After being rejected by investors nearly a hundred times, this education AI company has raised nearly 63 million US dollars in financing.

After nearly a hundred financing meetings and repeated rejections from investors, a former principal grew his education AI business and secured nearly $63 million in financing.

Recently, Crunchbase disclosed the latest entrepreneurial case of MagicSchool AI. Founded by Adeel Khan, a former U.S. principal, the company now has about 8 million registered educators worldwide, with investors including Bain Capital Ventures, Valor Equity Partners, Smash Capital, GSV Ventures and others.

In three years, MagicSchool has evolved from an AI tool for teachers' daily use to cover scenarios for students and school districts. A project that once struggled to gain capital recognition has gradually become one of the larger startups in the K12 education AI track.

01

8 million teachers registered in three years,

MagicSchool starts with lesson preparation tools

Before founding MagicSchool, Adeel Khan had worked in the U.S. public education system for many years. He started as a teacher in Atlanta Public Schools, then served as an assistant principal, and later founded a public high school in Denver. After leaving school, he was responsible for principal guidance at the school district office.

Adeel Khan, Founder of MagicSchool AI

In November 2022, Khan went on vacation. Coinciding with the release of ChatGPT, he returned to the high school he founded to test this new technology, and organized teachers to try to apply it to different teaching tasks.

Teachers' feedback was inconsistent. Most of them rarely used it, and some teachers believed that the time spent entering requirements and modifying generated results was no less than completing the tasks themselves; a small number of teachers quickly integrated ChatGPT into teaching. One teacher told Khan that this technology "completely changed the way I teach".

In 2023, Khan founded MagicSchool, breaking down the capabilities of general large models into specific teacher tasks. Teachers can directly generate grading rubrics, design differentiated assignments for different learning levels, and also create practice worksheets and reading materials. About three months after the product was launched, about 100,000 teachers were using it, and the platform provided more than 50 AI tools for teachers at that time.

Since then, MagicSchool's user base has expanded rapidly. By the time of its Series A financing in 2024, the platform had more than 2 million registered educators. According to data disclosed by Khan in this Crunchbase interview, there are currently about 8 million registered educators worldwide.

The product coverage has also extended from the teacher side to the student side. Teachers can create managed AI usage spaces for students in MagicSchool, where students can use tools such as algebra tutoring and writing assistants. Among them, the writing tool can be customized according to state-level examination grading standards, provide feedback based on corresponding standards, and assist students in revising their articles.

Apart from individual teachers, MagicSchool has also started to enter schools and school districts. Currently, the company has cooperated with large school systems such as Denver Public Schools, Broward County Schools in Florida, Hillsborough County Schools, and some private educational institutions. Khan said that about one in five children in the United States attends a school that has established a partnership with MagicSchool.

From teacher lesson preparation and material generation, to student use, and then to deployment in schools and school districts, MagicSchool's product coverage continues to expand.

02

Nearly $63 million in financing,

MagicSchool moves toward school districts

When it had nearly 1 million users, MagicSchool still did not have a mature business plan.

Khan's initial idea was to expand the user base first and then find a business model. The early team mainly relied on hourly-paid contractors to develop products, and he did not start to reach out to venture capital institutions intensively until the number of users approached 1 million.

The financing process was not smooth. Khan recalled in an interview with Crunchbase that he went through nearly a hundred financing meetings before getting institutional investment. Compared with Silicon Valley entrepreneurs familiar to investors, he had no background from top universities such as Stanford, was not a technical person, and had no work experience in large technology companies such as Google. Even some investment institutions focusing on edtech were initially cautious about the project.

This experience once confused Khan. At that time, some edtech companies of similar size to MagicSchool had received millions of dollars in seed funding before their products and sales reached scale; MagicSchool already had a large number of real users, but still found it difficult to obtain investment.

What finally changed the financing situation was the continuously growing user and business data. Khan said that in every round of financing, he relied on growth indicators to prove the company's market performance to investors.

In 2023, MagicSchool raised $2.4 million in seed financing led by Range Ventures. In 2024, the company completed a $15 million Series A financing led by Bain Capital Ventures. In 2025, MagicSchool completed another $45 million Series B financing led by Valor Equity Partners, with participation from Bain Capital Ventures, Adobe Ventures, Atreides Management, Smash Capital and other institutions. According to the data disclosed by Crunchbase this time, the company's cumulative financing has approached $63 million.

