The largest published count of AI agents is over 50 million, all registered in one company’s management tool, while a typical large organization runs about a dozen. With no official total, the answer depends on what you count. Forecasters expect billions in use by 2030.
Top 10 AI agent statistics for 2026
- Over 50 million AI agents are tracked in Microsoft’s agent registry, a total that rose by more than 30% in three months.
- The AI agents organizations use handle about 48 billion actions a year, by one research firm’s estimate.
- Companies built 3 million agents with Microsoft’s low-code tools in a single fiscal year.
- The official MCP Registry lists 59,574 active servers, the connectors agents use to reach external tools and data, by Hostinger’s count.
- An academic index of deployed agents profiles just 30 prominent ones.
- Large organizations run an average of 12 AI agents, according to their IT leaders’ own count.
- Companies with agents in production every month on one major platform went from five to 13 activated agents each in 15 months.
- The number of active AI agents is expected to nearly triple in 2027.
- IT leaders expect their organizations to run 67% more AI agents within two years.
- Organizations are projected to use 2.5 billion active AI agents by 2030.
How many AI agents are there?
Over 50 million AI agents were registered in Microsoft’s agent registry by the end of September 2026, up more than 30% in a single quarter. It’s the largest published count, but it covers only one company’s registry and includes agents that aren’t doing any work.
Microsoft says the registry includes its own agents, agents its customers build, agents found on local devices, and agents from other companies’ platforms, such as Amazon and Google.
Other figures measure activity instead of agents. IDC estimates that the AI agents organizations use today complete about 48 billion actions a year. The company doesn’t publish how many agents are behind that number.
Every published count measures something different: agents registered, created, studied, or used per company. Some counts can also overlap. An agent built on Salesforce’s platform and registered with Microsoft, for example, shows up in both companies’ numbers.
Before you compare any two of these figures, check what each one counts: a registry total, a yearly build count, and a per-company average answer different questions.
How many AI agents have companies built?
Companies built 3 million AI agents using Microsoft’s low-code tools in the fiscal year to June 2025. More than a third of those came in a single quarter: over 1 million custom agents were created with the same tools in January to March 2025, up 130% from the previous quarter.
Both figures count every agent built, including ones no longer running, and they mix different types of AI agents, from assistants that answer questions to agents that take actions on their own. Creation counts show how many agents people have tried building, not how many are at work.
How many MCP servers are there?
Hostinger counted 59,574 active servers in the official Model Context Protocol (MCP) Registry on Nov. 9, 2026. MCP servers are the connectors that agents use to access external tools and data, such as a calendar, a database, or a store’s product catalog.
The count uses the registry’s public data, includes each server once, and leaves out deprecated or deleted entries. Servers that were never registered aren’t in the registry, so the figure is a floor for what’s on offer.
How many notable AI agents do researchers track?
Researchers track 30 prominent deployed AI agents in MIT’s 2025 AI Agent Index. The list is smaller than before on purpose. The index’s first edition listed 67 agentic systems because it included every deployed system that met its definition of an AI agent, while the 2025 edition picks agents for their agency, impact, and practicality.
Of the 30, 24 launched or received major agentic updates since 2024, and only half of their developers publish an AI safety framework.
For a business owner, the index works as a shortlist: the 30 agent products researchers consider most significant.
How many AI agents does a typical company run?
Organizations use an average of 12 AI agents, according to a survey of 1,050 enterprise IT leaders across nine countries, conducted by Vanson Bourne with Deloitte Digital in October and November 2025. In the same survey, 83% said most or all of their teams and functions have adopted AI agents.
Respondents said their agents come from three places, on average:
- 36% are prebuilt agents in software-as-a-service (SaaS) tools.
- 34% are embedded in enterprise platforms.
- 30% are custom-built in-house.
Platform data shows the count rising. In Salesforce’s Agentic Enterprise Index, companies with agents in production every month saw the average number of activated agents increase from five in February 2025 to 13 in April 2026. Building one got faster too: creating an agent took 1.9 days on average, down 53% since February 2025.
Whether IT leaders count their own agents or a platform counts activations, the answer lands in the same place: a large organization runs about a dozen.
How widely do companies use their AI agents?
In a PwC survey from April 2025, 17% of U.S. companies that had adopted AI agents used them across almost all workflows and functions, and 35% used them broadly.
Larger companies are further along. In McKinsey’s 2026 State of AI survey, 40% of respondents at organizations with $1 billion or more in revenue said they’re scaling AI agents in at least one function, up from 27% a year earlier, while the share at smaller organizations stayed at 22%.
LangChain survey data in Hostinger’s AI agent for business report shows the same pattern: among professionals already working with agents, 67% at organizations with 10,000 or more employees had agents in production, compared with 50% at organizations with fewer than 100 employees.
Having agents and using them across the business are two different steps, and smaller companies are further behind on the second. For broader adoption data, from pilots to production, see the latest agentic AI statistics.
How many AI agents will there be by 2030?
According to IDC’s Worldwide Agentic AI Adoption Guide, organizations will use 2.5 billion active AI agents by 2030, nearly 80 times the 2025 number. IDC also expects active agents to nearly triple in 2027. The forecasts below come from two kinds of sources: IDC’s model and what IT leaders in Salesforce’s survey expect for their own organizations.
Whichever forecast you look at, the direction is the same: more agents, doing more of the work.
How to start using AI agents in your business
- Start with one job. The surveys above cover large enterprises running about a dozen agents each. For a small business, one agent that handles one task, like answering customer questions, is enough to learn what works before you add a second.
- Check the tools you already pay for. Prebuilt and embedded agents are two of the three sources of agents in Salesforce’s survey, so look at what your current software includes before building anything new.
Your Hostinger account can be one of those tools. Hostinger Agent, for example, adds specialized AI agents for SEO, marketing, content, sales, and customer communication.
Expert tip
A dozen agents sound small compared to the platform totals, but each usually owns a single job, like answering support questions or drafting a weekly report. Adding them one task at a time, and checking each one’s results before the next, keeps that number manageable.
Ričardas Kudirka
Hostinger Agent Project Manager at Hostinger
However fast the totals grow, what matters for your business is whether each agent you add does its job.
All of the tutorial content on this website is subject to Hostinger's rigorous editorial standards and values.
Daniela is an Off-Site SEO Specialist with extensive expertise in link building, digital PR, and content optimization. She has led international outreach campaigns across the U.S., Brazil, France, and Spain, securing high-quality backlinks. Follow her on LinkedIn.
