New Enterprise Technology Research data, reported by the Wall Street Journal today, shows how enterprise AI adoption has shifted over the past twelve months.
The percentage of enterprises currently using AND planning to continue using each model:
Source: Enterprise Technology Research survey of ~500 respondents, via Wall Street Journal
1. OpenAI Lost Ground YoY For the First Time, Albeit Modesly. It’s Still #1 in Enterprise with 62% Share
OpenAI peaked at 62% in September 2025 and dropped to 56% by March 2026. It’s still #1 in enterprise share, but the lead over Anthropic has compressed from 41 percentage points a year ago to 8 points today.
2. Claude Adoption More Than Doubled From 21% to 48% Share
21% to 48% in twelve months. ETR attributes much of this to coding assistants, which the firm identifies as the hottest front for competition among the major labs and the area driving rapid revenue growth at the model companies.
3. Gemini Adoption Grew From 27% to 40% Share
Google moved from 27% to 40%. Distribution through Google Workspace, Vertex AI, and BigQuery is the most likely driver, since adoption inside those accounts requires no new vendor relationship.
4. Grok Remains a Rounding Error
7% adoption, up from 4%. Per ETR’s Erik Bradley, Grok is barely growing within enterprise organizations while Claude and Gemini use is soaring.
5. The Coding Assistant Market Is Driving the Numbers
ETR’s commentary in the WSJ piece is direct: corporate adoption of coding assistants is driving rapid revenue growth at the labs. That’s consistent with what we see across portfolio companies and at SaaStr internally. Code generation is high-volume, high-token, and the workload pulling enterprise spend forward fastest.
Multi-Modal Is The New Default
- Multi-model is the new default. With four credible enterprise AI providers and a clear #1, #2, and #3, single-model architecture is now a procurement liability.
- Second place is a real business. Anthropic’s enterprise share grew faster than any other model in the survey while still sitting behind OpenAI. The economics of being the credible alternative in AI are strong.
- Distribution still matters. Gemini’s growth happened inside accounts Google already had. The fastest enterprise AI motion in 2026 is still “we already have a contract.”
The model layer is no longer a one-horse race. It hasn’t been in quite some time.
