According to saastr.com, Menlo Ventures released its fourth annual State of Enterprise AI report, revealing that enterprises will spend

76.8 billion on AI deployment in 2026, a 5.4-fold increase from $33 billion in 2025. AI spending accounts for about 30% of public cloud revenue, up from roughly 6% one year prior.

Majority of AI spending comes from new budgets rather than software line items

Menlo’s survey of approximately 400 U.S. enterprise decision-makers controlling over $4 billion in AI spending found that 55% of incremental AI budget is net-new, with only 22% coming from existing software budgets, and the remainder from labor and other lines. Of the 22% software budget share, 13 percentage points go to incumbents converting existing seat licenses into AI products, while 9 percentage points (

2.6 billion) are captured by AI-native vendors.

AI spending distribution highlights competitive dynamics between incumbents and startups

The report indicates that competition between AI-native startups and legacy seat-based software vendors over software budgets is a relatively small segment, mostly won by incumbents. Menlo emphasizes that B2B AI products priced at least twice the traditional seat price and still accepted by customers generally tap into the 78% of budgets outside the software line. This aligns with SaaStr’s experience where AI agents replaced human headcount, boosting revenue despite a smaller human team.

AI application categories show strong growth in coding and workplace assistants

Menlo estimates coding-related AI spending at $34 billion (7.7x growth), and workplace assistants and chatbots at