The Great Reckoning: How Artificial Intelligence is Dismantling the SaaS Empire
Xpert Pre-Release
Language selection 📢
Published on: February 23, 2026 / Updated on: February 23, 2026 – Author: Konrad Wolfenstein
The Great Reckoning: How Artificial Intelligence is Dismantling the SaaS Empire – Image: Xpert.Digital
AI is killing software subscriptions: Why SAP, Oracle & Co. are suddenly in free fall
From tech darling to problem child: Why Silicon Valley is facing a historic crash
For over two decades, Software-as-a-Service (SaaS) was considered the gold standard in the tech world. The simple yet ingenious principle—renting software instead of buying it and billing per user—brought astronomical valuations to giants like Salesforce, Adobe, and SAP, and was seen as a crisis-proof money-printing machine. But this empire is currently crumbling at breathtaking speed. The trigger is a new generation of artificial intelligence that no longer just assists, but acts completely autonomously. If AI agents take over the work of entire departments in the future, or if companies simply generate their own code, the foundation of the lucrative subscription model will collapse. This realization has already caused unprecedented carnage on the stock markets: The "SaaSpocalypse" wiped out over a trillion dollars in market value in a very short time. But the dramatic price crash is just the beginning. We are facing a tectonic shift that will radically transform not only industry pricing models but also the labor market, the venture capital ecosystem, and a multi-billion-dollar credit market. Anyone who still believes that the classic software subscription is a model for eternity has missed the boat.
Related to this:
- What exactly is SaaS? Who are the largest SaaS providers and what are their most important products?
Anyone who still believes that software subscriptions are a model for eternity hasn't heard the news
The Software-as-a-Service (SaaS) industry is undergoing the most profound transformation since its inception. What was considered an unassailable business model for over two decades is being shaken to its core by the rapid development of artificial intelligence. In the first few weeks of 2026 alone, over one trillion US dollars in market capitalization was wiped out in software companies. The SEG SaaS Index, which tracks the share price performance of publicly traded SaaS companies, was already down 12.1 percent at the end of October 2025, while the NASDAQ 100 rose by 17.9 percent and the S&P 500 by 14.2 percent during the same period. This divergence has since intensified dramatically. To speak of a temporary correction would be a gross understatement. The software industry is facing a tectonic shift, the effects of which on company valuations, the job market, and the entire startup ecosystem are only beginning to become apparent.
Related to this:
The bloodbath on the stock markets: When former darlings go into free fall
The numbers speak for themselves. A basket of SaaS stocks compiled by Morgan Stanley saw a 15 percent decline by mid-January 2026 alone, following an 11 percent drop in 2025. Salesforce, long the epitome of a successful cloud company, lost 26 percent of its market capitalization. Adobe fell by 19 percent, Atlassian by 30 percent. ServiceNow, HubSpot, and other industry giants experienced declines of 30, 40, or even 50 percent from their peaks. The term "SaaSpocalypse," coined by a Jefferies trader, has proven apt, and hedge funds earned $24 billion by shorting software stocks.
The case of SAP, Europe's most valuable technology company, is particularly revealing. Its stock reached an all-time high in February 2025 with a market capitalization of €344 billion, temporarily making it the largest listed company on the continent. Since then, SAP has lost around $130 billion in market value, hitting its lowest level since August 2024 in January 2026 and experiencing its biggest one-day drop since October 2020, with a decline of over 15 percent, following the release of disappointing cloud forecasts for 2026. Analysts at Citi commented that SAP needed a significant acceleration to counter the negative industry sentiment, but the mixed signals in its latest update suggested underperformance.
Oracle, once buoyed by a $300 billion partnership with OpenAI, experienced an equally dramatic fall. Its stock plummeted 13 percent in a single day in December 2025 when AI costs exceeded expectations and revenue forecasts fell short. For the full year, analysts say Oracle lost roughly 53 percent of its market value, dropping from a valuation near $630 billion to about $550 billion before continuing its downward spiral. The irony is clear: even companies investing heavily in AI are being punished by the market, as investors doubt the enormous expenditures will ever pay off.
Why AI is attacking the foundation of the SaaS model
To understand the depth of this crisis, one must grasp the fundamental economic principle of the SaaS business model. For over two decades, the industry's success was based on an elegant mechanism: software is no longer sold as a one-time purchase, but rented as a subscription, typically per user per month. The more employees a company has, the more licenses it needs, and the higher its revenue. This so-called seat-based pricing scaled perfectly in a world where human labor was the primary driver of productivity.
