OpenAI's data just exposed the enterprise AI gap. It tripled in 6 months. Your firm is probably on the wrong side.

OpenAI dropped two reports yesterday. The numbers should alarm every executive still treating AI as a side project.

Frontier firms—top 10% of AI usage—now generate 8.3x more output tokens per active user than typical firms.

In January, that gap was 2.6x.

It tripled in six months.

The gap is not about access to models. Every enterprise can buy the same GPT tokens. It's about who actually delegated work to agents and who is still asking chatbots how to write an email.

Here's the structural shift most leaders are missing:

Codex—OpenAI's agent tool—generated 64% of all enterprise output tokens in June. That means agents are doing the work, not humans asking questions.

And the fastest growth is not in engineering. Legal usage grew 108x since February. Sales grew 41x. Recruiting 41x. Marketing 26x. Engineering? Just 5x.

The departments you assumed would be last are now the ones running agents at full speed.

The workforce implication is brutal. Early-career employees use AI more than executives. Not survey data—actual usage logs. Six months after adoption, junior staff send 13 more messages per week than leadership.

Your AI strategy is probably built on the assumption that senior leaders drive adoption. They don't. The 26-year-old paralegal reviewing contracts with Codex is your actual power user.

Meanwhile, 68% of enterprises still have no AI policy in place.

This is not an adoption problem. It's a governance crisis. Your junior workforce is deploying agents your board hasn't approved, your CISO hasn't audited, and your CFO can't price.

Audit your token spend today. Map which departments are actually using agents. If you don't know, you're already behind—and the gap is compounding weekly.

SOURCE: https://openai.com/signals/enterprise-data/
VERIFIED: OpenAI Enterprise Signals report (August 12, 2026), VKTR analysis (August 12, 2026), CIO.inc (August 13, 2026)
SIGNAL: The enterprise AI adoption gap is not closing—it's accelerating. Firms that don't move from chat to agents in the next two quarters will face a structural disadvantage they cannot buy their way out of.