Why AI Founders Must Move Beyond Models to Build Real Value
The velocity of growth in the generative AI sector is rewriting the venture capital playbook, leaving traditional internet and cloud paradigms in the dust. As Anthropic scales toward a staggering $47 billion revenue run rate, the lessons learned from their meteoric rise offer a vital blueprint for the next generation of AI entrepreneurs.
The Unprecedented Velocity of AI Growth
In 25 years of investing, Matt Murphy of Menlo Ventures has never witnessed growth at this scale. During his tenure, he has seen the rise of the internet, the mobile revolution, and the cloud boom, yet none compare to the current trajectory of AI leaders. A standout example is Anthropic, which surged toward a $47 billion revenue run rate by May, a massive leap from its projected 2025 figures.
Menlo Ventures notably led Anthropic’s $500 million Series D, having previously backed the company at a $4 billion pre-revenue valuation. While such a bet was unconventional, early indicators—often called "green shoots"—such as strategic investments from giants like Google and Amazon, signaled that the company was more than just another model provider.
Why the Model is Not the Moat
A critical takeaway for AI founders is that having a superior Large Language Model (LLM) is no longer a sustainable competitive advantage. Murphy argues that a "great model was never the point." In a landscape where model capabilities are rapidly commoditized, the true value lies in building a robust ecosystem around that intelligence.
Anthropic successfully transitioned from being a model-centric company to a platform-centric powerhouse by launching functional layers like Claude Code, the Model Context Protocol (MCP), and Claude Skills. These tools allow users to integrate AI into complex workflows, transforming a raw intelligence engine into a comprehensive productivity platform. For founders, this means the real "moat" is built through developer tools, integration capabilities, and user-centric applications rather than just parameter counts.
Competing in the Era of Hyper-Growth
The emergence of high-growth startups like Lovable and Legora further underscores a shift in the market. These companies are scaling at speeds that defy historical benchmarks, forcing founders to rethink their go-to-market strategies and product development cycles.
To compete, founders cannot simply iterate; they must build platforms that provide deep utility. Whether it is navigating the complexities of safety rollouts—such as Anthropic's Mythos—or managing the high expectations of enterprise clients, the winners will be those who move beyond the "intelligence" layer and solve specific, high-stakes problems through integrated software ecosystems.
Key Takeaways
- Platform over Model: Superior LLM performance is a baseline requirement, but long-term defensibility (the "moat") comes from building platforms, protocols, and specialized tools like Claude Code.
- Unprecedented Scale: The AI sector is experiencing growth velocities that far exceed the internet, mobile, and cloud revolutions, requiring founders to scale with extreme agility.
- Strategic Ecosystems: Early indicators of success often include deep integration with major cloud providers and the ability to transition from a pre-revenue bet to a platform-driven revenue powerhouse.