Artificial intelligence is disrupting this equation on several levels simultaneously. First, AI agents—autonomous software systems that perform tasks independently—enable a radical increase in productivity. If a single employee can do the work of five or ten colleagues with the help of an AI agent, the number of software licenses required decreases accordingly. Satya Nadella, CEO of Microsoft and thus head of one of the world's largest SaaS portfolios, has described this shift with remarkable candor: business applications are essentially CRUD databases with business logic, and this business logic will migrate entirely to the AI layer in the future.
Secondly, generative AI is increasingly enabling companies to develop their own software instead of renting it. So-called vibe coding, where AI systems generate complete code based on natural language instructions, dramatically lowers the barrier to entry for software development. If a marketing department can have its own analytics dashboards built by an AI system, why should it still subscribe to a specialized SaaS product?
Third, AI is changing the role of the interface itself. In the most ambitious version of this development, AI agents become the primary point of contact between humans and software. Instead of opening and operating multiple specialized applications individually, the user communicates with a central agent that accesses the APIs of various services in the background. In this scenario, SaaS tools become what the telecommunications industry calls "dumb pipes": they store data and provide interfaces, but the real value creation shifts to the agent level. For SaaS providers, this means a brutal dynamic: if the agent is the interface, the underlying software becomes interchangeable, price wars ensue, and the competitive advantages built up over years lose their relevance.
The Palantir shock and the awakening of the markets
The catalyst for the recent sell-off was a single earnings call. When Palantir CEO Alex Karp announced in early February 2026 that AI had become so powerful at writing and managing enterprise software that many SaaS companies were in danger of becoming irrelevant, it triggered a sell-off that wiped out $300 billion in market capitalization for Microsoft, Salesforce, ServiceNow, and others. Palantir itself had previously reported revenue growth of 70 percent in the fourth quarter of 2025 and projected 61 percent growth for 2026. The market message was clear: While Palantir prospers as an orchestrator of AI systems, traditional SaaS providers are losing their raison d'être.
Just a few days earlier, Anthropic had unveiled Claude Cowork, an AI agent that independently creates reports, generates tables from screenshots, and extracts information from various documents. Although the product was still in the preview stage, the announcement was enough to send Intuit's stock plummeting by 16 percent, while Adobe and Salesforce each lost more than 11 percent. The analyst at Mizuho Securities put it bluntly: many institutional buyers no longer see compelling reasons to invest in software stocks, no matter how low valuations fall, as they perceive no catalysts for a revaluation.
Klarna's failed experiment and the limits of AI euphoria
While markets are reacting with panic, reality paints a more nuanced picture. Perhaps the most discussed use case is that of the Swedish payment service provider Klarna. CEO Sebastian Siemiatkowski announced a radical AI strategy in 2024 and 2025: Klarna would halve its workforce, eliminate 1,200 SaaS tools, including Salesforce and Workday, and replace everything with internal AI solutions. According to the company, an OpenAI-based chatbot had taken over the work of 700 customer service representatives and enabled annual savings of $40 million. Average annual revenue per employee increased from $400,000 to $700,000.
But by early 2025, Siemiatkowski had already backtracked. He publicly admitted that the approach had gone too far and that quality was suffering as a result of the radical cost-cutting. Klarna began rehiring people, particularly for customer service, because the AI-powered systems had caused quality problems. Investing in human support quality is the way forward, Siemiatkowski told Bloomberg. The Klarna case illustrates an important truth: AI can achieve enormous efficiency gains, but the attempt to completely replace human labor reaches its limits faster than the industry's euphoria would like to admit.
The tectonic shift in pricing models
The structural threat to the SaaS business model extends beyond mere disruption by AI agents. The entire pricing structure of the industry is at stake. The percentage of companies still relying on seat-based pricing models fell from 21 to 15 percent within a year, while hybrid usage-based models increased from 27 to 41 percent. Gartner predicted that by 2025, more than 30 percent of SaaS solutions would integrate outcome-based pricing components.
The transformation follows a clear hierarchy: The old model calculated access to the software, i.e., the potential to work. The current transitional model calculates actual usage, such as API calls or tokens consumed. The future model calculates the achieved result, such as per generated lead or closed contract. Salesforce is already experimenting with so-called Agentic Enterprise License Agreements, flat-rate structures for companies that use AI agents on a large scale.
