2026: The State of Enterprise AI (Menlo Ventures, November 2026): enterprise AI spend hit $177B, up 5.4x

Menlo Ventures puts enterprise AI deployment spend at $177 billion by September 2026, against $33 billion a year earlier, and growth is accelerating. Only 65% of enterprises pay for AI (55% last year), so existing buyers drove the increase. Agents drove it: agent traffic on OpenRouter is up 37x since it overtook human traffic in February and now produces about 80% of tokens. Frontier labs take 69% of deployment dollars, and in API spend Anthropic went from 40% to 55%, OpenAI from 27% to 31%, Google from 21% to 9%. Open-source models handle 19% of usage but 4% of spend.

The application layer reached $77 billion, with coding ($34 billion) and assistants ($25 billion) making up 76% of it; coding is 96% AI-native, while incumbents hold IT, data science and HR. 55% of AI dollars are net-new budget, not cannibalised software or labor spend.

For founders the lesson is where agents can act, learn a workflow and verify results, spend follows. Pick a category with those properties, price against work done, and expect incumbents to defend the systems of record.