Op-Ed: SaaS isn't dead; you're merely being sold the idea of its demise.
The narrative suggesting that “AI has killed software” has a few vocal proponents and substantial quiet evidence to the contrary. The companies that will thrive over the next five years are those that do not revere hyperscalers as the new authorities.
Whenever I make a claim, I prefer to conduct thorough research first, avoiding the tone of a LinkedIn post. I wish more professionals in this field would adopt the same approach, as there is a common belief that large figures tell the complete story.
During the Black Death, people likely felt it was the end. Similarly, during wartime, societies felt it was the final chapter. Yet, in an unexpected manner, we possess an inherent ability to overcome challenges and leverage change to our advantage.
When AI began to permeate our work and personal lives, many asserted that “AI will replace people,” claiming this technology—though not particularly novel—would dominate our minds, hearts, and jobs, steering us wherever it desired.
However, we continue to work; individuals are still writing, thinking, creating, and building. The heart of EU tech has the latest updates, a tale from our wise founder Boris, and some questionable AI art. Subscribe for free, delivered to your inbox weekly!
In the past two years, a growing number of voices have stated that “SaaS is dead.” This assertion originated from someone influential enough to sway public sentiment, and many were already dressed in black for the impending funeral.
In August 2024, Klarna’s CEO, Sebastian Siemiatkowski, mentioned in an earnings call that the Swedish fintech had “shut down Salesforce,” with Workday following suit. Klarna intended to develop its own AI-driven alternatives, a streamlined system free from the excesses of traditional enterprise software. This statement influenced market movements, resulting in articles proclaiming the demise of SaaS. At Dreamforce, Salesforce’s Marc Benioff was asked to respond to a client who seemingly opted for AI over his product; he admitted to feeling embarrassed.
Six months later, Siemiatkowski clarified that Klarna had not replaced Salesforce with AI, but rather with other SaaS solutions: Deel for HR, third-party tools for CRM, and the Swedish graph database Neo4j for data integration. Klarna still utilizes Slack, which remains a Salesforce product. Siemiatkowski himself expressed on X that he felt “tremendously embarrassed” by how the narrative evolved.
“No,” he stated, “we did not replace SaaS with an LLM.”
This incident is perhaps the most insightful story in enterprise software from the past two years. The disparity between what was communicated and what actually transpired highlights the mechanics behind the entire “SaaS is dead” narrative. The headline gained traction, while the clarification did not.
An industry of analysts, venture capitalists, and foundation model CEOs built a year of marketing on the louder side of the story.
It’s important to consider who benefits from the notion that software-as-a-service is being supplanted by artificial intelligence; the answer is surprisingly limited. The hyperscalers gain, as AI workloads validate the $660 to $690 billion in capital expenditure the five largest U.S. cloud and technology firms plan for 2026, according to Futurum Group analysis—nearly double compared to the previous year.
The foundation model labs also benefit, as every dollar of enterprise software spending shifted to their APIs reinforces valuations that are otherwise challenging to justify. OpenAI concluded 2025 with around $20 billion in annual recurring revenue, while Anthropic surpassed $9 billion in January 2026. These figures are indeed substantial, yet they represent only about three percent and a little over one percent of the hyperscaler capital expenditure aimed at supporting them, respectively.
Venture capitalists profit as their portfolio valuations rely on the narrative that AI-native companies will surpass the incumbents they once supported. Meanwhile, Nvidia, a key supplier and financier of the boom, benefits until circumstances change.
In March 2026, CEO Jensen Huang indicated that his recent investments in OpenAI and Anthropic would likely be the last. The cycle of financing, where Nvidia invests in OpenAI and OpenAI purchases Nvidia chips, had reached a stage where even the chip manufacturer no longer termed it a virtuous cycle.
MIT’s Michael Cusumano, as quoted by Bloomberg, stated starkly: “Nvidia is investing $100 billion in OpenAI stock, and OpenAI is indicating they will buy $100 billion or more of Nvidia chips.”
This could be characterized as demand, or alternatively, as bookkeeping.
The 95% statistic that should have quelled the hype
The more difficult inquiry is whether any of this is leading to business results. Here, the data is less favorable than the marketing materials.
In July 2025, MIT’s Project NANDA issued a report titled “The GenAI Divide: State of AI in
Op-Ed: SaaS isn't dead; you're merely being sold the idea of its demise.
The idea that "SaaS is dead" serves the interests of hyperscalers and venture capitalists. However, the evidence indicates that software is evolving rather than disappearing, and those who adapt will remain successful.