BUSINESSDevelopments · · 2 min read · Source: Menlo Ventures
Menlo report: companies spend 2.5x more on AI than planned
Venture firm Menlo Ventures published its State of Enterprise AI report on November 3, 2026. It finds companies spent on average 4.7 times more on AI than the year before – while planning for roughly double.
My tool tip · AdWispr Flow – AI dictationTry it free
Key facts
- Menlo Ventures published "2026: The State of Enterprise AI" on November 3, 2026.
- Menlo estimates enterprise AI deployment spending reached $176.8B through September 2026 (prior year: $33B).
- Companies spent on average 4.7x last year's level against a planned 1.9x; the survey covered nearly 400 U.S. decision-makers.
The key numbers
- Market size: Menlo estimates enterprise AI deployment spending at $176.8B through September 2026, up from $33B a year earlier (5.4x).
- Budget vs. reality: companies planned on average for 1.9x last year's spend and ended up at 4.7x. The median gap was smaller: 2.2x actual vs. 1.8x planned.
- Use cases: coding ($34B) and chatbots/assistants ($25B) make up 76% of the AI application market.
- Vendors: in enterprise LLM API spend, Menlo puts Anthropic at 55%, OpenAI at 31% and Google at 9%.
- New money: 55% of AI spending comes from new budgets rather than existing software or labor budgets.
Why budgets are bursting
Menlo attributes the gap to AI agents: they work on their own, run in parallel and keep going after people step away. Costs therefore scale with work done, not with headcount. On the OpenRouter platform, agents now generate 80% of all tokens processed, according to the report.
Context you need
The figures come from a survey of nearly 400 U.S. enterprise decision-makers (Morning Consult, September 8–26, 2026) plus Menlo's own estimates. Menlo is an investor in several companies it mentions, including Anthropic and OpenRouter. The numbers aren't a direct benchmark for smaller businesses, but the shift to usage-based AI costs applies everywhere.
What this means for you
- AI costs now grow with the work agents do – not with the number of seats.
- If you roll out agents, you need a spending cap and monthly cost monitoring.
- The figures are U.S. estimates from an investor with stakes in several companies it names.
How to use this
Open the billing pages of your AI services (ChatGPT, Claude, Copilot and so on) and set a monthly cap wherever possible. Add the last three months' total to a dedicated "AI" budget line.
Luigi's take
The most interesting number to me isn't the $177 billion – it's the gap between plan and reality. Anyone who has managed marketing budgets for decades knows the pattern: new channels with usage-based pricing get out of hand fast when nobody sets limits. AI agents are no different.
My advice for smaller companies and solo professionals: set a fixed monthly cap on every usage-billed AI service and review the invoice once a month. And plan AI as its own budget line, not buried under "software" – that's how you see what it actually delivers.
Source
Menlo Ventures: 2026: The State of Enterprise AI
More sources
- GlobeNewswire – Pressemitteilung zum Bericht
This article was researched and written with AI assistance. Responsible: Luigi Lugmayr.
