SaaS Incumbents Use Buybacks to Defend Valuations

Salesforce is the clearest case. The company issued $25 billion of senior notes in March to fund a record accelerated buyback, lifting senior notes from $8.5 billion to $33.3 billion in one quarter. It then used its Anthropic stake to mark up earnings by $2.7 billion in the August quarter, helping drive a two-day stock jump of 22.58% after Marc Benioff unveiled “Claudeforce” at Dreamforce.

The market quickly faded that bounce. Salesforce shares closed at $234.02 on Sept. 25, below the Sept. 3 peak of $263.97 and near where the stock was trading before the latest AI rebrand. The 50-day moving average sat at $213.37, while the 200-day moving average was $200.36, showing the stock has recovered from its spring washout but is still trading well below the summer spike that followed the Anthropic mark-up.

The broader sector is showing the same pattern. HubSpot authorized $2 billion of buybacks within seven months. Workday added a fresh $4 billion repurchase while cash and marketable securities fell. ServiceNow executed a $2 billion accelerated repurchase. Adobe approved a new $25 billion authorization running to 2030. IBM is still responding to AI pressure with legacy messaging rather than a frontier model or a large-scale ecosystem bet.

The economic logic is straightforward: when software vendors cannot justify premium growth valuations, buybacks become a substitute for a real AI strategy. But repurchases do not create product differentiation, and they do not stop customers from moving to AI-native interfaces that can sit on top of, or replace, legacy systems in a weekend.

That is why investors care. The valuation reset across SaaS may be only partly about rates and credit conditions; it is also about demand destruction from cheaper, better AI tools. The 10-year Treasury yield was 4.96% on Sept. 22 and Treasury bond sentiment remained elevated, while high-yield spread readings near 2.68 suggest markets are not pricing a full credit event. Even so, the combination of heavy debt, slower growth and rising AI substitution risk leaves less room for error.

The narrative connecting the facts is that the sector’s defensive playbook has changed. Instead of reinvesting for a new platform cycle, incumbents are borrowing to buy time. That can support earnings for a few quarters, especially when stakes in companies like Anthropic can be marked up, but it also leaves shareholders more exposed if the AI replacement cycle keeps accelerating.

For investors, the next catalyst is whether the latest Dreamforce-era bounce holds or fades as the market looks past repurchases and asks which SaaS names actually own the AI substrate. If the answer remains “none,” the debt trap Stephen Messer described looks less like commentary and more like the base case.