The EdTech Bubble: Are AI Learning Platforms Overvalued in the Current Market?

There’s a certain energy in the room when you walk around any major EdTech conference right now; it’s partly real excitement and partly something more difficult to describe. Booths are modern. The demonstrations are well-done. “AI-powered” appears in the tagline of every third startup, typically above an upward-trending graph. It’s the kind of environment that exudes confidence until you start posing the precise queries that no one really wants to address, such as whether or not these platforms are truly improving student learning.

The industry has been debating this issue for years, and it is becoming more and more prevalent. In retrospect, the wave of startups and valuations that resulted from the EdTech boom of 2020–2022, which was fueled by pandemic school closures and an influx of venture capital that appeared to follow every Zoom call with a term sheet, bore more than a passing resemblance to something the financial world has seen before. In October of last year, The Guardian reported that the similarities to the dotcom bubble of the late 1990s were “genuinely close to madness.” In 1999 alone, the Nasdaq increased by 86%. All it took for companies to see a spike in share prices was to announce an online strategy. It turns out that the idea that something similar might be occurring right now with AI-labeled educational technology is not particularly contentious among those who work in the field.

The most obvious warning comes from India. Once valued at $22 billion and hailed as an example of what EdTech could achieve, Byju’s reportedly suffered losses of more than ₹8,000 crore. PhysicsWallah and Unacademy followed with their own eye-popping red ink. These were not fringe players; rather, they were the most well-known success stories in the sector, supported by some of the world’s most astute investors. The money came in incredibly quickly. Exam scores, graduation rates, and employment are the only metrics that truly matter, and the learning outcomes were, at best, inconsistent. At worst, the people who write the checks themselves silently disregard it.

Speaking with educators who use these tools on a daily basis gives the impression that the industry has been creating something a little different from what it promotes. In an online forum that has become something of a gathering place for people disillusioned with the sector’s promises, one university LMS administrator described nearly ten years of working in higher education technology. He wrote that technology in education was never really meant to improve student outcomes. It was constantly trying to defend its own existence and costly maintenance. It’s a hard read. However, the number of educators who support it in public and in professional settings has been rising.

Although the AI layer has significantly altered the valuations, it hasn’t clearly altered this dynamic in recent years. Due to generative AI integration, which is essentially a chatbot added to pre-existing content, platforms that could have raised a Series A at $50 million two years ago are now able to command multiples of that amount. Some of these products might eventually prove to be truly helpful. Additionally, it’s possible that investors have once again mistaken a strong pitch for a long-term company. Last October, the Bank of England warned about the wider risk associated with AI valuation. It’s still unclear if those who are specifically investing in EdTech are receiving that warning.

It is evident that the educational institutions and school districts that are actually using these tools are coming to their own conclusions, and those conclusions aren’t always positive. Teachers in many nations are reporting a quiet return to analog methods, such as pencil, paper, and physical textbooks. This is not because they are afraid of technology, but rather because AI tools are causing more issues than they are resolving. Students are turning in unexplainable AI-generated work. classrooms where students’ focus is more dispersed than concentrated. Districts that heavily invested in platforms during the pandemic are now unable to identify a single quantifiable improvement in the intended results of the funding.

Any sincere EdTech investor will admit off-the-record the deeper structural issue: the customers who purchase EdTech, such as district technology officers, procurement departments, and school administrators, are often not the same people who use it. Additionally, it is rare for the satisfaction of its users—teachers and students—to determine whether the contract is renewed. In enterprise software, there has always been a mismatch between the buyer and the user. It carries a weight that a subscription renewal rate doesn’t adequately convey in education, where the stakes are someone’s capacity for reading or math.

This does not imply that all AI learning platforms are worthless or that the industry as a whole is doomed to fail. Tools that actually save teachers time when grading assignments, provide a patient and accessible resource for struggling students at two in the morning, or enable a college course to be taken by someone who would not otherwise be able to afford it are truly valuable. The issue is that “is valued correctly by the current market” and “has real value” are two quite different statements. Observing the market at the moment, it seems as though the market is pricing the dream rather than the delivery. That gap has a tendency to close—not always gently, but eventually.