For established SaaS companies, this shift presents a fundamental paradox: the better their software becomes thanks to AI integration, the fewer licenses are needed. A CRM system that automates lead qualification, contact, and appointment scheduling with embedded AI agents might reduce the number of users from twenty to two. In the old pricing model, this translates to a 90 percent drop in revenue, even though the product has become ten times more powerful. This is the so-called incentive misalignment trap: product improvement leads to revenue loss.
The time bomb in the credit market: Private credit and the $600 billion bet
While the public focuses on falling stock prices, a potentially even more dangerous problem is brewing behind the scenes. The private credit market, which has financed over 1,900 software companies in private equity acquisitions worth more than $440 billion in the past ten years, has an estimated $600 to $750 billion invested in the software sector. Between 20 and 25 percent of all private credit deals involve SaaS companies, and according to UBS, as many as 25 to 35 percent are threatened by AI disruption.
The fundamental assumptions on which this lending was based are now all being called into question: stable recurring revenues, high margins, predictable cash flows, and high switching costs. At the beginning of February 2026, $17.7 billion in technology-related corporate loans fell to distressed levels—below 80 cents per dollar—within just four weeks. The total volume of distressed technology loans amounted to approximately $46.9 billion, dominated by SaaS companies. In the leveraged loan market, a record $25 billion in software loans are now trading below the distress threshold.
Particularly alarming is the fact that Apollo, one of the most experienced lenders in the market, nearly halved its software exposure in 2025, from about 20 percent to around 10 percent. When Apollo de-risks a sector, it should be seen as a serious warning sign. Deutsche Bank was left with $1.2 billion in loans for a software acquisition that it couldn't sell to investors—a so-called hung deal that illustrates how quickly lender demand evaporated.
The parallels to the 2023 banking crisis, when Silicon Valley Bank collapsed due to a concentration in the technology sector, are unsettling. A single market segment, massively overweight in an asset class whose underlying assets are losing value due to technological upheaval, and an investor base that simultaneously withdraws its funds at the shift in sentiment. There will likely not be a single, spectacular weekend of government bailouts, but rather a creeping process of silent defaults, write-downs, and credit tightening affecting hundreds of software companies.
🎯🎯🎯 Benefit from Xpert.Digital's extensive, five-fold expertise in one comprehensive service package | BD, R&D, XR, PR & Digital Visibility Optimization
Benefit from Xpert.Digital's extensive, five-fold expertise in a comprehensive service package | R&D, XR, PR & Digital Visibility Optimization - Image: Xpert.Digital
Xpert.Digital possesses in-depth knowledge across various industries. This allows us to develop tailored strategies precisely aligned with the requirements and challenges of your specific market segment. By continuously analyzing market trends and monitoring industry developments, we can act proactively and offer innovative solutions. The combination of experience and expertise generates added value and provides our clients with a decisive competitive advantage.
More information here:
The great software crisis is here: What comes after the golden age of SaaS
The devastation in the venture capital market
For startups traditionally built on the SaaS path, the funding environment has deteriorated dramatically. In the first half of 2025, 53 percent of all global venture capital investments went to AI startups, and in the US, the figure was even higher at 64 percent. Yet AI companies represent only 29 percent of all funded startups, meaning they absorb a disproportionate amount of capital per transaction. By October 2025, venture capitalists had invested a total of $192.7 billion in AI startups, and 2025 is expected to be the first year in which more than half of all VC funding flows into a single technology category.
For traditional SaaS companies, only the scraps remain. Jason Lemkin, founder of SaaStr and one of the most influential voices in the SaaS ecosystem, analyzed over 1,000 pitch decks and reached a damning conclusion: The venture capital market has fundamentally split in two. The winners are AI-native companies with unprecedented growth rates, even with negative gross margins, and exceptional traditional SaaS companies with over 100 percent growth and more than $25 million in annual recurring revenue. All others find themselves in a funding desert, regardless of the quality of their business or the satisfaction of their customers.
Private market valuations for SaaS companies have plummeted from multiples of around 18 times revenue in 2021 to 3 to 6 times. A company growing from $20 million to $35 million in annual revenue—a truly impressive feat with 75 percent growth—is considered virtually unfundable in this market. The recommendation from industry experts to non-AI-focused startup founders is unequivocal: extend the runway, postpone fundraising, and hope for a more favorable market environment.
