Citrini Research envisions a future where AI renders software tools and human expertise obsolete, leading to widespread job losses and economic instability.
IBM’s stock experienced a significant decline on Monday after Anthropic suggested its Claude Code tool could automate the modernization of COBOL, an antiquated programming language that has remained a crucial part of IBM’s business for six decades.
The company saw its stock price plunge by 13.2%, marking its steepest drop since the dot-com bust a quarter-century ago.
However, IBM isn’t alone in facing challenges due to the AI buzz. A wave of panic appears to be sweeping across various SaaS companies as AI’s potential to automate software creation and management becomes clearer. Firms like Salesforce, Atlassian, ServiceNow, and Snowflake have all witnessed substantial sell-offs in what one trading company dramatically called the “SaaSpocalypse.”
Yet, according to a hypothetical scenario outlined in The 2028 Global Intelligence Crisis by Citrini Research and co-author Alap Shah of Lotus Technology Management, the situation for SaaS companies, and by extension the broader US and global economies, could soon deteriorate much further.
A Thought Experiment
The report contemplates what the rise of AI might look like from the perspective of 2028, just two years after an imagined peak of 8,000 points for the S&P 500 in October of this year.
“Productivity was surging. Real output per hour escalated at rates unparalleled since the 1950s, propelled by AI agents that neither rest, take sick days, nor require health benefits,” stated Citrini Research in what it labeled a thought exercise. Concurrently, another phenomenon unfolded: “The proprietors of computing power witnessed an explosion in their wealth as labor costs disappeared. Meanwhile, real wage growth stagnated.”
As businesses pursued enhanced AI-driven productivity, they began dismissing employees. However, with workers laid off, demand for goods and services diminished, prompting companies to seek even greater productivity and automation. “In every aspect, AI surpassed expectations, and the market revolved around AI. The sole issue…the economy did not,” Citrini speculates.
SaaS platforms will be the first to experience this impact. In the future, AI agents will handle the SaaS tasks currently performed by businesses and consumers, diminishing the appeal of these platforms and compelling vendors to offer substantial customer discounts to remain viable.
The Subscription Model in Jeopardy
This pessimistic forecast comes shortly after OpenAI’s early February announcement of its new Frontier enterprise agentic platform, which was reportedly pitched to investors as a direct competitor to SaaS offerings.
As Xpert Digital consultancy bluntly asserted: “If AI agents assume the responsibilities of entire departments in the future, or if companies simply develop their own code, the very foundation of the profitable subscription model will crumble.”
Will these difficulties for SaaS providers mean that customers can negotiate better license prices in 2026? And what is the likelihood that some will abandon existing solutions entirely, opting instead for systems from AI companies rather than the AI-enhanced software that SaaS firms will attempt to entice them with?
AI Can Write, But Running is Harder
Experts consulted by CIO.com remain unconvinced that AI’s ascendancy will be as straightforward as Citrini Research suggests. Firstly, AI’s capability to execute business tasks as effectively as SaaS platforms is unverified. Secondly, even if they can replicate these systems, none of it will be auditable. Furthermore, businesses must consider AI agent security concerns and whether they will ultimately be cost-efficient.
“AI makes the process of writing software significantly easier. It does not, however, simplify the operation of enterprise software. These are distinct challenges, and the majority of the cost resides in the latter,” observed SAVVI AI CEO, Maya Mikhailov.
“The moment you decide to build software internally, you also assume responsibility for security, compliance, uptime, integrations, and round-the-clock support. While it sounds appealing in theory, the actual costs and complexities will directly impact the bottom line,” she elaborated.
According to Collin Hogue-Spears, a technical expert at Black Duck Software, concerns about AI’s reliability also persist. “OpenClaw proliferated from zero to 135,000 exposed instances in mere weeks because it executes workflows rapidly. However, it does not generate audit evidence, fulfill license obligations, or produce the compliance documentation a regulator requires before that code is deployed,” he noted.
“AI agents perform tasks,” he remarked. “They do not establish the verifiable trail required to protect your company from regulatory enforcement. This pattern holds true across all industries: AI streamlines commodity features but simultaneously heightens governance responsibilities.”
Nevertheless, SaaS customers should not passively accept any price hikes in 2026. The current vulnerability within this sector presents a unique opportunity. “SaaS vendors are facing the most significant pricing pressure since cloud computing supplanted on-premises licensing, and the opportunity for negotiation is available right now,” Hogue-Spears advised.
This report was originally published on CIO.com.
