Menlo Ventures has released its 2026 State of Enterprise AI report, revealing that spending on artificial intelligence across businesses climbed to $176.8 billion by September 2026, representing a 5.4-fold increase from the $33 billion recorded in 2025.
The report highlights the growing influence of autonomous AI agents, which are increasingly driving how companies deploy and consume artificial intelligence technologies.
The research combines responses from nearly 400 US enterprise decision-makers responsible for more than $4 billion in AI expenditure with a bottom-up assessment of the enterprise AI market.
Authored by Tim Tully, Joff Redfern, Deedy Das and Derek Xiao of Menlo Ventures, the report examines spending patterns, AI agents, model market share, enterprise applications, infrastructure, open-source development, industry adoption and the balance between purchasing and building AI solutions.
AI Agents Drive a Sharp Increase in Model Usage
One of the report’s major findings concerns the growing consumption of AI models by autonomous agents.
Agent-driven token consumption has increased 37 times since February 2026, when agents overtook human users as the largest source of model traffic.
By comparison, human-driven AI usage grew fivefold over the same period.
The findings indicate that autonomous agents have become a major component of enterprise AI activity, with businesses increasingly relying on them to perform tasks and support operations.
Enterprise Adoption Expands Alongside Spending
Despite the substantial increase in expenditure, the proportion of enterprises paying for AI services recorded a more modest rise.
Menlo Ventures found that the share of companies spending on AI increased from 55% in 2025 to 65% in September 2026.
The report also examined how businesses financed their growing AI investments.
New budgets accounted for 55% of additional enterprise AI spending, while 22% came from existing software budgets.
Most of the expenditure redirected from existing software budgets remained with established providers as they incorporated AI capabilities into their products.
Microsoft Office’s transition towards Copilot was cited as an example of this development.
The research found limited evidence of companies replacing major business systems, including Salesforce’s customer relationship management platform and Workday’s human resources information system.
Enterprise AI Application Spending Climbs to $77 Billion
Spending on enterprise AI applications reached approximately $77 billion in 2026, compared with $19 billion the previous year.
Coding tools represented the largest category, attracting $34 billion, while AI assistants and productivity applications accounted for another $25 billion.
Together, the two categories represented 76% of total enterprise AI application expenditure.
Healthcare emerged as the largest industry-specific market for enterprise AI, with spending reaching $3.7 billion.
AI-Native Providers Lead New Markets as Established Companies Retain Existing Customers
The report identified differences in market share between AI-native companies and traditional software providers across various enterprise functions.
AI-native providers accounted for 96% of enterprise spending on coding applications, demonstrating their dominance in that category.
However, established software companies continued to command substantial shares in more traditional business functions.
Incumbent providers held 91% of IT spending, 88% of data science spending and 68% of HR spending.
The findings show that AI-focused companies have gained a strong position in emerging enterprise workflows, while established providers continue to lead in several existing software categories.
Anthropic, OpenAI and Google Account for 69% of Enterprise AI Spending
Frontier AI laboratories increased their combined share of enterprise AI deployment spending to 69% in 2026, compared with 50% in 2025.
Anthropic, OpenAI and Google collectively accounted for the majority of expenditure in this segment.
Within the enterprise large language model API market, Anthropic increased its share from 40% to 55%, strengthening its position as the leading provider.
OpenAI also recorded growth, with its share rising from 27% to 31%.
Google’s share, meanwhile, declined from 21% to 9%.
Open-Source Models Gain Usage Despite Lower Spending Share
Open-source AI models continued to expand their presence in enterprise deployments, handling approximately one in every five tokens processed.
However, their share of model expenditure remained considerably smaller, accounting for approximately one in every 25 dollars spent by enterprises on AI models.
The report’s findings provide an overview of how enterprise AI spending, application adoption and model consumption evolved throughout 2026, with autonomous agents emerging as a major driver of activity.
