Interview, Fireside Chat
Windsurf: The Making of a Billion-Dollar AI Company | Leigh Marie Braswell, Kleiner Perkins
- The team plans to launch "Windsor," a product forked from VS Code, by owning user identity to serve both consumer and enterprise users, aiming to create a "much better product."
- Ambitions include making developers "10X better" and expanding the user base to "Make Everyone a Developer" through AI-driven capabilities.
- Financial deployment involves an $800 million KP21 venture fund and a $1.2 billion Select 3 growth fund, with a strategy to maintain similar fund sizes due to the relationship-focused and unscalable nature of early-stage venture capital.
- AI model costs are projected to decrease by a factor of 10 annually, unlocking new automation opportunities across various sectors.
- The firm intends to frequently update its mental models of companies by tracking engineer movements to identify "talent vortexes" and invest in or incubate companies in battleground spaces like enterprise knowledge discovery, medical scribes, and customer support.
- "Seedstrapped" companies, defined as entities raising a seed round, achieving massive growth with small teams, and skipping to Series C or higher, are expected to become a new standard where the $100M ARR milestone effectively becomes a $1M ARR mark.
- Despite self-sufficiency claims, "Seedstrapped" companies are predicted to eventually raise additional capital to build sales teams and compete in crowded markets.
- Significant investment opportunities are anticipated in domain-specific automation for finance, supply chain, biotech, and legal workflows, including the automation of financial analyst tasks like PowerPoint and Excel generation.
- In the near term, "speed" is expected to be the only reliable moat for AI coding tools like Windsurf, with hopes that this translates to switching costs or network effects later.
- Strong engineering cultures and high employee retention are viewed as critical "unfair advantages" for navigating crowded markets.
- Investors anticipate frequent instances of companies fudging numbers or misrepresenting ARR as the sector enters "bubble territory," citing specific examples such as 11X and Frank.
- The only reliable "moat" for AI coding tools in the immediate future is speed, though the firm hopes this will translate to switching costs or network effects later.
- "Vibe coding" is predicted to become powerful for non-technical users to build complex workflows quickly, not just for developers.
- A product-market fit is characterized by a founder's sense of constant urgency, described as feeling like "everything is on fire."
- Successful participation in the fastest-moving deals requires a prepared mind and an existing read on the founder; without these, investors "will just not be in the mix."
- The firm intends to maintain relationship-focused investment strategies through activities such as "poker nights" with modest buy-ins ranging from a few hundred dollars or less to foster an environment focused on building connections.
- The team expects to maintain relatively similar fund sizes in the future because they believe early-stage venture capital is relationship-focused and "unscalable."