The “SaaSpocalypse” narrative says AI makes software so cheap to build that SaaS incumbents die. Konstantine Buhler’s counter: AI is a work revolution, not a communication revolution. Work revolutions (like the PC/Excel shift) give new tools but don’t change distribution. Communication revolutions (like the internet) change distribution entirely. Since AI doesn’t change who has enterprise relationships, incumbents keep their distribution advantage.
The evidence: Freshworks was founded in Chennai with a significant cost advantage. Yet Freshworks is a $4.5B company while ServiceNow is $200B. Cheaper software development — even dramatically cheaper — does not overcome deep enterprise moats. The same logic applies to AI-generated software.
Buhler projects we could see “five or ten trillion-dollar enterprise SaaS businesses” — individual companies at $1T+ market cap, concentrated like the Magnificent Seven in tech. The winners: ServiceNow, Palantir, Palo Alto Networks — companies entrenched in large enterprises with decades of integrations. “The biggest moats at the biggest enterprise companies are not software moats, they are people moats; they are relationships with the C-suite.”
This follows the pattern from Technology transitions create more of the 'dying' thing, not less: Oracle grew alongside Salesforce during the SaaS transition. Both old and new expanded. AI will do the same — the market grows but concentrates upward toward platform winners, consistent with the thesis that infrastructure ownership decides AI winners.
Separately, Buhler frames AI’s real opportunity as the massive US services market (labor budgets, not existing software spend). Cloud expanded the software market significantly. AI targets the much larger labor pool — and Already-outsourced tasks are the autopilot wedge — vendor swap beats reorg identifies the entry point: already-outsourced tasks where budget exists, external delivery is accepted, and substitution is a frictionless vendor swap. This resolves the tension between middleware compression and SaaS survives as the governance and coordination layer — determinism still rules: middleware SaaS dies, but platform SaaS that provides governance and deep enterprise workflows absorbs that value and grows.