
"What precipitated the latest selloff- dubbed 'SaaSpocalypse' by Jeffries analyst Jeffrey Favuzza-was the one-two punch of Anthropic and then Open AI launching agentic AI systems for enterprises that appear to perform some key functions currently provided by SaaS players, undermining their business models. Will Anthropic's Claude Cowork and OpenAI's Frontier doom the SaaS giants? It depends. Check out my colleague Jeremy Kahn's take on that."
"Data: Mission-critical software that's integrated with sensitive, regulated or high-value proprietary data is hard to dislodge. I've yet to meet a leader who's eager to outsource their payroll or enterprise security to a large language model. What we pay our people is more sensitive than what we pay for products. That makes me wonder if HR companies like Workday, for example, are in a stronger position than some investors might think."
"Volume: SaaS companies tend to charge by user, not usage. Who hasn't seen a new software contract spark lobbying efforts in the office for access to one of the "seats" authorized under the new license. That's different from the model used by Milan Shetti, CEO of Rocket Software, which serves clients in highly regulated industries like financial services, insurance and healthcare. "We have usage‑based pricing, not seat‑based pricing," Shetti told me earlier this week. "The SaaS companies with user‑based pricing have taken a hit, because if AI improves productivity, the number of users goes down.""
Agentic enterprise AI systems from Anthropic and OpenAI have begun to replicate functions provided by many SaaS offerings, triggering investor concern and market selloffs. Software tightly integrated with sensitive, regulated, or high-value proprietary data presents higher switching costs and remains harder to replace. HR, payroll, and security systems face particular resistance to outsourcing to general LLM agents due to privacy and regulatory concerns. Pricing models matter: user‑seat pricing is vulnerable to productivity gains that reduce headcount, while usage‑based pricing can be more resilient. Integration depth, data sensitivity, and pricing structure will differentiate winners and losers in this shift.
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]