Software spend is weakening for three reasons, and the cycle is only the fourth

UBS put the question of why software spending has weakened to 12 of its Fortune 500 CIO and CTO contacts, and published the initial unvarnished feedback ^[3]. The answers matter because the sector has already de-rated materially over the last few months ^[1] on a cause that was not obvious ^[2], and because the ranked list the CIOs produced describes a deliberate cleanup rather than a soft quarter.

There was no single explanation, according to UBS, but a combination of issues ^[4]. First and most widely acknowledged: a more concerted rationalization of software spend, especially at the SaaS layer, in reaction to what respondents called "excessive" price increases and following an over-investment phase ^[5]. Second: AI crowding out other software spend, both because dollars have to be freed up and because CIOs are hesitating or de-prioritizing while they work out what AI does to a given category ^[6]. Third: a mix shift inside budgets, with some SaaS projects de-prioritized and dollars moving to cloud infrastructure, cyber-security and open source ^[7]. Macro uncertainty, rates and weak consumer spend, showed up fourth, further down the list than many might expect ^[8]. Not one of the CIOs cited unusual 4Q-1Q seasonality ^[9], which is to say zero of the 12 ^[2]. Three of the four causes on the list are structural ^[1].

The uncomfortable part for vendors is the mechanism. UBS says the repeated message was that customers have over-bought, suffering from enterprise license agreement fatigue, and that software firms over-played their hand on price ^[10]. Buyers are now actively rationalizing in a still-tight budget backdrop ^[11]. The comparison UBS draws is exact: companies are doing to their SaaS budget what they spent the last 18 months optimizing away in AWS and Azure ^[12]. That earlier cleanup took four to five quarters, and with it largely finished at many large enterprises, attention has moved to other budget lines ^[17].

Application software is the target because it is where the fat was. Enterprises historically overprovisioned there relative to infrastructure, data and security ^[13]. It is also where price increases were sharpest, in part because SaaS firms facing weaker customer headcount growth, and therefore weaker seat expansion, tried to offset it through pricing and bundling ^[14]. And it is where the AI stories are still evolving or unclear ^[15], which makes deferral cheap. Timing may reflect the start of a new year, or a reaction to aggressive sales pushes at 4Q renewals ^[16]. UBS also raises the possibility that some customers are only now moving AI work from small proofs of concept into production, so the crowding-out effect is just starting ^[18].

The distinguishing tests UBS is applying to individual names are the same ones an operator should apply to their own vendor list: how much shelfware exists, which suppliers pressed hardest on pricing and bundling, and which are premium-priced versus low enough per seat to be rationalized last rather than first ^[19]. Adobe is offered as the counter-example, not immune but possibly bucking the trend in its May quarter given less over-provisioning in creative software, no real competition, and a credible AI story in Firefly ^[20].

What to watch: whether renewal negotiations in the coming quarters produce seat reductions rather than deferrals, since seat expansion is already weak ^[14]; whether the money leaving SaaS keeps landing in infrastructure, security and open source ^[7]; and whether AI moving into production widens the squeeze on non-AI lines ^[18]. If macro were the driver, seasonality would have been mentioned by someone ^[9].