In addition to user growth, MagicSchool has also started to generate revenue. Khan revealed this time that by the end of last year, the company's revenue increased by about 3 times year-on-year. Currently, its main customers include school districts that hope to deploy generative AI in a unified manner, where schools can provide managed AI tools for teachers and students in an environment that complies with data privacy rules and local curriculum requirements.

MagicSchool's business model has thus gradually become clear. Currently, the product retains both a free version and paid subscriptions for individual teachers, and also provides enterprise versions for schools and school districts. Compared with the generation tools used by individual teachers, the school-side version adds functions such as unified management and deployment, and the enterprise version is priced separately according to the scale of schools and school districts.

After receiving the new round of funding, MagicSchool is still continuing to expand its investment for schools. During the Series B financing in 2025, the company stated that the funds will be used to expand the product and engineering teams, strengthen the school cooperation and customer experience teams, and continue to invest in product safety and testing.

From struggling to obtain institutional investment with nearly a million users, to successively completing seed round, Series A and Series B financing, MagicSchool's financing path has changed, and its business has gradually moved from acquiring teacher users to procurement by schools and school districts. As its customer base expands from individual teachers to educational institutions, the product demands faced by MagicSchool are no longer limited to lesson preparation, question setting and material generation.

03

Beyond the model,

educational experience becomes a new competitive advantage

As the capabilities of general large models improve, MagicSchool is also adjusting its product priorities.

Khan said in an interview with Crunchbase that the model itself is no longer the main differentiator, and what is more important is how to combine the model with problems in specific work and professional knowledge. MagicSchool positions itself as a "vertical AI" solution for K12 schools. Khan compares it to Harvey and Legora in the legal industry: the latter provides AI products around the specific work of law firms, while MagicSchool hopes to cover the teaching and management scenarios of K12 schools.

Since the beginning of this year, MagicSchool's product updates have continued to extend to teachers' daily work.

In June, the platform updated Class Writing Feedback, allowing teachers to use more flexible grading rubrics and support importing student assignments from Canvas, Schoology and OneDrive. For Enterprise customers, schools can also create and share customized student AI tools within the organization.

MagicSchool is also strengthening its connection with the original teaching systems of schools. Currently, the platform already supports commonly used school tools such as Google Classroom, Canvas, Schoology and Microsoft. Teachers can bring existing courses and teaching materials into MagicSchool, and can also continue to use the generated content in the original teaching environment, reducing repeated content processing across different tools.

School management scenarios are also included in the product scope. During the Series B financing in 2025, MagicSchool launched MagicSchool Labs, with its first project focusing on school data analysis, allowing educators to query and analyze school data using natural language. The tasks covered by the product have further extended from the initial lesson preparation and material generation to student usage management and school data analysis.

These product changes are also related to Khan's personal experience. Before founding MagicSchool, he successively served as a teacher, assistant principal, principal, and worked in a school district office. Even after leaving the school management post, he maintained contact with classrooms through school district visits. Khan said that the high school he founded is still led by the original team members, and remains one of the better-performing public schools in Denver.

This experience was once a disadvantage for Khan when raising funds. Investors were more familiar with entrepreneurs who had backgrounds in technology companies or top universities, while Khan's professional experience came almost entirely from the education system. As MagicSchool enters more schools and school districts, these years of educational work experience have instead become an important foundation for the product to go deep into specific teaching scenarios.

As Khan said, "The model itself is no longer a differentiator. The key lies in how to combine the model with the specific problems people encounter in their actual work and specific professional knowledge, which is the key to unlocking the next wave of innovation and impact in the field of generative artificial intelligence."

Original link:

"No one wanted to invest in a former principal": How an educator used $63 million in venture capital to build an edtech AI startup

https://news.crunchbase.com/venture/educator-built-edtech-startup-ai-magicschool-kahn/

This article is from the WeChat official account "Duojing" (ID: DJEDUINNO), Author: Siluo, published by 36Kr with authorization.