NYT: OpenAI puts run rate near $50B as safety fights escalate

Key Takeaways

1. DealBook reported OpenAI’s annualized revenue near $50 billion as of late September.
2. Axios said the $50 billion figure is about $20 billion below earlier reports.
3. AP reported OpenAI fired three safety researchers on Oct. 9.
4. Valley buyers of ChatGPT seats and GPT credits should expect attention on pricing and compliance.

$50 billion is the number OpenAI gave DealBook on Oct. 9. The number is big. For Valley CFOs and IT directors paying for ChatGPT or building on its APIs, the growth story and the safety headline matter because they point to where pricing, enterprise controls, and contract terms could move this fiscal year.

What DealBook reported

The Times’ DealBook column said OpenAI’s annualized revenue rate is about $50 billion as of September, up from roughly $28 billion at the end of June. The company had talked earlier this summer about a push toward a $100 billion pace by year end, so the new figure reads as both a surge and a step short of the loftiest chatter. Investors and procurement officers read those gaps differently, since ARR is a blunt metric that can include very different contract shapes.

Axios, citing prior reporting and accountant caveats, framed the new $50 billion run rate as roughly $20 billion below some earlier expectations tied to looser counting. Translation for buyers here: volatility in headline numbers doesn’t change what shows up on your invoice, but it often feeds sales tactics and discount windows.

Safety questions in the same week

On the same Oct. 9 news cycle, AP reported OpenAI fired three safety researchers after internal disputes about risk. Less than a week later, multiple states sued over alleged consumer harms, and Axios reported new FTC subpoenas seeking information on model safety and real‑world impacts. OpenAI’s own posts in recent weeks have acknowledged security incidents and promised tighter controls, which tell you the company knows regulators are watching closely. A lot, if you’re the one signing the check.

Why it matters along Highway 99

Across Fresno and Kern counties, employers I talk to are piloting small batches of ChatGPT Enterprise seats or gating GPT access behind Azure credits, usually 25 to 200 users, which is where pricing tiers and compliance promises matter most. UC Merced labs and Fresno software shops are experimenting with language models for code review, summarization, and customer support, which pushes data‑handling questions to the front of the contract. If OpenAI keeps chasing rapid top‑line growth while fielding safety probes, expect two immediate local effects: procurement will ask for stronger logging and retention controls, and finance will press vendors on multi‑year price locks to protect 2027 opex. And it all lands in procurement.

For the Valley’s mid‑market buyers, the practical checklist is simple this quarter: confirm data isolation terms in writing, pin down support SLAs, and price out an exit ramp to an alternative model in case policy or pricing shifts. One Fresno co‑working kitchen had a half‑full Jarritos by the keyboard when I stopped in, which tells you people are working through lunch on this stuff.

The headlines might be national, but the decisions get made in small rooms, under fluorescent lights, with a redlined MSA on the table and a sticky note that says “tokens?” taped to a laptop.

Central Valley AI is produced by the CVAI Business Desk team and developed by Kaweah Tech, a regional firm that builds, deploys, and integrates AI solutions for businesses across California's Central Valley.

Source

https://www.nytimes.com/2026/10/09/business/dealbook/openai-revenue-safety.html