A new a16z report looks at which AI companies startups are actually paying for

Andreessen Horowitz's inaugural AI Spending Report, leveraging Mercury transaction data, indicates a highly fragmented and rapidly evolving AI application market where startups utilize diverse tools, with no single dominant solution per category. The report highlights substantial spending on "human augmentors" or copilots, led by OpenAI and Anthropic, though a future transition to fully agentic workflows is expected. A key trend is the increasing adoption of consumer-oriented tools for enterprise use, blurring market boundaries and suggesting faster professionalization for AI companies, underscoring the market's dynamic nature and high churn.

Analysis

The inaugural Andreessen Horowitz AI Spending Report, based on Mercury transaction data, reveals a highly fragmented and dynamic AI application market among startups. The data shows no consolidation around single winners in specific categories; instead, companies are adopting a diverse range of tools, indicating a 'best-of-breed' approach is currently prevalent. While foundational model providers OpenAI and Anthropic command the top two spending positions, the primary use case remains 'copilots' that augment human productivity, rather than fully autonomous 'agentic workflows'. A key trend identified is the rapid consumerization of enterprise AI, where tools initially designed for personal use are being adopted in corporate environments, thereby blurring the total addressable market (TAM) and compelling startups to 'professionalize' and build enterprise-grade features sooner. Horizontal applications represent at least 60% of the top companies, with popular vertical software concentrated in sales, recruiting, and customer service. The report underscores the market's extreme velocity, suggesting high churn where today's leaders could be displaced within 12 months, making long-term predictions challenging despite the sector's overall optimistic growth trajectory.

Key Decisions for Investors

• Given the market's high fragmentation and rapid churn, investors should consider a portfolio or 'basket' strategy for the AI application layer to mitigate single-company risk, rather than attempting to pick one definitive winner.
• Monitor the shift in enterprise spending from 'copilot' tools to 'end-to-end agent' platforms, as this will be a critical inflection point indicating market maturation and a potential re-allocation of value towards companies that successfully monetize automation.
• When evaluating potential investments, prioritize companies that demonstrate a clear strategy to bridge consumer adoption with enterprise monetization, as this 'prosumer' trend is identified as a key accelerator for growth.
• The sustained high spending on foundational models like OpenAI and Anthropic suggests that investing in the underlying infrastructure and large-model providers remains a core strategy to capture broad ecosystem growth, acting as a 'picks and shovels' play on the entire AI application boom.