How the AI bubble ate Y Combinator (www.inc.com)

Since ChatGPT’s November 2022 debut, Y Combinator has become a near-monoculture for AI startups — a reflection of Silicon Valley’s sprint to create the next billion-dollar AI company. YC, which offers mentorship and a $500,000 seed check, packed its Summer 2025 batch with 170 companies, 154 of which are focused on AI. The cohort spans AI agents, voice assistants, video production tools, AI-native browsers and infrastructure tooling like Interfere, an automated debugging startup, underscoring an unprecedented concentration on a single technical domain in YC’s 20-year history. That concentration matters for the AI/ML community because it accelerates both innovation and systemic risk. On the upside, intense focus drives rapid iteration on models, data, inference stacks and developer tooling — boosting demand for compute, specialized datasets, and model-evaluation frameworks. On the downside, the herd amplifies duplication, talent and capital misallocation, and valuation bubbles that could distort long-term research priorities. For practitioners and investors, the takeaway is pragmatic: expect vibrant tooling and fast product cycles but also heightened competition, narrower exploration of non-AI problems, and an increased need for robust benchmarking, infrastructure, and risk-aware funding strategies.