Andreessen Horowitz (a16z), in collaboration with fintech firm Mercury, released its first AI Spending Report on Thursday, providing an analysis of how startups are allocating capital to AI-native application layer companies. The report, which draws on transaction data from Mercury, examines the top 50 such firms, revealing a fragmented market and a significant focus on tools that augment human productivity.
The data indicates that companies are adopting a variety of AI products for specific tasks, with the market not yet consolidated around a few dominant tools in each category, according to a16z partners Olivia Moore and Seema Amble. Amble stated, \"There\'s a proliferation of tools. It hasn\'t just coalesced around one or two in each category.\" The report also highlights substantial spending on \"human augmentors\" or \"copilots,\" suggesting that startups currently prioritize enhancing workforce productivity over fully autonomous \"agentic\" workflows. Amble projected a future shift towards end-to-end agent tools as technology advances.
Leading AI labs dominated the top of the list, with OpenAI securing the first position and Anthropic following at second. Vibe-coding tools also featured prominently, including Replit at No. 3 and Lovable at No. 18. Other notable coding-oriented entries included Cursor at No. 6 and Cognition, which offers enterprise-focused tools like Devin, at No. 34. While Lovable previously demonstrated higher consumer traffic than Replit, the report shows startups allocate more spending to Replit, partly due to its advanced enterprise features.
The report noted a growing integration of consumer-oriented tools into enterprise environments, citing instances where startups adopt applications like CapCut and Midjourney for business use. Horizontal applications comprised at least 60% of the listed companies, while vertical applications accounted for 40%, with popular sectors including sales, recruiting, and customer service. Moore observed that AI is enabling what were once service firms to evolve into software companies, with examples such as Crosby Legal demonstrating the rapid review of legal contracts.
Amble and Moore anticipate rapid shifts in the AI application landscape. Older companies are integrating AI features to maintain relevance, while new entrants frequently emerge with novel solutions. Amble remarked on the fast pace of innovation, suggesting that even current leading applications may be replaced within a year. This dynamic environment supports a varied selection of AI tools, allowing organizations to select solutions best suited to their operational needs.
