The narrative in the enterprise software sector has shifted from "growth at all costs" to a desperate, televised hunt for a pulse. For months, a quiet panic—the so-called SaaSpocalypse—has been hollowing out valuations. The thesis was simple and devastating: if an autonomous AI agent can do the work of ten people, why would a company keep paying for ten software "seats"? This fear wiped nearly $2 trillion in market value from the sector since the start of 2026.
But then came the February earnings calls.
RingCentral and Five9, two veterans of the cloud communications trenches, didn't just survive their quarterly post-mortems; they provided a roadmap for how legacy SaaS might actually avoid obsolescence. By the time the markets closed following their reports, RingCentral (RNG) had surged over 30%, and Five9 (FIVN) followed with a double-digit rally. The takeaway for the casual observer was that AI concerns were "quelled." The truth is more surgical. These companies aren't just adding AI; they are cannibalizing their own traditional business models before someone else does.
The Death of the Seat and the Rise of the Agent
For decades, the software industry lived on the "per-seat" model. It was a gold mine. You hire more people, you buy more licenses. AI broke that gears-to-growth linkage. Investors began pricing in a future where headcount—and thus software revenue—would collapse.
RingCentral’s response was to stop fighting the ghost of seat compression and start selling the ghost in the machine. CEO Vlad Shmunis has pivoted the company toward what he calls an "agentic voice AI" firm. It is a subtle but foundational shift. Instead of just providing the pipes for humans to talk to each other, they are selling the intelligence that replaces the conversation entirely.
The numbers suggest the gamble is paying off. RingCentral reported that annual recurring revenue (ARR) from customers using its AI tools has doubled year-over-year. It now accounts for nearly 10% of their total ARR. This isn't just a "feature" anymore; it is the new floor of the business.
Five9 is playing the same high-stakes game. Its enterprise AI bookings more than doubled, hitting a $100 million ARR run rate. When you consider that Five9's total quarterly revenue was roughly $300 million, the math becomes clear. AI is no longer a peripheral experiment. It is the only thing keeping the growth engine from stalling.
The Margin Trap
There is a darker side to this rally that many analysts are glossed over. While the revenue growth is real, the cost of maintaining it is shifting.
Moving from a static software-as-a-service model to an AI-as-a-service model is expensive. Standard SaaS has legendary gross margins because once the code is written, the cost of serving the next customer is nearly zero. AI is different. Every query, every "agentic" decision, and every real-time translation costs money in the form of compute power.
RingCentral managed to expand its non-GAAP operating margin to 22.8% this quarter, but it is also leaning heavily into a partnership with OpenAI to power its GPT-5.2 integrations. This creates a new dependency. These software companies are no longer the masters of their own cost structures. They are increasingly middlemen for the chipmakers and the large language model providers.
The market is rewarding them today because they proved they can monetize the tech. But the long-term question remains: can they keep enough of that money to justify their valuations, or will the "AI tax" eventually eat their lunch?
Defensive Dividends and Stock Buybacks
One of the most telling moves of the month came from RingCentral’s board: the initiation of a quarterly dividend. This is a classic "old tech" move. It is a signal to the market that says, "We are a mature, cash-generating machine, not a speculative growth play."
Between the $250 million increase in share repurchases and the new dividend, RingCentral is trying to floor its stock price with cold, hard cash. It is a defensive posture. It acknowledges that the days of 30% organic revenue growth are over—total revenue grew a modest 4.8% this quarter—and that the company must now compete on financial engineering as much as product innovation.
Five9 is also using the buyback lever, completing a $50 million repurchase recently. For a company that was recently flirting with 52-week lows, these moves are designed to project a confidence that the underlying fundamentals haven't been broken by the AI wave.
The Infrastructure Advantage
The real reason these two companies are rebounding while other vertical SaaS players are still bleeding is moat.
An AI agent is only as good as the data it can access and the systems it can control. RingCentral and Five9 sit on billions of minutes of proprietary voice data and deeply embedded integrations with enterprise workflows. You can't just drop a generic chatbot into a regulated healthcare contact center and expect it to work. You need the "plumbing"—the security, the compliance, and the low-latency connectivity—that these incumbents have spent fifteen years building.
Nvidia’s Jensen Huang recently argued that AI agents will run inside platforms like these, not around them. This quarter's results support that theory. The "SaaSpocalypse" might be real for companies that only provide a thin UI on top of a database, but for those that provide the actual infrastructure of communication, AI is becoming a massive upsell opportunity rather than a replacement.
The Price of Admission
Investors should remain wary. The rally we're seeing is a relief rally, not a return to the bull market of 2021.
Five9’s dollar-based retention rate actually dipped slightly to 105%. Pricing pressure in the enterprise segment is real, and it’s being driven by customers who realize they have more leverage than ever. They are demanding more AI capability for the same price, or they are threatening to consolidate their "seats."
The software industry has survived the first contact with the AI agent revolution. RingCentral and Five9 have shown that by pivoting toward automation and returning capital to shareholders, they can keep the wolves at bay. However, the transition from "selling seats" to "selling outcomes" is just beginning. The companies that can't make that shift will find that their recent rallies were nothing more than a dead cat bounce in a much longer decline.
The next phase isn't about who has the best AI. It's about who can afford to run it.
Would you like me to analyze the specific debt-refinancing risks RingCentral faces with its $609 million convertible maturity in 2026?
