EnterpriseOCT 15, 2026

OpenAI Data: EU Small-Business Agent Users Up 8.1x Since May, Trail North America by 6.5x

OpenAI Economic Research says weekly active agentic users at EU SMBs grew 8.1x by Oct. 5, with Codex and ChatGPT Work generating 55% of enterprise output tokens — but North America still houses 65% of frontier firms.

BY GRETA REINHART · ENTERPRISE SAAS

Weekly active users of OpenAI’s Codex and ChatGPT Work at EU small and mid-sized businesses grew 8.1x between early May and Oct. 5, according to an OpenAI Economic Research report published Oct. 14. Every EU member state cleared a 5x increase. Belgium led at 12.1x, with Italy at 10.8x and Luxembourg at 10x.

The report defines SMBs as accounts with 200 or fewer associated users, which captures the small-team operator band (roughly 5–30 employees) where owner-founders typically run sales and marketing themselves. Agentic tools now account for more than 50% of output tokens across the EU ChatGPT Enterprise and ChatGPT Business sample, with Codex and ChatGPT Work together generating 55% of enterprise output tokens. OpenAI attributes the step-change to the July launch of ChatGPT Work, which bundled the agentic layer into the consumer-facing superapp.

Country-level intensity varies sharply. Lithuanian firms route 85% of output tokens through agents, Cyprus 81%, Hungary 79%. At the bottom end, Slovenia sits at 22%, Austria at 26%, Portugal at 29%.

“Europe is now in a new era of AI, with tools that can carry out tasks for people,” said OpenAI CFO Sarah Friar.

The headline growth number masks the structural story. EU-based companies account for roughly 10% of global “frontier firms”, the top 10% by output tokens per active user. North America holds 65%, APAC 13%, broader EMEA 11%. Between May and September, non-EU frontier firms grew 12.9x. EU frontier firms grew 6.1x. By September, frontier firms outside the EU were producing more than twice the output per active user of their EU counterparts.

Read against OpenAI’s own corporate chronology, including the managed-agents and always-on assistant rollout at DevDay and this month’s $50B correction, the EU data documents a mainstream threshold crossed. Small operators still hand-prompting a chat window are now measurably behind peers whose software executes the work unsupervised. The 6.5x concentration gap is the compounding version of that lag.

Sources

About the author

Greta Reinhart

ENTERPRISE SAAS

Greta Reinhart tracks the enterprise software stack from San Francisco — data platforms, AI bundling, seat pricing, and channel checks across the largest SaaS vendors. She files on go-to-market shifts, packaging changes, and quarterly enterprise reads.