This also has consequences for the exit side of the startup ecosystem. Many SaaS startups planned to exit through acquisition by larger SaaS companies. If these potential buyers themselves are under pressure, their valuations plummet, and their credit lines become more expensive, a crucial component of the entire startup cycle disappears. While SaaS M&A activity reached a record high of over 2,500 transactions in 2025, the median deal size shrank from $67 million in 2021 to just $41 million. Transactions under $500 million accounted for 82 percent of the total volume. The market has shifted from transformative mega-deals to small, strategic bolt-on acquisitions, significantly reducing overall value creation.
Related to this:
The job market bombshell: Still quiet, but already live
The potentially most significant economic impact of the SaaS crisis is the labor market. The figures point to a development whose full force is yet to be felt. According to industry trackers, 170,630 jobs were lost in the US technology sector in 2025, a significant increase from 95,000 the previous year. In the first half of 2026 alone, over 62,000 more jobs were lost, with Amazon alone eliminating 14,000 positions in middle management, customer service, software development, and human resources.
The survey data is even more alarming than previous layoff figures. According to a survey of US business leaders, around 37 percent of companies plan to replace employees with AI by the end of 2026, and almost 30 percent have already done so. The World Economic Forum predicts that 85 million jobs worldwide will be displaced by AI. A study published by the Harvard Business Review shows that a majority of the organizations surveyed have already implemented minor to moderate (39 percent) or major (21 percent) staff reductions in anticipation of AI.
Particularly worrying is that companies are cutting jobs based on projected efficiency gains rather than on the actual, proven performance of AI systems. This means that even if AI agents fall short of expectations, the jobs will not return because companies need to maintain their public narratives about AI-driven productivity. Unemployment among 20- to 30-year-olds in technology-exposed jobs has risen by almost three percentage points since the beginning of 2025.
SaaS companies, in particular, face a double burden. On the one hand, they themselves are reducing staff to realize AI-driven efficiency gains and defend their shrinking margins. Workday reported layoffs of 8.5 percent of its workforce, explicitly citing AI. On the other hand, every staff reduction at their customers leads to fewer software licenses being needed, which reduces their own revenue. This creates a vicious cycle: AI reduces workforces, fewer employees require fewer software licenses, and declining revenue forces further cost cuts, which in turn lead to layoffs.
Gartner predicts a 20 to 30 percent reduction in customer service and support positions by 2026 solely due to investments in generative AI. The average number of SaaS applications used by an organization decreased from 112 to 106 by April 2025, with 82 percent of organizations actively reducing their number of software vendors.
Related to this:
BofA's counter-argument: A logical contradiction in the market?
Not all analysts share this doomsday pessimism. Bank of America, in a widely discussed analysis, argued that the current sell-off is based on two mutually exclusive scenarios. On the one hand, the market fears that AI investment spending will deteriorate and returns will remain weak. On the other hand, it fears that AI adoption will become so ubiquitous and productivity-enhancing that established software workflows and business models will become obsolete. Both cannot happen simultaneously.
If AI is powerful enough to disrupt established industries, the infrastructure investment supporting it cannot collapse. Conversely, if spending collapses due to poor returns, the technology cannot be ubiquitous enough to threaten traditional software models. BofA forecasts that AI investment spending will quadruple to $1.2 trillion by 2030 and sees the current valuation compression as an attractive entry point.
The objection has intellectual merit, but it may fall short. Markets don't operate according to logical coherence, but rather by repricing risks under uncertainty. Investors discount future cash flows, not conceptual contradictions. Both scenarios can certainly coexist if they are understood as expressions of fundamental uncertainty about the speed and extent of AI adoption, not as binary predictions. Current volatility may reflect less a logical error on the part of market participants than the rational realization that the range of possible developments has become extremely wide and that almost all scenarios are unfavorable for traditional SaaS providers.
Why even AI SaaS companies are not safe
A common assumption is that at least those SaaS companies that successfully integrate AI into their products will thrive. This assumption deserves critical examination. SAP directly contradicted the thesis of Palantir CEO Karp, stating that AI agents would massively expand the performance limits of SaaS solutions, but not replace them. However, the market reaction to SAP's own AI strategy was anything but encouraging: the share price continued to fall despite significant investments in AI integration.
The problem runs deeper. When AI becomes a commodity—that is, when the major language models from OpenAI, Anthropic, Google, and others become increasingly powerful and interchangeable—then the competitive advantage shifts away from the AI technology itself and toward implementation, control, integration, and compliance. In this scenario, it's not the SaaS providers who integrate AI into their products who benefit, but rather the orchestration platforms that manage AI agents across various systems. Palantir has positioned itself precisely this way and is being rewarded accordingly by the market.
