Vijay Raina is a seasoned veteran in the world of enterprise SaaS, known for navigating the complex intersections of software architecture and market shifts. As the industry grapples with the promise and the perceived threat of artificial intelligence, Raina offers a grounded perspective on why the predicted “SaaS-pocalypse” might be more of a slow evolution than a sudden collapse. This conversation explores the strategic moves of industry giants, the reality of developer efficiency, and why the “vibe-coding” revolution faces significant hurdles in the corporate boardroom.
The interview covers the massive financial commitments being made toward AI coding agents and the resulting shifts in corporate headcounts. It addresses the tension between the ease of “vibe-coding” custom solutions and the high risks associated with abandoning established platforms. Finally, the discussion dives into the evolving commercial models of SaaS vendors and how they plan to monetize the AI-driven future over the next two decades.
How does a $300 million investment in AI coding agents fundamentally change the way developers and enterprises interact with software building?
This level of investment, specifically the circa $300 million planned for Anthropic in 2026, signals a pivot where “everything is gonna be cheaper to make” and far more efficient. We are moving into an era where coding agents and humans work together to break through obstacles that previously stalled innovation for months or even years. The internal impact is already visible; for instance, despite continued growth, there was a complete halt on hiring new software engineers in 2025. At the same time, we saw the cutting of around 4,000 support staff, which illustrates a shift toward a leaner, “LLM-guzzling” future. This isn’t just about writing code faster; it’s about the ability to implement software and sell it at the exact same time, creating a dizzying speed of delivery that was previously impossible.
With the rise of “vibe coding” and the ability for individuals to generate their own tools, why aren’t we seeing a mass exodus of companies building their own CRM systems to avoid high licensing fees?
The idea of the “SaaS-pocalypse” suggests that by June 2028, US unemployment could hit 10 percent because former application vendor staffers are being replaced by AI-coded tools, but that overlooks the reality of enterprise inertia. While some people are already “vibe-coding” CRM systems to avoid big-name fees, the risk of switching off a major platform remains incredibly expensive and technically daunting. Most organizations only spend about 3-10 percent of their total revenue on IT, and the actual software portion of that is even smaller. A CIO would have to be incredibly brave to face the “wrath of users” who notoriously hate change, especially when the move involves the nightmare of migrating data and replicating complex security. For a mere fraction of a percent in savings, most businesses would rather stay in the “commercial bear traps” they know than gamble on a home-grown AI solution.
How should enterprises interpret the shifting commercial models, like the move away from capped agreements toward more flexible AI-driven pricing?
There is a “very high margin opportunity” for vendors here, and they are playing a very long game. We are seeing instances where a company is happy to make a loss on a capped-price deal in the short term because they believe they have 20 years to monetize that customer relationship. Gartner has warned that these capped enterprise agreements might not be there when it comes time to renew, which creates a sense of unpredictability regarding future costs. The vendors argue that renewals remain flexible and that they are simply “tailoring terms” because AI compute costs are constantly shifting. It creates a tension where the customer is promised “unprecedented” outcomes and “awesome” results, but they must accept a pricing structure that makes it harder to calculate the long-term value.
What does the internal architecture of a modern SaaS giant look like as it pivots toward AI, and how does this affect the human element of support and engineering?
The architecture is becoming much more automated, with a focus on “high margin opportunities” that require fewer human touchpoints. When you see a company cut 4,000 support staff while simultaneously leaning into AI agent platforms, you are seeing a business that is shedding its traditional human-heavy skin. The goal is to reach a state where the software is “more efficient” than ever before, essentially using AI to build the very tools that will then sell themselves. It creates an environment that feels fast and futuristic, but it also places a heavy burden on the remaining humans to work in perfect tandem with these coding agents. The focus has shifted from maintaining a massive workforce to harnessing LLMs to ensure the company can grow its revenue without growing its headcount.
What is your forecast for the so-called SaaS-pocalypse?
The forecast for a total industry collapse is likely premature, as the structural foundations of enterprise software are much deeper than just the code itself. While we will certainly see more specialists using LLMs to build custom extensions or “vibe-coding” peripheral interfaces, the heavy lifting of security, data integrity, and ingrained user habits will keep the major players in power. By 2028, we likely won’t see 10 percent unemployment driven by AI-coded CRM replacements; instead, we will see a more stratified market where the giants have integrated AI so deeply that they become even more indispensable. The real shift will be in the margins, where companies leverage their $300 million investments to maintain “eye-watering” profits while the “vibe-coding” movement remains a useful, yet ultimately peripheral, tool for niche applications.
