The SaaS Apocalypse: Who Lives & Who Dies | Insight Partners Co-Founder, Jerry Murdock

Market Status & Strategic Shift

• Most current AI-native startups view "bolt-on" AI as obsolete; the market is shifting toward "AI-native" architectures.
• Jerry Murdoch explicitly states that for many startups, the $27–30 billion valued coding platform Cursor is "obsolete" in its current form due to the rise of autonomous agents.
• Emerging companies (e.g., E2B, Eventual, Lotus AI, GetDynasty, Ovens) are already utilizing autonomous agents to write code, a capability developed over the past two to six weeks by some teams.
• Murdoch warns against "being caught on the beach," urging companies to treat the current AI wave as a "tsunami" requiring immediate movement to "higher ground."

The Autonomous Agent Wave & Technical Evolution

• Definition: Murdoch identifies "autonomous agents" as the core tsunami event, distinct from general AI, driven primarily by the open-source OpenClaw and NanoClaw communities.
• The "Claw Stack": Analogous to the 2003–2004 LAMP stack explosion, the industry is building a new stack for autonomous agents featuring:
  • Reasoning Layer: Currently dominated by Claude, Codex, and Gemini.
  • Orchestration Layer: A future layer allowing agents to triage workflows by dynamically assigning tasks to different models (e.g., using expensive models for complex tasks, open-source models like DeepSea or Llama3 for cheaper tasks).
• Infrastructure Impact: The shift to agents will drive demand for ASIC chips (custom chips) over traditional GPU architectures (NVIDIA), as ASICs allow models to run directly on the chip with integrated memory, reducing costs and increasing tunability.
• Speed Requirements: Agent-driven systems require sub-400ms response times; E2B sandboxes operate at ~80ms, a speed critical for agents spinning up hundreds of thousands of instances per second.

Hardware & Competitive Landscape

• NVIDIA: The acquisition of Grok by NVIDIA is interpreted as a strategic move to ensure CUDA remains viable for the upcoming ASIC explosion, particularly by integrating memory directly onto chips.
• Meta: Meta's decision to reject NVIDIA chips ("no" to Jensen Huang) is a direct bet on ASICs.
• Execution Risk: Murdoch notes that NVIDIA's future value depends entirely on execution and the ability to migrate CUDA support to ASICs; without this, value leakage to ASICs could occur.

Investment Strategy & Valuation

• Sourcing: Murdoch's best deal sourcers are his own partners at Insight (Jeff Lieberman, Devon Parekh), followed historically by Peter Fenton and John Doerr.
• Decision Making: Investment success relies on intuition (distinguishing it from wishful thinking) and the ability to identify founders with an "edge" who may be socially challenging but highly driven.
• Risk Profile: Murdoch estimates 80% of his investments have returned less than 1.3x, with the 20% "home runs" driving all returns; no investment path is currently "safe."
• Remote Management: Murdoch argues against remote management, citing his early mistake of hiring a partner for Insight Europe too early; he asserts that proximity and face-to-face interaction are critical for high-stakes decision-making.
• Timing: He identifies timing as the single most correlated factor for VC success, noting that a 2005–2006 fund had a distinct advantage over a 2009 fund for capturing mobile and social waves.

Labor Market & Societal Impact

• Job Displacement: Autonomous agents will replace white-collar roles involving data entry, scheduling, customer support, bookkeeping, legal, and coding, with small and medium businesses adopting first due to immediate efficiency gains.
• Hiring Trends: Companies will begin halting entry-level hiring in white-collar sectors, creating a "hollow" in the career pipeline for junior developers and assistants.
• Political Implications: Murdoch predicts Universal Basic Income (UBI) or a "minimum viable income" will become a major ballot question in the next US presidential election (2.5 years away) as a response to mass white-collar displacement.
• Corporate Structure: The ratio of agents to humans will shift dramatically, with potential for "billion-dollar single-person companies" if agents are deployed effectively; companies like Klarna may reduce headcount from 7,000 to 2,000 by 2030.

Public Market Volatility & SaaS Valuation

• Market Reaction: Recent stock drops in companies like Cloudflare and CrowdStrike (10% declines) are attributed to "cautious sideline buyers" waiting for clarity rather than panic selling, similar to the malaise of March 2000.
• System of Records: The value of traditional SaaS "systems of record" (e.g., Salesforce, Carta) depends on their ability to integrate autonomous agents; those that fail to adapt may see value eroded, while those that leverage agent-driven tokenization or workflow management could increase in value.
• Pricing Models: The buyer base is shifting from humans to agents, driving a move toward consumption-based pricing where agents authorize spending based on memory or compute usage (e.g., Docker's shift to AI consumption).

Personal Wisdom & Lessons

• Failure as Data: Murdoch's primary wisdom stems from repeated failures, including losing money every year for a period and personal struggles with alcoholism in his early career, which taught him to respect the "energy" of money.
• Intuition vs. Wishful Thinking: He learned to distinguish true intuition from wishful thinking, noting that he often failed by investing in founders he personally liked who lacked the necessary "obsession" or "chip on the shoulder."
• Legacy: Murdoch views his career as a series of experiences; he emphasizes that money does not come with instructions and must be respected as energy to be used constructively.
• Future Outlook: He expresses optimism about AI's potential to significantly extend human longevity (potentially adding 5+ years per 5 years survived) and remains committed to investing in AI-native companies from the ground up.