Salesforce has become a central figure in what many are now calling the SaaS-apocalypse — a generational reckoning for software-as-a-service companies whose pricing power was built on a labor-intensive era that AI is rapidly dismantling. The company sits in the crosshairs of investors, analysts, and former champions alike, with sentiment scores that place it near the bottom of its peer group. Understanding why requires looking past the headline numbers and into the structural problems facing the legacy CRM business.
A Revenue Model Built for a Different Era
The core issue is mechanical. Salesforce's billing and revenue are based overwhelmingly on a per-seat unit: the more employees at a client use Salesforce products, the more the client pays. That arrangement worked beautifully when enterprises were expanding their headcount. It works far less beautifully in an environment where AI is actively replacing people. Each automated workflow, each agent that handles what a human used to handle, removes a seat from the billing roster. Companies may become more efficient per remaining person, but the absolute number of seats shrinks — and that shrinkage attacks the heart of Salesforce's revenue model.
The company is aware of the problem and is experimenting with alternative pricing structures, including flex credits and conversation-based billing. The goal is to keep extracting high revenue per customer even as headcounts contract. But every model on offer points in the same direction: declining revenue from existing customers as AI compresses the human footprint they used to pay for.
A Slow Slide That AI Will Accelerate
The trajectory was already visible before AI hit full stride. Fiscal year 24 grew at roughly 11%. Fiscal 25 came in around 9%. Fiscal 26 looks like 7 to 8% once the acquisition of Informatica is stripped out. Sentiment-driven indicators have the underlying business running at roughly +5% on a year-over-year basis. The deceleration is steady, and AI is poised to accelerate it to the downside.
On a sentiment score where readings under 30 lean bearish and readings over 70 lean bullish, Salesforce sits at 17 — effectively tied for last place among major SaaS companies, with the likes of Dropbox keeping it company at the bottom. Microsoft, by contrast, scores 90, Adobe 89, and Datadog 85. The gap is not a rounding error; it reflects a market that increasingly sees clear winners and losers as AI sorts the software landscape.
The Hidden Threat: New Customers Never Arriving
Most investors understand the squeeze on existing revenue. Far fewer have thought hard about what is happening with new customers. This is arguably the bigger blind spot.
A wave of well-funded CRM-style startups is being built from scratch — by AI, for AI. They are cheaper, simpler, and natively designed to work with modern AI workflows, in contrast to a fifteen-year-old platform layering integrations onto a legacy foundation. New companies setting up their customer relationship management for the first time are not defaulting to the incumbent; they are picking the new entrants. The premium price tag that defined Salesforce's market position is now a deterrent for the next generation of buyers.
Software itself has been commoditized by AI. Building a polished CRM platform no longer requires the engineering army it once did. Distribution and network effects are the only durable moats left in software — the kind of moat a platform with four billion users enjoys, where every additional user makes the product more valuable to every other user. Salesforce has no such moat. It is a sales engine and a marketing engine wrapped around a software platform that, given enough funding, almost anyone could now build.
When Customers Become Competitors
The pressure is not only external. Salesforce itself reportedly did not hire new engineers over the past year, yet productivity in its engineering and development organization rose roughly 30%. That figure is probably conservative compared to what is happening elsewhere; in many development environments, productivity gains of 5x to 10x are now realistic. Large customers are seeing the same dynamic inside their own engineering departments — and noticing how much they pay Salesforce.
The natural next thought, especially for developers under pressure to justify their roles, is to build internally. A motivated team with modern tooling can replicate a meaningful subset of CRM functionality without the seat-based bill. Expect to see more of this. It represents a second front in the revenue assault: not only fewer seats from headcount shrinkage, but entire customer accounts being clawed back in-house.
When Showcase Customers Turn Bearish
The signals from the investor and customer side are striking. Bridgewater has exited its position. Starboard has exited its position. Bank of America has tagged the stock as an underperform. The irony is biting: Bank of America is one of Salesforce's marquee case studies, featured for results that included a 10% lift in sales and a 15% improvement in cross-sell. And yet that same institution's research arm has placed the company on its underperform list, citing fewer new customers, limited upsell, and weak AI monetization.
There is room for the right-hand-left-hand caveat at an organization of two hundred thousand employees — the investment side and the operating side may not be in perfect dialogue. But the fact that a flagship reference customer's analysts are skeptical tells a story on its own. Bank of America has also said something that is being echoed across many large enterprises: every time an employee leaves, they reassess whether to refill the role. That is a structural commitment to fewer seats over time.
The Math on Alternative Pricing Doesn't Save It
Salesforce's pivot toward conversation- or productivity-based billing runs into another problem: the underlying inference costs for large language models are falling by roughly an order of magnitude per year. The unit economics of AI-flavored services are being commoditized as fast as the seat economics are being eroded. Pricing power has nowhere stable to land.
This is why a forward P/E in the 13 to 14 range, which some interpret as a maturing company shifting into cash-generation mode, may not be the bargain it looks like. Salesforce is buying back significant amounts of stock — by some accounts deploying roughly half of its capital that way — rather than reinvesting heavily for growth. That is the behavior of a company whose own leadership may not see a clear path to growing the platform aggressively. It is defensive capital allocation, not offensive.
A Spotify-to-iTunes Moment?
The right analogy may be the kind of generational shift that turned iTunes from category-defining to category-displaced when Spotify came along with a more modern model. Salesforce risks ending up as the worst mousetrap of the group: a once-leading product whose architecture, pricing, and incentives belong to a previous wave of technology.
Reversing the narrative would require Salesforce to demonstrate something concrete: a credible case study showing that customers are thriving on its software while sustaining high per-customer spend even as seat counts decline. That is a tall order, because every pricing model the company is testing points in the opposite direction — declining revenue from existing customers and an inability to capture new ones from the well-funded, AI-native startups now hoovering up the next generation of business.
The Bottom Line
The near-term earnings story may not be catastrophic. Expectations are low, and a clean print is plausible. But the smart listeners on the call will be focused on a single metric: new client signings. Defending big legacy contracts is not enough. Without a steady inflow of fresh customers, the long-term arc bends downward.
What is unfolding is bigger than one quarter or one company. It is the moment when the AI era stops being a rising tide for everyone in software and begins separating winners from losers in earnest. Companies with genuine distribution moats and AI-native architectures are pulling ahead. Those whose business models depended on per-seat economics in a world that no longer needs as many seats are discovering, sometimes painfully, that they are on the wrong side of the wave.