Published: August 16, 2026
OpenAI published a 69-page working paper with five academic co-authors analysing 17+ million real ChatGPT Enterprise usage logs, showing token consumption grew approximately sevenfold between June 2025 and March 2026. The paper, “How Organizations Use AI: Evidence from ChatGPT,” is the largest empirical study of workplace AI adoption published to date. Its 2 counter-narrative findings: junior employees — not senior staff — are the heaviest users, and AI adoption is widening corporate inequality, with adopting companies’ median income, market cap, and R&D spending all 10 times higher than non-adopters.
What the 17 Million Sessions Show About Enterprise Growth
Total token consumption in ChatGPT Enterprise increased approximately sevenfold from June 2025 to March 2026. Most of that growth came from existing customers deepening use rather than new signups: among companies that adopted before June 2025, internal token usage rose approximately fourfold. The study, co-authored by Aaron Chatterji and David Holtz with three additional scholars, links the usage logs to employee seniority, task categories, and financial data of publicly traded companies. Token growth measures usage intensity, not revenue or productivity — the paper’s financial linkages are correlational, and the study does not prove AI increases profits.
Junior Employees Are the Heaviest Users
Employees with just a few years’ experience are the most intensive ChatGPT users across all seniority levels, according to the paper. The finding inverts the common assumption that senior workers drive enterprise AI value: the data shows junior staff running experiments and building AI habits fastest, while senior employees use the tools less intensively. For companies planning rollouts of AI tools for business, the data argues for provisioning junior teams first rather than reserving licences for leadership.
Which Companies Adopt AI: The 10x Inequality Gap
AI-adopting companies are not a representative cross-section of businesses. The paper reports that adopters’ median income, market cap, and R&D spending are each 10 times higher than non-adopters — larger, richer, more research-intensive firms are the ones deploying AI at scale. The gap between AI-using and non-AI-using companies is widening, not narrowing. The study covers ChatGPT Enterprise only; usage patterns for competing platforms, such as Claude Enterprise and Gemini for Workspace, are outside its data.
Our Take: The headline number is 7x token growth, but the real story is the inequality signal: companies that are already rich and well-resourced are the ones betting on AI, and the gap is widening. If you run a small business not yet on AI tools, this study says you are falling further behind every quarter — not catching up.
For Context: Enterprise AI Adoption at Scale in 2026
The paper quantifies a trend WithO2 has tracked through individual company moves. Cognizant certified 30,000 staff in the largest single enterprise AI adoption program to date, and Ode with Anthropic built a $5B business on enterprise AI implementation alone. OpenAI’s dataset covers the paid enterprise tier; on the consumer side, ChatGPT free users received unlimited chats this month — a separate adoption channel the working paper does not measure.
Related: 15 Best AI Tools for Business in 2026 · 12 Best AI Agents for Business Tasks
