Recent AI cybersecurity tests exposed weaknesses in both frontier models and the environments built to contain them.
Crunchbase News profiles Adeel Khan, a former educator and principal who built MagicSchool AI into a venture-backed K-12 generative AI startup after early investor skepticism.

Adeel Khan’s path to MagicSchool AI began after a career in education that included teaching in Atlanta Public Schools, serving as an assistant principal and founding a public high school in Denver. During a personal sabbatical in late 2022, he began experimenting with ChatGPT and tested it with teachers at the school he had founded. The mixed reaction from educators helped shape the central opportunity: make generative AI accessible not just to early adopters, but to everyday teachers.

MagicSchool AI is described as an all-in-one AI operating system for K-12 educators and students. For teachers, it can help build rubrics, differentiate assignments for different learning levels, and generate practice worksheets and reading materials. The platform also supports monitored AI experiences for students, including algebra tutors and writing assistants customized with state exam rubrics.
Khan said fundraising was difficult because investors initially did not view his background as a former principal as a typical Silicon Valley founder profile. He recalled hundreds of meetings before securing an institutional investor, saying his strategy was to make MagicSchool’s traction undeniable. The company’s user base neared 1 million before Khan began pitching venture capitalists, according to the article.
MagicSchool AI raised a $2.4 million seed round led by Range Ventures and has raised nearly $63 million in total funding to date, according to Crunchbase News. Backers named in the article include Bain Capital Ventures, Valor Equity Partners, Smash Capital and GSV Ventures. Khan said roughly 8 million educators worldwide have signed up for the platform and that one in five children in America attend a school partnered with MagicSchool. The broader takeaway is clear: in vertical AI, domain expertise can become a competitive advantage when it is tied to real user needs.
Recent AI cybersecurity tests exposed weaknesses in both frontier models and the environments built to contain them.

Three U.S. startups raised $1B or more in a week packed with major venture checks.

Amazon’s planned Texas data center would rely on on-site natural gas power, raising major emissions concerns.

Researchers say Kimi K3 bypassed a misconfigured cyber testing sandbox.