OpenAI expanded ChatGPT for Teachers on Wednesday to 55 additional US school systems across 20 states, reaching more than 100,000 additional educators and staff. That takes the total to more than 100 K–12 organisations across 30 states and over 300,000 educators. The product stays free for verified US K–12 educators through June 2028.
The new cohort reads like a distribution map rather than a pilot programme: Los Angeles County Office of Education, Baltimore City Public Schools, Fort Worth ISD, Northside ISD, Frisco ISD, Orange County Public Schools, Virginia Beach City Public Schools, St. Vrain Valley Schools. OpenAI says it includes “1 in 5 of America’s 20 largest public school districts.”
The number that actually matters is 16. That is how many states are now covered by a National Data Privacy Agreement OpenAI signed through the Student Data Privacy Consortium framework: Illinois, Iowa, Maine, Massachusetts, Missouri, Nebraska, New Hampshire, New Jersey, New York, Ohio, Rhode Island, Tennessee, Texas, Vermont, Virginia and Washington, with California covered by a separate agreement.
Anyone who has sold software into K–12 knows why that line is the story. The bottleneck was never model quality. It was that every district’s counsel evaluates student-data terms independently, on its own timetable, against its own state statute. A vendor either negotiates thousands of separate agreements or waits. A signed general offer under a consortium framework collapses that into one document a district can simply accept.
Our take: This is a procurement move dressed as a product announcement. Free through June 2028 covers three school years — long enough for lesson plans, gradebook workflows and district training programmes to be rebuilt around one tool, and long enough that switching later feels like a curriculum change rather than a software change. The privacy agreement removes the friction that was slowing that down. Anthropic began offering free Claude access to educators in July. The model comparison is not what decides this.
What the usage data actually shows
OpenAI’s own back-to-school analysis, covering January 1 to July 16, found educators sent more than 1.9 million messages tied to time-saving tasks: roughly 900,000 about report cards and progress reports, 800,000 about lesson planning, and more than 100,000 each about substitute plans and teacher-evaluation materials.
Note what is not on that list. Almost none of it is teaching. It is the administrative layer around teaching — the paperwork that eats evenings and weekends. That is the wedge, and it is a durable one, because it is the part of the job nobody defends.
OpenAI also reports that more than 1,600 teachers, administrators and district leaders joined its AI Skills Jams this summer, that 93% said they left with something they could use again, and that 96% planned to apply it within 30 days. Those are self-reported figures collected by the vendor running the session, and worth reading as such.
What to watch
- How quickly the other 34 states join the consortium framework. That count, not the district count, is the real adoption curve.
- June 2028. A free tier with an expiry date printed on it is a pricing decision waiting to happen.
- Whether districts push to extend access to students. OpenAI has deliberately kept the product to administrators, faculty and educators only — that boundary is where the political risk sits.
- How Anthropic and Google respond. Free educator access is table stakes now, so the fight moves to legal paperwork and procurement, which is a slower and stickier place to compete.