# After U.S. artificial intelligence (AI) model startup Anthropic released the enterprise AI agent "Claude Co-Work" on Jan. 23 (local time hereafter), more than $280 billion (about 421.82 trillion won) in market value was wiped out from software (SW) corporations worldwide, including Salesforce, Inc., Microsoft (MS), and Adobe, over two days through the 24th. In January alone, $1 trillion (about 1,507 trillion won) in market value disappeared. The S&P North American Software Index's 15% drop in January was the largest since the 2008 financial crisis. That's because fears spread of a "SaaSpocalypse" (software as a service + apocalypse) that AI would cannibalize the SW industry.
# Nvidia Chief Executive Officer Jensen Huang said at "GTC 2026," Nvidia's annual developer conference that closed on Mar. 19, that fears of a SaaSpocalypse are "completely wrong," adding, "AI will not replace SW corporations but will boost SW corporations' development productivity." Huang also said at the "Cisco AI Summit" held on Feb. 4, "There is a perception in the SW industry that the role of SW tools will decline and AI will replace them," adding, "This is the most illogical idea in the world." He added, "On the contrary, as more individuals use AI agents for their work, usage of SW tools will increase exponentially." He flatly refuted the SaaSpocalypse.
The global software as a service (SaaS) market, which surpassed $400 billion (about 602.6 trillion won) in 2025 and has grown 13-fold over the past decade, is at an inflection point. This is not the first time crisis theories have emerged for the SaaS market, which began to take off thanks to the cloud shift of information technology (IT) systems more than 20 years ago and then grew explosively as if it would devour the world amid the acceleration of digital transformation during the COVID-19 pandemic in 2020. In May 2024, when Salesforce, Inc., a leading SaaS provider, reported revenue that fell short of market expectations for the first time since 2006 and its stock plunged about 20% in a single day, Jeffrey Favuzza, vice president of trading at U.S. securities firm Jefferies, dubbed it the SaaSpocalypse.
Concerns became reality as Sweden's leading Fintech Klarna canceled Salesforce, Inc. in 2024 and Workday in 2025. The SaaS industry is trying to break through the fear of a SaaSpocalypse head-on by improving results, but the vibe coding trend in 2025 (AI writes programs when spoken to in everyday language) and the subsequent AI agent onslaught signaled the dawn of an era when customers build and use SW themselves, reigniting SaaSpocalypse worries and fueling debate over the end of SaaS. Regarding Anthropic's Claude Co-Work, the New York Times (NYT) said on Feb. 3, "Now clicks are made by AI, not humans," and "a SaaS killer has emerged," but there are also projections "that AI will open a new market in the SW space (Kyunghee University Professor Lee Kyung-jun)."
Goldman Sachs predicted in a June 2024 report that AI agents will account for more than 60% of the global SaaS market, which will reach $780 billion (about 1,175 trillion won) by 2030. Economy Chosun covered the future of the SaaS industry that AI will transform and the field response.
1 AI is a tool to strengthen SaaS competitiveness
This is a scenario where AI becomes a "booster engine" that amplifies the value of existing SaaS rather than destroying its core functions. Domains where decades of data have been accumulated, such as cost control at construction sites or clinical data management in pharmaceuticals, and complex regulatory environments will act as barriers that prevent AI from replacing SaaS.
U.S. SW company Veeva Systems is a representative case. The company provides SW that manages clinical data for pharmaceutical companies worldwide and helps them obtain regulatory approval from governments. The company's "Veeva AI Agent" analyzes and delivers specialized regulations and data that general-purpose AI struggles to handle. It targets the fact that while general-purpose AI can mimic the look of SaaS, it is difficult to replicate the sophisticated business rules and closed data engineered into its core. Therefore, AI becomes a tool that maximizes productivity rather than completely replacing users' work. This is why CEO Jensen Huang argued, "AI is a tool to strengthen the competitiveness of SaaS and will remain a booster engine that increases productivity."
2 SaaS is the conduit through which AI ferries data
This is a scenario where SaaS's individual brand power weakens and cedes the lead to AI agents. Instead of accessing individual SaaS, users handle all tasks in AI agents like Gemini or Claude. For example, instead of logging into the Hangul or Word program, a user orders an AI agent, "Compile the existing documents, write a report, and send it to the person in charge," and the AI automatically writes the report, creates it as a Hangul or Word file, and sends it. This is an example of Google Gemini Enterprise using SaaS apps on an integrated AI platform.
When you access Google Gemini Enterprise, you can sync and use data from SaaS systems such as MS SharePoint, Jira, and Confluence. This aligns with Tomasz Tunguz, CEO of Theory Ventures, saying, "As powerful AI models that lord over existing SaaS emerge, existing SW is likely to be pushed behind the screen as a backend System of Record that stores data." As the user touchpoint shifts to AI agents, existing SaaS providers' profitability begins to deteriorate. To prevent this, SaaS providers start restricting AI agents' access to data.
Salesforce, Inc. erected a technical barrier called "masking" in June 2023 to block data leakage to external AI providers, and Adobe fully revised its terms in June 2024 to prevent creators' work from being leaked externally and used for AI training. SaaS providers compete with AI firms by launching their own AI agents to defend user touchpoints or by partnering with AI firms.
3 The per-account billing model of SaaS collapses
This is a scenario where AI completely overwhelms SaaS's existing operating model. Here, AI moves beyond a booster engine that strengthens SaaS competitiveness to serve as the core engine. As SaaS's proprietary data combines with AI automation, an "end-to-end" format takes root in which AI agents handle business processes from start to finish on their own.
Once this stage is reached, the SaaS industry's per-seat billing model collapses. What matters is not how many accounts subscribed to SaaS, but tangible outcomes such as "how many insurance claim determinations were completed" or "how many lines of code were written." A representative example is U.S. online customer support provider Intercom, which fully implemented "performance-based billing" for AI-resolved outcomes at the end of 2025. Intercom had purchased SW licenses based on the number of agents, but as AI agents took over agents' work, it applied performance-based billing instead of per-account billing. Aventis founder Marcin Majewski said, "SaaS as we know it—SaaS as a distinctly separate, independent category—will probably cease to exist."
4 AI that fully engulfs SaaS
This is a scenario in which AI completely swallows existing SaaS, making SaaS virtually unnecessary. The case of Klarna replacing SaaS services such as Salesforce, Inc. and Workday—subscribed since 2024 for two years—with AI agents and laying off more than 700 agents is regarded as a preview of this scenario. The Financial Times (FT) said, "SW has begun to directly replace labor, going beyond being a tool," and "the post-SaaS era has opened, in which the power dynamic between suppliers and consumers is reversed."
In standardized work areas with little human intervention that are repetitive, such as customer support (CS) and simple attendance management, full encroachment begins as AI replaces SaaS. Vice President Favuzza said of this situation, "A wave of SW corporations will degenerate into zombies." The only ways for SaaS providers to survive are to replace their own services with AI or to develop their own AI agents. Examples include ServiceNow embedding Anthropic and OpenAI's AI agents into its platform and Salesforce, Inc. developing its own AI agent to block access by general-purpose AI.