OpenAI leads enterprise agent platforms, but Anthropic has the stronger pipeline

OpenAI is more likely than Anthropic to become an enterprise's primary AI agent platform, according to an August survey reported by VentureBeat. The more interesting wrinkle is that Anthropic has the stronger pipeline of organisations considering a switch or addition, so the platform race is not quite the tidy victory lap the headline numbers suggest.

Watch Desk analysis

What happened

VentureBeat's VB Pulse wave surveyed 221 respondents and qualified 169 people at organisations with at least 100 employees. Among the 75 enterprises using OpenAI's Agents SDK or Responses API, 52, or 69%, named OpenAI their primary orchestration platform. Anthropic's Claude Platform and Agent Skills were primary for 17 of 45 users, or 38%.

Across the 162 respondents who named a primary platform, OpenAI led with 53 selections, or 33%, followed by Google at 39, or 24%, and Anthropic at 18, or 11%. Read the full VentureBeat analysis for the survey's question bases and methodology.

Key findings

• OpenAI wins the installed-base conversion
  69% of its users in the survey called it primary, compared with 38% for Anthropic.
• Anthropic has the livelier pipeline
  36 organisations were considering Anthropic against 45 current users, the strongest consideration-to-use ratio among platforms with a substantial installed base.
• The market is choosing control planes
  26% expected a provider-managed agent service to host their primary control plane by the end of 2026, while 33% preferred a hybrid of provider-native and external orchestration.
• Lock-in remains the chief worry
  Inflexibility across models and tools was the most common concern about provider-controlled orchestration, selected by 28% of respondents.
• This is a snapshot, not a league table
  The wave was self-selected, skewed towards mid-market organisations outside technology, and used a different respondent mix from earlier months.

Why it matters

The useful distinction is between installing an AI platform and building the company's agent machinery on top of it. The latter is the stickier decision: it determines which tools agents can call, how workflows fail and how difficult it becomes to move later.

OpenAI's lead therefore says something about current commitment, while Anthropic's consideration figures say something about future pressure. Enterprises may be adding Anthropic alongside an existing platform rather than replacing it, which makes the market look less like a single-winner contest and more like an increasingly expensive committee meeting.

Our read

Treat this as evidence about platform positioning, not proof that OpenAI has won enterprise AI. OpenAI currently converts more users into primary-platform customers in this survey. Anthropic has a credible route to close the gap if consideration turns into primary deployments.

For buyers, the practical question is not simply which model answers best today. It is which orchestration layer leaves room to change models, tools and governance without rebuilding the house.

What to watch

• Whether Anthropic's consideration converts into primary-platform deployments.
• Whether OpenAI can retain its lead as enterprises adopt multi-provider stacks.
• Which vendors publish clearer portability and interoperability commitments.
• Whether later VB Pulse waves reproduce this pattern with comparable samples.

Discussion spark: If you were choosing an enterprise agent platform today, would current primary use or future flexibility matter more?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.