SaaS Is Thriving Despite Industry Pessimism
▼ Summary
– The narrative that AI is killing SaaS is driven by a narrow group of beneficiaries like hyperscalers and foundation model labs, who gain from redirecting enterprise spending.
– A key example is Klarna, which was reported to have replaced Salesforce with AI but actually switched to other SaaS tools, showing a gap between hype and reality.
– Research indicates most enterprise AI projects fail to deliver profit, with only 5% reaching production and success tied to practical back-office automation, not sales tools.
– Technological shifts like AI rarely cause extinction; instead, they force adaptation, as seen with SaaS vendors integrating AI features rather than being replaced.
– Companies that survive are those that adopt useful technology without succumbing to hype, maintaining focus on existing customers and measurable business outcomes.
The narrative that artificial intelligence has rendered traditional software obsolete is pervasive, but a closer look reveals a more nuanced reality. The SaaS industry is not dead, it is undergoing a significant metamorphosis. The most successful companies will be those that strategically integrate new capabilities without abandoning proven business logic or ceding their future to a handful of dominant platform vendors.
Consider the instructive case of Klarna. In 2024, its CEO declared the company had “shut down Salesforce,” sparking a wave of headlines proclaiming the end of enterprise software. The truth, clarified months later, was far less dramatic. Klarna had replaced one SaaS platform with a collection of others, alongside specialized tools. The company still uses Slack, a Salesforce product. This gap between the sensational headline and the mundane correction reveals the mechanics behind the entire “SaaS is dead” narrative. The story spread instantly, while the factual retraction barely made a ripple.
It’s crucial to ask who benefits from promoting the idea that AI is replacing SaaS. The list is short. Hyperscalers like AWS and Google Cloud benefit, as AI workloads justify staggering capital expenditures. Foundation model labs such as OpenAI and Anthropic benefit when enterprise spending shifts to their APIs, supporting valuations that are otherwise challenging to justify. Venture capitalists benefit by repricing portfolios around AI-native startups. This circular financing, where a chipmaker invests in an AI lab that then buys its chips, creates an illusion of organic demand that can obscure real business value.
The critical question is whether this spending translates into tangible results. Data from MIT’s Project NANDA offers a sobering perspective. Their 2025 study found that despite billions in enterprise generative AI investment, a staggering 95% of pilots showed no measurable impact on profit and loss. Only 5% reached full production. The successful minority shared key traits: they were built by specialized vendors, focused on back-office automation, and integrated deeply with existing workflows. In contrast, most budgets flowed to flashy sales and marketing tools with the lowest return on investment. This isn’t a revolution, it’s a pattern of wasted expenditure alongside pockets of genuine, quiet efficiency gains.
Admittedly, the SaaS market shows real stress. A sharp selloff in software stocks in early 2026, dubbed the “SaaSpocalypse,” wiped out hundreds of billions in market value. The underlying rationale that AI tools can reduce per-seat software costs has merit. However, technological transitions rarely cause extinction, they create heterogeneity. Desktop computing survived the rise of mobile. Cloud coexists with on-premise solutions. History shows that software evolves in accumulating layers, not through total replacement.
Established SaaS vendors are adapting by becoming agent-orchestration platforms. Salesforce launched Agentforce, HubSpot embedded AI tools, and Snowflake formed key partnerships. The transformation in pricing and interface is real, but the core enterprise needs, auditing, compliance, and data gravity, remain unchanged. The extinction event is a marketing fiction.
Every major tech wave briefly anoints new companies as the reinventors of reality. The rhetoric around Nvidia, OpenAI, and others follows a familiar cadence, promising to rewrite civilization itself. There is a grain of truth, as agentic AI represents a real technological advance. The companies positioned to thrive are those disciplined enough to recognize this historical pattern. They adopt what is useful, ignore pure hype, and measure outcomes against costs. They refuse to treat platform vendors as infallible gods, instead viewing them as large commercial entities with specific agendas.
A troubling pattern has emerged among mid-size SaaS companies: a rush to ship AI features while net revenue retention quietly collapses. One firm, eighteen months after an “AI-first” pivot, saw its NRR plummet and lost millions in renewals. The AI features were well-built, but core customers churned because the focus on the future came at the expense of the present product. This is not an argument against AI adoption, but a warning against strategic imbalance.
The previous AI cycle ended with research winters and shuttered startups, while the survivors were those doing quietly useful work. This cycle will likely conclude similarly. Some hype will prove real, but most revenue projections will not. The enduring companies will avoid both extremes. They won’t dismiss the trend, as ignoring AI today is a grave strategic error. Nor will they drown in it, emptying their engineering teams into rebrands while their revenue base erodes.
For the record, Klarna still pays for SaaS subscriptions. It also pays OpenAI. This is the probable shape of the future, not a dramatic death but a quieter rearrangement. The winners will be the pragmatic operators who kept their feet on the ground, integrating new tools while maintaining their core business, while everyone else was fixated on the sky. The funeral for SaaS was well-attended, but the patient, upon examination, is very much alive and adapting.
(Source: The Next Web)