SaaS companies adding AI capabilities face an additional cost challenge. Every agent-driven action incurs token, compute, and API costs. In a seat-based model, a heavy user deploying hundreds of AI agents can quickly turn a profitable account into a loss-making business. SaaS industry gross margins, traditionally at 70 to 80 percent, are under significant pressure with the addition of AI inference costs.
The death of uniformity: The bifurcation of the software landscape
What is emerging is not a uniform erosion, but a fundamental split in the software market. On the one hand, AI-native companies are emerging, built from the ground up for a world where agents do the majority of the work. These companies are scaling at an unprecedented rate, absorbing the lion's share of venture capital, and defining new valuation metrics. On the other hand, established SaaS providers are struggling with the transformation of a business model they have built and optimized over decades.
Orlando Bravo of Thoma Bravo, arguably the most important private equity investor in the software sector, offered a remarkably sober assessment in Davos: AI will disrupt some software companies—less than half, he predicted—but for many, especially those whose core competencies are technical, it will be disruptive. Less than half as a bull's baseline assumption. That should give us pause.
The industry is clearly evolving towards a new architecture in which the agent layer replaces the old application layer as the primary value creation layer. SaaS companies that can position themselves as the platform of platforms, as the central nervous system where all agent decisions are validated and logged, have a chance of survival. ERP providers like SAP could find themselves in this role if they manage the transition quickly enough. Point solutions that serve a single workflow, on the other hand, face the greatest challenge.
The dilemma of key performance indicators: When the old metrics lie
Over two decades, the SaaS industry has developed a sophisticated system of key performance indicators (KPIs): Annual Recurring Revenue, Net Dollar Retention, Customer Acquisition Cost, Lifetime Value, and the Rule of 40. This system of KPIs served not only for internal management but also as the common language between founders, investors, and analysts. Valuations were calculated based on these metrics, financing rounds were structured, and acquisition prices were negotiated.
In an AI-transformed world, these metrics lose their significance. When the value of software is no longer measured by the number of users accessing it, but by the results it achieves, then ARR is no longer a reliable indicator of company value. If AI agents increase customer retention but simultaneously reduce the number of licenses required, net dollar retention can rise while absolute revenue falls. The entire metrics-based management and evaluation logic of the industry needs to be rethought.
This is a profound disruption for the venture capital market. For over a decade, VCs have refined a repertoire of due diligence methods, valuation models, and portfolio strategies built upon the principles of the SaaS business. When these principles no longer apply, an entire ecosystem loses its navigation system. Not only are many startups on the SaaS path, and most success metrics are geared toward it, but acquisition targets also frequently include SaaS companies, which are now themselves under pressure.
Looking ahead: An uncomfortable prognosis
The hope that the current situation will quickly reverse is likely to be disappointed. Several structural factors point to a worsening of the situation. First, the performance of AI models continues to accelerate. The release of Anthropic's Claude Cowork was not an endpoint, but an intermediate step. Each new generation of models expands the range of tasks that can be performed autonomously, and thus the range of SaaS products that could become obsolete.
Secondly, the feedback effect on the labor market is only just beginning to unfold. If 37 percent of companies actually replace employees with AI by the end of 2026, as announced, this will lead to a measurable decline in SaaS license revenues, which has not yet been adequately considered in any industry revenue forecast.
Third, the private credit wall looms. Twenty-three of the 32 business development companies assessed have $12.7 billion in unsecured debt maturing in 2026, a 73 percent increase over 2025. Refinancing this debt in an environment of falling software valuations and rising risk premiums will be extremely difficult for many companies.
Fourth, companies are actively reducing the number of software providers they use. This consolidation trend is further fueled by AI-powered alternatives and particularly affects specialized SaaS companies that serve individual niche functions.
The software industry isn't going anywhere. Software remains the cornerstone of the digital economy. But the business model that has defined the industry for over two decades—seat-based cloud subscriptions—is undergoing irreversible change. The winners will be those companies that reinvent themselves as data and orchestration platforms, implement outcome-based pricing models, and embrace AI not as a feature, but as a foundation. All others are entering an era of structural contraction, where overcapacity must be reduced, valuations compressed, and business models fundamentally transformed. The uncomfortable times have only just begun.
Your global marketing and business development partner
☑️ Our business language is English or German
☑️ NEW: Correspondence in your native language!
Konrad Wolfenstein
I and my team are happy to be available to you as your personal advisor.
You can contact me by filling out the contact form here or simply call me at +49 7348 4088 965. My email address is: [email protected]
I'm looking forward to our joint project.
