VC: 'Autonomous Agents' Are the Real AI Tsunami
Insight Partners co-founder Jerry Murdock argues the coming AI revolution isn't about better generative models, but autonomous agents that can act independently and orchestrate complex workflows. He compares the shift to the LAMP stack revolution of the mid-2000s, stating that companies must be "AI native" to survive. This sentiment is echoed by analysis of new models like OpenClaw, which are designed for persistent, autonomous tasks.
An "AI-native" company is built from the ground up with artificial intelligence as its core, rather than adding AI features to a traditional software product. This means AI is embedded in everything from product architecture to the business model, shaping the entire system rather than being a component within it. The core idea behind autonomous agents is to move beyond single-task AI and create systems that can independently achieve goals. These agents perceive their environment, make decisions, and take actions with minimal human intervention, capable of handling complex, multi-step workflows. Venture capital is increasingly targeting these "AI-native" startups focused on autonomous agents. In 2025, an estimated 30% of all venture funding is expected to go to AI-native companies, a significant increase from 12% in 2022. Projections suggest that by 2026, over half of all VC investments could be directed towards AI-powered startups. OpenClaw is an open-source autonomous AI agent that has gained significant popularity. It functions as a personal assistant that connects large language models (like those from OpenAI, Anthropic, or local models) to a user's applications via messaging platforms such as WhatsApp or Slack. This allows it to manage emails, browse the web, and run scripts to execute tasks. The shift towards autonomous agents is also evident in the enterprise sector. Companies are developing specialized agents for tasks like customer support, financial trading, and IT automation. These agents can proactively resolve issues, prioritize sales leads, and even manage complex back-office processes without direct human command. However, the rise of powerful autonomous agents like OpenClaw introduces significant security and privacy concerns. Because these agents require broad permissions to access sensitive data across emails, calendars, and other services, they are vulnerable to prompt injection attacks where malicious instructions could be hidden in the data they process.
