Enterprise software, long the boring backbone of corporate IT budgets, is having its most volatile year in decades. Investors are trying to separate the products AI could make obsolete from those too deeply embedded to replace. That divide is creating some of the sharpest dislocations in the market right now.
Chips up, software down
The pattern has become almost mechanical. When chip and networking stocks rally on strong earnings, software stocks sell off.
Palantir ($PLTR) and Microsoft ($MSFT) both slid in premarket trading recently as AI infrastructure names surged. Salesforce, ServiceNow, and Workday followed them lower.
The logic is simple: chips power the large language models that investors believe will hollow out the need for expensive enterprise software subscriptions.
But the tape is not one-directional. PLTR and MSFT had both surged over the prior three months when chip stocks looked overvalued. The sector swings hard in both directions depending on which AI narrative dominates the week.
The earnings do not match the fear
The SaaSpocalypse debate intensified after Airtable agreed to be acquired for under $1.3B — roughly one-tenth of its 2021 peak valuation of nearly $12B.
HubSpot and Datadog both dropped on earnings. Datadog said its largest AI client cut usage since June.
Then the mood shifted. Atlassian surged to its best single day since going public after reporting its most profitable quarter since 2021. Twilio and Cloudflare also posted big beats.
The iShares Expanded Tech-Software Sector ETF plummeted 24% in the first quarter, but has since recovered to down just 3% for the year while the Nasdaq is up 15%.
"People were pencils down in the space. They just didn't feel like it was worth their time to meet with software companies back then, and that was even high-quality companies."
Matt Hedberg, RBC Capital Markets
The split tells you something important. Vibe-coding tools from OpenAI and Anthropic are a real threat to generic software. But software deeply embedded in regulated workflows is proving far harder to dislodge.
Private equity is the hidden pressure point
The stress is not just in public markets. AI looms over software buyouts structured on the assumption of steady subscription revenue growth.
Private equity firms loaded these companies with debt in the low-rate era, and that debt now sits on balance sheets facing higher interest costs and an AI disruption narrative that makes refinancing expensive.
Across leveraged loans, junk bonds, and private credit, there is over $150B of software company debt coming due between now and the end of 2029, according to Barclays.
Thoma Bravo's Medallia is already a cautionary tale: lenders took control after the loan balance swelled to ~$2.8B, and BDC funds marked the debt as low as 54 cents on the dollar.
Fitch Ratings found that 9% of the software companies it rates are at high risk of AI disruption, while half are at low risk.
Public software companies carry far less debt relative to their size, giving them more room to absorb the disruption. The danger zone is leveraged private companies with floating-rate loans and maturing debt walls.
Where the dislocation is sharpest
For investors looking at beaten-down public names, the key distinction is between software AI can replace and software running the operational core of a business.
ServiceNow's IT management tools and Veeva Systems pharma-specific clinical trial software both sit in that harder-to-replace category. ServiceNow trades at a forward multiple in the mid-20s, well below the ~60x it historically commanded.
Adobe is the harder call. AI-native rivals are chipping away at content creation tools, and analyst opinion is split between downgrades and double-digit upside targets. At a forward multiple near 10, the market is pricing in a slowdown.
On the infrastructure side, Morningstar flags Broadcom as a key secondary AI compute vendor to Nvidia, and Arista Networks as a direct beneficiary of AI networking buildout. Both sit in the hardware layer that benefits when software fears spike.
The sector is not going to zero. But the AI disruption is landing unevenly, and the debt sitting inside private software buyouts may be the part of this story that has not fully shown up in public markets yet.