Jerry Murdock: Autonomous Agents Are the SaaS Tsunami

Why a $90B Fund Manager Says Agents Will Rewrite Software

Jerry Murdock co-founded Insight Partners in 1995 and built it into one of the world’s largest tech-focused venture firms, now managing over $90 billion. In this conversation with Harry Stebbings, Murdock draws on three decades of pattern recognition to argue that autonomous agents — not AI broadly — are the defining wave of this era, and that most of the software industry is dangerously unprepared.

The tsunami metaphor: “A tsunami is harmless when it’s out at sea. It’s only dangerous when it hits the beach.” Murdock’s framing is precise: AI itself has been building in the open ocean for years. The destructive force — the wave that hits the beach — is autonomous agents. Not copilots, not chatbots, but agents that write code, make purchasing decisions, and operate as employees. The pre-waves are already visible in coding, customer support, and scheduling. The main wave is still approaching.

Cursor is already obsolete: Murdock reports that multiple AI-native portfolio companies — E2B, Eventual, Lotus AI, Get Dynasty — have told him directly that Cursor is obsolete in their view. These companies have deployed autonomous agents (via OpenClaw, NanoClaw, or custom implementations) to write code independently. “Most of the companies, their view as they’ve told me, is Cursor is obsolete. That’s where the product is today.” He gives Cursor credit for smart people and capital, but the directional shift is clear: from assisted coding to autonomous coding.

The agent orchestration stack: Murdock draws a parallel to the LAMP stack (Linux, Apache, MySQL, PHP) that enabled the web explosion of 2004-2005. He predicts an equivalent “Claw stack” for autonomous agents — a reasoning layer dominated by Claude, Codex, and Gemini, topped by an orchestration layer that routes workloads across models based on cost and capability. “They’re going to triage workflows and say, ‘For this part, let’s use Claude — expensive tokens but they’ll get the job done better.’ And for this other part, let’s use an open-source model like DeepSeek.” This probabilistic routing, driven by agents themselves rather than human developers, will accelerate open-source model adoption and ASIC chip demand.

Agents as the new buyers of software: The most provocative thesis: software will no longer be purchased by humans. Agents become employees with credentials and identity. You manage them the way you manage staff — reviewing what they bought, what they spent, what they accomplished. “An autonomous agent becomes an employee. You give it credentials. You give it identity. And then it’s up to the agent to make the decision.” This fundamentally changes pricing models toward consumption-based billing, sales motions, and the entire SaaS value chain.

Labor displacement becomes a political issue: Murdock predicts that AI-driven white-collar displacement will be a defining issue in the next presidential election, roughly two and a half years away. The first impact is not on current jobholders but on new hires — companies simply stop adding junior developers, assistants, and marketing staff. SMBs adopt first; enterprise follows. He raises the possibility of a minimum viable income becoming a ballot question.

5 Takeaways from Murdock on the Autonomous Agent Revolution

• Autonomous agents are the tsunami, not AI generally — The distinction matters: copilots assist, agents replace. The wave that destroys SaaS moats is agents that write code, buy software, and manage workflows without human involvement
• Systems of record survive only through adaptation — Salesforce won’t melt overnight, but its value is now a function of how well the ecosystem built on top of it survives the agent wave. If satellite companies get picked off, the foundation erodes
• Billion-dollar single-person companies are real — Murdock sees this as inevitable given how well autonomous agents work. The constraint isn’t technology; it’s the founder’s ability to deploy and review agents effectively
• Timing is everything for investors — The best time to start a new fund is now, precisely because incumbent investors are too wealthy and comfortable to move fast. Sea changes reward new entrants
• Intuition is the one thing agents lack — The best investment decisions still require human intuition. Murdock distinguishes between genuine intuition and wishful thinking: the latter cost him more than any market crash

What the SaaS Apocalypse Means for AI-First Organizations

Murdock’s thesis is not abstract — it’s grounded in real portfolio data from companies already running autonomous agents in production. The implication for every software company is binary: either you are building for agents as your primary customers, or you are building a product that agents will route around. The speed of this shift — two weeks to six weeks from pilot to full autonomous coding at multiple companies — suggests the window for adaptation is shorter than most boards realize. Organizations that treat this as a bolt-on strategy will find themselves, as Murdock puts it, caught on the beach when the wave hits.