Why AI Rewrites the Rules of Business Moats
Key points
Key takeaways from Source interview with Mighty Capital Managing Partner Sc Moatti (September 2026):
AI-era moats. Mighty Capital mapped 550 companies. It ranked each firm by moat defensibility and capital raised. Two firms combine heavy funding with strong defensibility: OpenAI and Anthropic. The moats of the SaaS era, such as switching costs and data moats, no longer hold. Network effects and counter-positioning now dominate.
Capital-efficient serial founders. Mighty Capital invests in serial entrepreneurs. It targets capital-efficient, fast-growing, fairly valued companies. Many founders prefer raising less money to limit dilution. They see AI as a 10-15 year platform shift. Several M&A exits can beat a single IPO.
The IPO bar has moved. The IPO market is shrinking. Passive indexing pulls capital from the active managers who underwrite IPOs. Analyst coverage concentrates on large caps. Companies now need 30-50 billion USD of value before listing. A 5-10 billion USD floor worked for early Mighty Capital exits, which include Amplitude, DigitalOcean, Net Scope. Its latest exit was a 20 billion USD licensing deal between Groq and Nvidia.
Power-law strategies fit mega funds. The power-law model suits mega funds. These funds need large tickets per company. Funds under 1 billion USD in assets under management can write smaller tickets. Such tickets can still yield strong returns when companies exit for 100 million to 1 billion USD.
M&A is increasingly product-led. Chief product officers at Salesforce, Disney, Walmart, Johnson and Johnson must grow fast and adopt AI. M&A follows product innovation, not EBITDA multiples. Acquirers weigh culture, roadmaps, innovation synergies. They use acquisitions to change culture from within.
The two AI KPIs. Two AI metrics matter. First, revenue per employee: keep headcount flat, triple revenue using AI. Second, inference cost per unit of customer value: it should fall 20-30x per year. Token costs drop roughly 10x per year.
Signals from product buyers. Mighty Capital runs a Product Signal Intelligence System. It listens to conversations among roughly 600,000 product leaders, about one third of product managers worldwide. ML, NLP, LLM turn these conversations into human-readable signals. The fund invests in buyers close to innovation, not the builders.
The future VC firm. Venture firms adopt AI in three ways. Good firms automate current workflows. They face maximum resistance. Great firms build AI-native roles around product builders. The best take an allocator view. They deploy people, agents, capital to priority problems. Sourcing stays human. Underwriting becomes AI-assisted. Firms that celebrate imperfect first versions outperform.
Read more: YouTube - How I Invest Podcast
