Menlo Ventures released its 2026: The State of Enterprise AI report, finding that enterprise AI spending reached $176.8 billion, up 5.4x in a year as autonomous agents transformed how businesses use and pay for AI.
“Wall Street is questioning whether the massive investments in AI will pay off, but our latest data shows just how quickly enterprise demand is growing,” said Tim Tully, Partner at Menlo Ventures. “Enterprise AI spending has jumped more than fivefold, from $33 billion last year to $177 billion this year. Based on what we’ve heard from enterprise buyers, companies aren’t pulling back. They’re finding more ways to put AI to work and agents are driving a lot of that demand. If anything, we may still be underestimating just how big this market can become.”
“Agents are changing the economics of software. When AI can do the work itself, enterprise spending is no longer constrained by the number of employees using it, but by how much work it can get done. Companies entered 2026 expecting to spend more on AI, so they budgeted nearly twice what they’d spent in 2025. But even that wasn’t enough! Our survey found they ended up spending 2.5 times what they’d budgeted.” said Derek Xiao, Principal at Menlo Ventures. “That demand is growing faster than even the investment required to build these models. For every dollar spent on AI model training this year, enterprises are spending $1.70 putting AI to work.”
Key Findings
- Enterprise AI spending more than quintupled. Spending surged from $33 billion in 2025 to $176.8 billion by September 2026, accelerating from 3.9x growth the year before. Yet the share of enterprises paying for AI rose only modestly, from 55% to 65%, suggesting that growth is being driven primarily by greater spending among adopters rather than a wave of new buyers.
- Agents are now the biggest consumers of AI models. On OpenRouter*, agents generate 80% of all tokens, roughly seven times human-driven usage. OpenRouter is one channel, but it shows how quickly AI consumption is shifting from humans asking questions to machines doing work. Since overtaking human traffic in February, agent token consumption has grown 37x, compared with 5x growth in human-driven usage.
- AI is expanding the software market, not replacing it. New budgets funded 55% of incremental enterprise AI spending, while 22% came from existing software budgets, most of it staying with incumbents as they convert licenses to AI offerings, such as Microsoft Office to Copilot. The research found little evidence of enterprises replacing core systems such as Salesforce’s CRM or Workday’s HRIS.
- Applications are becoming a major engine of growth. Enterprise AI application spending reached approximately $77 billion, up from $19 billion last year. Coding ($34 billion) and AI assistants and productivity tools ($25 billion) account for 76% of application spending. Healthcare remains the largest vertical industry market at $3.7 billion.
- AI-native companies dominate emerging workflows, while incumbents retain an advantage in established ones. AI-native providers capture 96% of coding spending, while incumbents lead IT (91%), data science (88%), and HR (68%). The divide illustrates how AI is creating new markets even as established software providers maintain their positions in existing workflows.
- Frontier labs capture 69% of enterprise AI spending. Anthropic*, OpenAI, and Google together account for 69% of deployment spending, up from 50% in 2025. Within enterprise LLM API spending, Anthropic extended its lead from 40% to 55%, OpenAI rose from 27% to 31%, and Google fell from 21% to 9%. Open-source models now handle approximately one in five tokens but account for only about one in 25 dollars of model spending.
“Right now, the best models are capturing the lion’s share of enterprise,” said Deedy Das, Partner at Menlo Ventures. “Our data shows that frontier labs capture nearly seven of every ten dollars spent on enterprise AI, up from half last year. Anthropic grew its share of enterprise API spending from 40% to 55%, while Google fell from 21% to 9%. That’s how quickly the competitive landscape can change. The labs have a tremendous advantage because building frontier models is incredibly hard, but no one’s position is guaranteed. Every major model release is an opportunity to shift market share.”
“AI isn’t disrupting every software category in the same way,” said Joff Redfern, Partner at Menlo Ventures. “In coding, AI-native companies have captured 96% of the market because they’re fundamentally changing how software gets built. In established categories like IT and HR, incumbents still have the advantage. Our data shows when AI changes how work gets done, new companies have an opening to reshape a market. When AI improves the way people already work, established players are harder to displace.”
