With annual tuition as high as $75,000 and only two hours of study per day, can Alpha School's AI school model work?

Can students achieve better results than traditional schools by studying only two hours a day?

In the United States, a private school called Alpha School is using AI to try to change the familiar way of education.

There are no teachers in the traditional sense standing at the front of the classroom lecturing. Students spend each morning completing core subjects such as math and reading on computers, and in the afternoon they take part in sports, public speaking, coding, and various hands-on activities. The school claims that this model allows students to learn at twice the speed of traditional schools.

Even more notable is its price. According to Scientific American, annual tuition at most Alpha School campuses is about $40,000 to $75,000. Even so, this model, which emphasizes AI and personalized learning, still attracts many families willing to try a new approach to education.

However, as the school gradually became a star example of AI education, doubts followed. On October 23, The New York Times published an investigative report revealing that some of Alpha School's internal learning data differed from its promoted claim of "doubled learning speed."

At the same time, former students' descriptions of how learning is managed have also prompted outsiders to re-examine the school. When AI begins to take on teaching work that originally belonged to teachers, will learning really become more efficient? And can a school with such high fees extend its model to the broader education market?

How can a school without traditional classrooms get students to study only two hours a day?

The story of Alpha School began in 2014.

The school was founded in Austin, Texas, and its co-founder MacKenzie Price believed that traditional classrooms have an obvious problem: dozens of students sit in the same classroom and learn at the same pace, but each person's comprehension ability and knowledge base are not the same. Some students have already mastered the course content but still have to wait for their classmates; some students have not yet understood the knowledge point, but the class has already moved on to the next part.

Alpha School hopes to change this situation through AI.

Here, students no longer need to follow a unified classroom pace; instead, they complete core subject learning through an adaptive learning system called TimeBack. The system adjusts learning content based on students' answering performance. If a student repeatedly makes mistakes on a question, the course will return to the relevant foundational knowledge; if the student has already mastered a knowledge point, they can continue learning forward. According to the school's vision, this approach can reduce repetitive practice and let students spend time on what they truly need to learn.

Therefore, Alpha School concentrates daily core subject learning into about two hours in the morning. However, this does not mean students only need to stay at school for two hours a day. In the afternoon, they still need to participate in sports, public speaking, coding, and other hands-on activities. The school hopes that this arrangement will allow students to complete subject learning while also having more time to develop communication, cooperation, and problem-solving skills.

As the teaching method changes, the role of teachers has also been redefined.

Alpha School calls the adults on campus responsible for students' daily activities "Guides." They no longer lecture on subjects such as math and English like traditional teachers, but instead focus more on students' motivation, behavioral habits, and growth activities. When students encounter complex academic problems, a remote professional team also provides help.

In other words, Alpha School does not completely eliminate teachers, but rather distributes the different tasks traditionally undertaken by teachers among the AI learning system, remote academic staff, and on-site guides.

For some students, this change has indeed brought a different learning experience. In a report updated on October 22, CBS interviewed a student at Alpha School's Chicago campus. He believed that studying independently on a computer made it easier to concentrate than sitting in a classroom with more people. His father also said that the child was more willing than before to talk proactively about school life.

But this model is not cheap.

CBS reported that annual tuition at the Chicago campus for fall 2026 is $55,000. At other Alpha School campuses, fees can reach up to $75,000 per year.

This also means that Alpha School is currently more like an experiment taking place in the high-end private education market. What parents buy is not just a set of AI learning software, but an entirely different school life: efficient learning in the morning and developing interests and abilities in the afternoon.

The question is whether this seemingly attractive arrangement can really deliver the learning results the school promises.

Have scores really doubled? Internal data spark controversy

"Learning at twice the speed of traditional schools" has always been Alpha School's most closely watched promotional claim. For parents, this means their children may complete more learning tasks in less time; for the school, this is also an important basis for proving the value of AI teaching.

But The New York Times investigation on October 23 showed outsiders another set of data. The report disclosed that in the internal statistics involved, about 42% of students simultaneously met the school's promoted "double growth" standard in both math and reading. Another set of data showed that among some students whose math scores were at or below the 50th percentile nationally in the fall of the previous year, 37% achieved the goal of double growth in math by winter.

These figures contrast with the image of high efficiency that the school emphasizes to the outside world. Of course, 42% does not mean that only 42% of students are performing at grade level, nor does it show that the remaining students have made no progress. What it reflects is that, in the relevant statistics, not all students are able to reach the high-growth targets set by the school.

Because the full sample size and raw data have not yet been sufficiently disclosed, outsiders currently cannot judge the overall quality of Alpha School's teaching effectiveness based on these figures alone. But one question has already emerged: when "learning at twice the speed" becomes the school's most important selling point, does it have enough data to prove that this promise can be broadly fulfilled?

In fact, Alpha School has consistently emphasized students' outstanding performance on standardized tests. In February of this year, the school also published an analysis of test scores based on NWEA MAP data, and used AI tools to explain the results. However, better student performance does not necessarily mean that the AI teaching model is the only reason.

