In 2025, the generative AI market reached $53.7 billion, and OpenAI held 23.6% share, according to Global Market Insights. That momentum will hit classrooms next. The sales pitch will be slick, the promises big. Districts will see the familiar pattern of Big Tech in schools: integrate now, figure it out later.
What the market data says about the AI push
Global Market Insights expects the market to grow from $83.3 billion in 2026 to $988.4 billion by 2035, a 31.6% compound annual growth rate. North America is the largest market; Asia Pacific is the fastest growing. The top five vendors—OpenAI, Microsoft, Google, Amazon Web Services, and NVIDIA—controlled 71.03% of the market in 2025, per the same report. That level of concentration shapes pricing power, support terms, and how quickly new features get bundled into education suites.
The report also notes that enterprises are adopting generative AI through managed APIs, SaaS copilots, and retrieval-augmented generation tied to internal data. Those channels already exist in education platforms from the same companies. Expect more district-facing bundles where a productivity suite quietly becomes an AI suite, with usage-based pricing layered on top.
Schools should read this trend line as a procurement signal, not a fait accompli. When a small group of vendors controls most supply, lock-in risks grow and exit ramps shrink. That matters in classrooms where switching tools midyear can derail instruction time.
Why this matters for budgets and classrooms
Budgets are finite. A high-growth market tends to push for rapid adoption, then upsell. For districts, the real cost of an AI feature is rarely the first-year license. It’s the training hours, the integration work, and the data governance you must maintain for years. In a concentrated market, those recurring costs can escalate faster than instructional value.
There’s also the data question. Student records are protected by U.S. law through FERPA, which restricts disclosure without consent. Many AI tools improve by learning from user inputs. Districts must lock down contracts so no student content becomes model training data, and that logs are purged on a clear schedule. Groups like the Electronic Frontier Foundation have warned for years that education tech can drift into surveillance. AI features raise the stakes, because they capture more context, faster, and often by default.
Risk management frameworks can help. The NIST AI Risk Management Framework outlines practical steps for identifying and managing risks across the AI lifecycle. It’s written for broad use, but districts can adapt it: define intended use, stress-test for failure modes, and assign a data steward. UNESCO has urged public bodies to set guardrails before scaling AI in education; its guidance emphasizes human oversight and transparency over hype (UNESCO policy guidance).
Put plainly: Big Tech in schools is a procurement problem wrapped in a pedagogy promise. If the value isn’t clear, the safest timeline is slower than the vendor’s.
Big Tech in schools: five guardrails for district buyers
- Prove learning value before scale: Run short, bounded pilots with defined success metrics tied to instruction time, equity, or teacher workload. Share the rubric publicly. No expansion without measured gains.
- Lock down student data in writing: Require a data processing addendum that bans training on student data, spells out retention and deletion windows, and commits to breach disclosure timelines that exceed legal minimums. Tie access controls to least privilege.
- Insist on interoperability: Demand exportable data in open formats, clear API documentation, and single sign-on via open standards. Make renewal contingent on proof that your content, logs, and models can migrate without penalty.
- Control usage and costs: Cap monthly spend, turn features off after hours, and require real-time dashboards for usage, prompts, and error rates. Put price increases behind a multi-year ceiling.
- Require transparency and audit trails: Vendors should disclose model versions, update dates, content filters, and human-in-the-loop points. Keep an internal risk register and review it before each feature toggle.
These steps won’t stop innovation. They make sure innovation serves instruction, rather than the other way around. They also give districts practical leverage if a vendor pushes premature features or surprise fees.
Keeping classrooms about learning as AI vendors target schools
There’s a right order to this work: set goals, then pick tools. When the market swells as fast as Global Market Insights projects for 2026 through 2035, the temptation is to backfill goals later. Resist that. Start with a curriculum problem statement. Define what “better” looks like and how you’ll measure it. Then ask whether the same outcome can be reached with existing tools, improved teacher planning time, or targeted professional development. Many districts will find that the first 80% of value comes from clearer workflows and data hygiene, not a new license.
Vendors will tout time saved, fewer clicks, and slick copilots. Ask for the evidence, not a testimonial. If a product claims to cut grading time by 30%, require a study design you can repeat. If it claims to raise reading outcomes, demand disaggregated results and methods.
None of this argues against AI forever. It argues for agency. Districts should own the decision-making cadence, the data boundaries, and the definition of success. That’s how you keep Big Tech in schools aligned with a school’s mission to teach, not the other way around. For more on this, see bloomberg.com and nytimes.com.