. If an agent resolves the ticket or closes the deal, the seat it replaces does not renew. So the identical catalyst that makes a chipmaker’s order book swell makes a software vendor’s pricing model look obsolete. Investors are not confused. They are funding the AI trade by selling the companies they believe AI will disrupt, and the four names here sit squarely in the crosshairs. The capital lifting the index is, in part, capital raised by liquidating these very stocks.
The Rotation Is Mechanical, Not Emotional
This is not panic. It is reallocation, and reallocation has its own gravity. When large funds decide software is the funding source for an AI-hardware and cyclical overweight, the selling is systematic and price-insensitive. It does not wait for a bad earnings report. It sells the group because the group is the source of cash for the trade the manager actually wants on.
The pressure has been compounded by supply. The market has been digesting an enormous calendar of new issuance, headlined by the largest debut in history, which pulls liquidity toward the new name and away from the incumbents. When fresh, exciting paper competes for the same dollars, the boring profitable software franchise loses the bid. None of this requires anyone to dislike the business. It requires only that the marginal dollar has somewhere better to go.
Why Beating Earnings Did Not Help
The most revealing detail of the year is that all four companies reported good quarters and fell anyway. Salesforce crushed its earnings estimate and sold off. ServiceNow grew subscription revenue 22% and dropped. Atlassian beat by roughly a third and remains deep in the red. Monday.com topped expectations and got cut.
That tells you the market is not repricing the quarter. It is repricing the multiple. The debate is no longer about this year’s revenue, which everyone concedes is fine. It is about the durability of the model that produces next decade’s revenue. When the argument moves from earnings to terminal value, a single good print cannot fix it, because the print was never the question. The multiple compression reflects a lowered conviction in the longevity of seat-based growth, and only sustained evidence — not one beat — can rebuild that conviction. This is why the stocks can be cheap on every backward-looking metric and still fall. The market is discounting a future it has decided to distrust.
What Breaks the Divergence
A gap this wide eventually closes, in one of two directions. Either the software names prove the disruption fear is overblown, or the AI-hardware leaders prove the spending cannot be monetized and the whole complex re-rates downward together.
For the four here, the bull resolution is specific and measurable. It requires net revenue retention to stabilize, the new consumption and work-unit pricing models to demonstrably offset seat erosion, and AI products to convert from pilot to paid at scale. ServiceNow showing half its new business is no longer seat-based, Salesforce scaling Agentforce past the billion-dollar mark, Atlassian reporting that AI customers expand twice as fast — these are the early data points that, if they compound, force the market to admit the thesis was wrong. Until then, the divergence persists, because the market does not reward what might be true. It rewards what has been proven.
The record highs and the software wreckage are not a contradiction. They are two readings of the same instrument. The index is telling you the market believes in AI. These four stocks are telling you the market has not yet decided whether they survive it. Both statements are true at once, and the year will be decided by which one the next four quarters confirm.