Interview, Fireside Chat
Nabeel Hyatt, GP @ Spark Capital: To Win in AI, Investors Need to Change Their Approach | E1255
- Nabil highlights a critical divergence between the current AI market and the previous B2B SaaS boom: the former is characterized by "mysteries" (unpredictable, requiring judgment and navigation through fog of war) while the latter was a series of "puzzles" (solvable via raw horsepower and standardized metrics).
- The VC industry is currently dominated by principals and associates driven by promotion rather than long-term exits, leading to a "packaging" mentality focused on quick markups rather than deep alignment with founders.
- Traditional heuristics like "18 months to $10M ARR" are obsolete in AI, where products can reach that milestone in months but may become obsolete within two years.
- Nabil argues that the "industrialization of venture capital" is a losing strategy because the market is defined by exceptions and rapidly shifting landscapes rather than pattern matching or Brita-filter efficiency.
- Spark Capital maintains a small team structure (7 partners) to preserve the ability to make subjective, "artisanal" bets and deeply engage with founders, rejecting the scale-up model of larger firms.
- The firm distinguishes between Early Stage and Growth investing as "two different sports," utilizing separate teams with different processes (e.g., early stage focuses on "falling in love" with founders, while growth involves more hierarchical diligence and customer calls).
- Nabil rejects the notion that "the hottest deals don't turn out to be the best," stating that venture capital is inherently about exceptions and that avoiding competitive, high-priced deals often leads to mediocrity.
- The firm evaluates founders based on a balance of "taste" and "execution speed," prioritizing founders who can reinvent themselves continuously rather than those who simply execute on a static roadmap.
- Nabil classifies AI startups into three categories: Adaptation (slapping AI onto existing products, which Spark avoids), Evolution (new workflows/native behaviors like Descript or Granola), and Revolution (new platforms like Uber that only exist due to the tech).
- Spark views ownership of the consumer interface as critical for data advantage, arguing that user insights and "data exhaust" from a well-designed UI are more valuable for model iteration than raw capital alone.
- While acknowledging the risk of DeepSeek-like disruptions, Nabil maintains that capital is not the only barrier to entry; execution speed, product taste, and direct customer relationships remain the primary competitive moats.
- The firm believes that too much capital can be detrimental to a company, potentially destroying value by disrupting hiring velocity, execution speed, and the organic growth trajectory.
- Nabil advises founders to focus on solving "hard jobs" that will remain relevant for a decade, rather than pursuing near-term arbitrage that AI models could solve tomorrow.
- European founders face significant structural disadvantages, specifically regarding the scarcity of "all-in" talent willing to sacrifice for success, making San Francisco the default choice for high-risk, high-reward ventures.
- Nabil suggests that traditional VC metrics (like specific ARR targets) can be counterproductive, advocating instead for helping founders define a visionary story for their next 18 months that exceeds investor expectations.
- In the Q&A, Nabil notes that successful VCs often possess a "nothing to lose" mentality, derived from having the optionality to return to being a founder, which allows for better risk-taking than those focused on fund survival.
- He advises young investors to abandon rigid frameworks and excelists; instead, they must learn to navigate uncertainty with confidence, treating venture as a creative exercise rather than a math problem.