VentureBeat Survey Finds OpenAI More Often Becomes the Primary Agent Platform

Installing an AI-agent platform is not the same as making it the system that coordinates production work. In VentureBeat’s August VB Pulse survey, 69% of surveyed OpenAI users called its platform their primary orchestration layer, versus 38% of Anthropic users—while Anthropic drew the strongest consideration relative to its current user base.

The distinction is consequential for companies running agents through multistep tasks. An orchestration platform acts as a control layer: it coordinates which agents run, what tools and data they can access, and what happens if a step fails. A company can test several platforms in its stack without choosing all of them for that central role.

OpenAI’s current foothold, Anthropic’s possible opening

Among the 75 organizations that said they use OpenAI’s Agents SDK or Responses API, 52 identified OpenAI as their primary orchestration platform. Anthropic was primary for 17 of the 45 organizations using its Claude Platform and Agent Skills. Organizations using both platforms could be counted in both groups.

Looking only at the 162 respondents that named one primary platform, OpenAI led with 53 selections, or 33%. Google’s Enterprise Agent Platform followed with 39, or 24%, and Anthropic had 18, or 11%. That measure captures a narrower decision than simple platform availability: where a buyer says the operational center of its agent system sits.

69%OpenAI users naming OpenAI primary

52 of 75 surveyed organizations using OpenAI’s Agents SDK or Responses API named OpenAI as their primary orchestration platform.

38%Anthropic users naming Anthropic primary

17 of 45 surveyed organizations using Anthropic’s Claude Platform and Agent Skills named Anthropic as their primary orchestration platform.

Interest has not yet become a primary role

Anthropic’s counterweight in the survey is its potential pipeline. Of 121 enterprises considering whether to adopt, add, or replace a platform within 12 months, 36 named Anthropic, compared with 45 current users. That consideration-to-user ratio was the highest among platforms used by at least 40 surveyed enterprises. OpenAI drew 41 considerers against 75 users; Google had 30 against 65, and Microsoft 15 against 31.

The survey also suggests that buyers are still making foundational choices. Sixty percent of respondents said they plan to adopt, add, or replace an agent-orchestration platform within a year, including 47 organizations planning a move within the quarter. Those changes could reshape the current rankings, but the August data does not establish a market trend.

Buyers want a control layer without losing flexibility

The findings do not point to a simple rush toward provider-owned control. When asked where their primary agent control plane would sit by the end of 2026, 33% chose a hybrid of provider-native and outside orchestration, while 26% chose a provider-managed agent service. The most cited concern about provider control was inflexibility across models and tools, selected by 28%; vendor lock-in was selected by 16%.

A useful signal, not a market census

VentureBeat’s August wave received 221 responses and qualified 169 people at organizations with at least 100 employees. The sample was self-selected and skewed toward mid-market companies outside technology, and each monthly wave draws a different respondent mix. The results therefore describe this buyer panel’s August cross-section, not the entire enterprise market or a confirmed shift over time.

Still, the snapshot clarifies the competitive question. OpenAI has the stronger reported primary-platform position among its surveyed users. Anthropic has the relatively fuller pool of companies considering a change. The next test is whether that interest turns into control-plane wins—or whether enterprises continue to use multiple providers while reserving the central role for another platform.