For example, Alpha School families generally have strong financial capacity, and students may also have different learning foundations and family support conditions.

The school says it does not screen students solely based on past test scores, but it still pays attention to whether applicants are suited to this way of learning. Therefore, to determine how much help AI actually provides, it is still necessary to compare these students with students of similar circumstances who attend traditional schools. This is also an important question that education researchers have raised about Alpha School.

In addition, the controversy is not limited to test scores. In September of this year, education researcher Benjamin Riley published an investigation based on interviews with former students. Some interviewees said that the school had used software or cameras to track learning behavior, and used strict task management and incentive mechanisms to urge students to complete their studies.

The technology media outlet Futurism subsequently reported these allegations. It should be noted that these descriptions come from the personal experiences of some former students, and it cannot currently be concluded from them that the same situation exists at all Alpha School campuses.

But they still raise a question worth paying attention to.

When a large amount of students' daily learning activities take place on computers, the system can not only record how many questions they answered correctly, but may also grasp study duration, answering speed, and task completion. These data can help the school identify learning problems, but they also draw more attention to student privacy, learning pressure, and daily management methods. Especially for younger students, learning is not only about completing tasks and mastering knowledge; it also requires communication, encouragement, and emotional support.

AI can adjust the curriculum based on answer performance, but why students lose interest and why they feel confused sometimes cannot be judged by data alone. Therefore, what Alpha School truly needs to prove is not only whether students can complete the curriculum within two hours, but also whether this learning method can maintain stable results over the long term.

From a high-priced private school to a larger market, can Alpha School replicate its success?

Despite the ongoing controversy, Alpha School has not slowed the pace of expansion.

According to an investigation jointly published by The Texas Tribune and ProPublica in September of this year, Alpha School has established private campuses in multiple U.S. states and is accelerating expansion in Texas. The school is clearly not satisfied with serving only a small number of high-income families; rather, it hopes to promote its teaching method to more regions.

From a business perspective, this model has a clear expansion logic. Traditional schools require teachers to organize courses by class, while Alpha School hopes to provide core subject learning through a unified software system, with instructors at each campus responsible for student management and practical activities.

If the system can operate stably, the curriculum and technology have the opportunity to be reused across different schools. Compared with assembling a complete subject teaching team anew every time a school is opened, this method theoretically has certain expansion advantages. But once it truly enters the broader education market, the problems also become more complex.

A joint investigation in September of this year disclosed that an Arizona charter school associated with the Alpha model had estimated during the application stage that the proportion of students reaching proficiency in English and math in the first year would be 65% and 60%, respectively. However, preliminary state assessment results cited in the report showed actual proportions of 28% and 10%.

It should be noted that this set of data only concerns the phased performance of the relevant charter school and cannot represent the teaching level of all Alpha School private campuses. But it at least shows that after the same teaching model enters different schools, actual results may differ significantly. There is already a considerable difference between high-fee private schools and ordinary public education environments. The former can attract families who agree with its educational philosophy and have strong ability to pay; the latter must face students with different learning foundations, family backgrounds, and support needs.

For some students, learning independently on a computer may be more efficient; but for students who need more face-to-face explanation, supervision, and help, this method may not be equally suitable.

In addition, school expansion is far more than copying a piece of software. Even if core courses are provided by an AI system, new campuses still need space, management personnel, safety guarantees, and student services. Software development, technical maintenance, and remote academic support also require continuous investment.

At present, Alpha School has not yet disclosed sufficiently complete financial data, making it difficult for outsiders to judge whether this model can truly reduce school operating costs, nor can they conclude solely from the growth in the number of campuses that it has achieved profitability at scale.

The policy environment also affects expansion. According to the investigative report, the Texas State Board of Education previously rejected Alpha School's charter school application, but most eligible Alpha private campuses subsequently entered the list of selectable schools under the state education subsidy program. This provides new possibilities for some families to choose Alpha School, but it does not mean that the school's teaching effectiveness has been officially recognized.

As AI education moves from an option for a small number of families to the broader school market, companies will no longer face only the question of whether parents are willing to pay. Whether the curriculum meets local requirements, whether different students can receive sufficient support, how learning data is protected, and whether public funds are used reasonably will all become practical issues affecting their development.

For the entire AI education industry, Alpha School's attempt still has reference value. In the past, AI was mostly used for answering questions, grading, and personalized practice, mainly undertaking auxiliary teaching work. Alpha School is trying to go further: making technology the main vehicle for core subject learning and rearranging the relationship among teachers, students, and schools.

This exploration does indeed offer new ideas for traditional education. But from assisting teachers to undertaking the school's core teaching tasks, what AI needs to prove is also increasing. It must not only enable students to learn faster, but also enable students with different foundations to truly learn; it must not only achieve results in a small number of campuses, but also be able to adapt to a broader educational environment.

Alpha School has already demonstrated the possibility of AI changing the way schools operate, but whether the story of "two hours of learning a day" can become a sustainable education model still requires time and data to answer.

This article comes from the WeChat public account "Duojing" (ID: DJEDUINNO), author: Xu Miaomiao, published by 36Kr with authorization.