Alpha School Internal Data Shows Most Students Miss Stated AI Learning Benchmarks
Only 42% of students at the premium private network met targets to double learning speed in math and reading as software expands into public districts.
Alpha School, a private school operator charging families up to $75,000 annually on the premise that artificial intelligence can compress core academics into a fraction of the standard school day, is falling short of its central academic targets, according to reporting by fortune.com (https://fortune.com/2026/11/14/alpha-school-ai-students-tuition-learning-goals/) based on internal company records first disclosed by The New York Times.
Alpha promises parents that its software allows students to master academic material twice as fast as traditional school peers in two hours of computer-based learning each morning. However, internal network records reveal that only 42% of students achieved the goal of doubling their learning pace across both mathematics and reading.
The performance deficit is concentrated in mathematics. Approximately half of kindergarten through eighth-grade students at Alpha's private campuses failed to hit the network's goal of doubling math progress, while roughly 20% missed their reading targets. Outcomes were lower for pupils who began behind grade level: of the estimated one-fifth of private school students who scored at or below the 50th percentile nationally on fall math assessments, only 37% met the double-speed growth milestone by winter.
Founded by MacKenzie Price, Alpha operates without conventional classroom teachers. Instead, campuses employ adult "guides" tasked with motivation and supervision while students complete coursework via TimeBack, an application Alpha markets as an adaptive AI tutor. The compressed academic schedule is pitched as freeing afternoons for non-academic seminars, including public speaking and life skills. The network has expanded to roughly 1,500 students across about 50 campuses and has attracted prominent tech and financial backers, including software billionaire Joe Liemandt, with public endorsements from LinkedIn cofounder Reid Hoffman and hedge fund manager Bill Ackman.
The operational reality behind the model has drawn skepticism beyond internal test data. Parents and guides previously told Wired that students routinely worked late into the night outside school hours to meet algorithmic targets, creating significant academic stress. Former personnel also told 404 Media that the network's automated software generated flawed curriculum sequences that staff warned caused educational harm, while allegedly incorporating scraped third-party coursework without licensing.
The internal metrics surface as Alpha accelerates a transition from high-fee private campuses into state-funded public education, where more than 2,600 public school students across Texas and South Carolina now use the software. Public systems operate under strict retention mandates and cannot easily filter or dismiss underperforming students, presenting an unproven operational test for Alpha's automated model. In the Houston Independent School District, which is running a two-school pilot, Superintendent Mike Miles told The New York Times that participating students spent longer on the software than scheduled, encroaching on non-digital school activities.
Sources
Written by
The Company Wire
Inside the companies building what’s next. Reporting on startups, technology, funding and the people shaping them.
