AI & EdtechSeptember 19, 2026

The Great EdTech Compression

Sahra-Josephine Hjorth lying in an open coffin wearing sunglasses and a blue gown, holding a bouquet of white lilies

Part five of how 90% of edtech disappears. Learning is not disappearing. The companies, valuations, service layers, and independent categories built around delivering it are being merged, repriced, absorbed, and bypassed.

Bankruptcies are the public funerals of EdTech. They are dramatic, recognizable, and relatively easy to count: a company files for Chapter 11, the debt gets restructured, the assets are divided, and everybody agrees that something died. But they are not where most of the industry is disappearing.

An industry can lose enormous amounts of mass without producing a corpse. Two companies can merge. A functioning business can survive while most of its valuation does not. A service layer can remain necessary while losing the economics that once made it enormously valuable. An educational product can be absorbed into consulting, HR, or a general-purpose technology platform. A company can even remain standing while the learner finds a shorter route around it.

For the past twenty years, we have been adding industrial mass around learning. Assessment could be a company. Authoring could be a company. Credentials could be a company. Student success could be a company. Helping a university put a degree online could become an entire industry of its own. Around each educational function, we built products, service businesses, implementation layers, sales teams, integrations, headcount, and increasingly enormous amounts of capital.

A lot of that mass is now coming out. The educational capabilities frequently survive. What gets smaller is the collection of independent companies, valuations, service layers, and commercial relationships built around delivering them. That is the Great EdTech Compression.

Sometimes it looks exactly like a funeral

2U is the obvious place to start because the numbers are difficult to miss. In 2021, it agreed to acquire substantially all of edX’s assets, including its brand, website, and marketplace, for $800 million in cash. Three years later, 2U entered a restructuring process that would cut its debt by more than half to approximately $459 million and inject roughly $110 million of new capital. The company was equally explicit that its educational programs and services would continue without interruption. [1]

Anthology gives us an almost physical illustration of the same idea. The company had assembled Blackboard together with student information, ERP, CRM, engagement, and student-success products. After Chapter 11, Blackboard’s core Teaching & Learning business emerged in March 2026 as a standalone, debt-free company. Anthology’s SIS and ERP business moved to Ellucian, while Encoura acquired the Lifecycle Engagement and Student Success portfolio. [2]

Universities still need learning platforms, student information systems, and tools for retention and student success. Bankruptcy did not make those functions useless. It reduced or dismantled the corporate and financial structures that had accumulated around them.

Nobody has to die. Two can simply become one.

Bankruptcy makes good coffin photography, but it is not actually the most interesting version. Coursera and Udemy completed their combination in May 2026. The combined company says it now encompasses 290 million learners, 18,000 enterprise customers, and more than $1.5 billion in 2025 annual revenue. It also expects $115 million in annual run-rate cost synergies within 24 months, with a significant majority expected during the first year. [3]

I love the word synergy because it makes subtraction sound like addition. The new company is enormous, and online learning has not disappeared. But apparently, delivering all of that learning does not require the full cost structures Coursera and Udemy previously maintained separately.

PowerSchool shows the same process happening across product categories rather than between two direct competitors. According to its SEC filings, it completed 19 strategic acquisitions between 2015 and June 2024 while building a platform covering student information, enrollment, learning management, assessment, special education, finance, HR, analytics, communications, and talent management. Bain then acquired PowerSchool in October 2024 for approximately $5.6 billion in enterprise value. [4]

That is very obviously not a company dying. It is several things that could once be bought independently increasingly sitting behind one commercial relationship. The category is not emptying. It is being folded into fewer balance sheets.

Sometimes the company survives and the valuation gets buried

Unacademy is useful because nobody can plausibly describe what happened as a simple corporate death. At its peak in 2021, it was valued at $3.44 billion. On September 1, 2026, upGrad acquired it in an all-stock transaction valuing the company at approximately $206 million, around 94% below the peak. Yet Unacademy reportedly still had about $95 million in cash, most of its businesses were profitable or close to profitability, the brand was retained, its CEO remained in charge, and there were no planned layoffs tied to the transaction. [5]

The company did not disappear. Most of the valuation did. This is why counting bankruptcies badly underestimates what is happening. A company can keep its customers, employees, brand, and login page while the financial story previously told about it is buried almost completely. Compression can happen underneath a business that remains visibly alive.

The strange disappearance of the OPM fortune

Some of the most dramatic compression is not happening to software at all. For years, universities built an enormous service industry around putting programs online. Online Program Management companies handled combinations of marketing, recruitment, enrollment, instructional design, technology, and student support. These businesses attracted prices based on the belief that this intermediary layer would remain extraordinarily valuable.

Wiley paid $220 million for Deltak in 2012 and later acquired Learning House and XYZ Media as it expanded its education-services operations. When Wiley sold University Services on January 1, 2024, the fair-value selling price was $122.6 million, and Wiley recorded a $107 million pretax loss on the sale. [6]

Pearson offers an even cleaner before-and-after story. In 2012, it paid $650 million in cash for EmbanetCompass, which provided online-learning services to more than 100 university programs. When Pearson sold Pearson Online Learning Services in 2023, there was no equivalent cash price. The deferred consideration consisted of 27.5% of positive adjusted EBITDA each year for six years, plus 27.5% of the proceeds from a future monetization event. [6]

Universities did not stop needing help delivering online education. They stopped supporting the same economics around the companies standing in the middle. The educational capability remained. The OPM fortune did not.

EdTech can disappear without anybody admitting that it left EdTech

This is where Accenture and Udacity belong, and yes, the acquisition is real. Accenture announced in March 2024 that it had agreed to acquire Udacity and completed the transaction on May 20. The purchase price was not disclosed. More than 230 Udacity employees joined Accenture LearnVantage, taking Udacity’s content, services, and learning technology into a much broader consulting and workforce-transformation business.

At the same time Accenture launched LearnVantage, it announced a $1 billion investment over three years in the business. That $1 billion was not the Udacity acquisition price. It was Accenture’s commitment to building the much larger learning capability into which Udacity was absorbed. [7]

That makes the example better, not worse. Udacity did not become worthless. Accenture decided its capabilities were valuable enough to acquire, but their natural home was no longer necessarily an independent EdTech company. They became part of a consulting giant selling talent transformation, AI skills, and workforce change. Sometimes a category wins the argument for its own importance and loses the argument for its independence.

Even Photomath now belongs to Google. The European Commission cleared Google’s acquisition of sole control of the company in 2023. The specialist educational tool and the general-purpose interface now sit under the same owner. [8] These are not distressed-company stories. That is precisely why they belong here. A category can compress through success.

And then there is the company nobody needs to acquire

Chegg is different. Its total revenue fell from $617.6 million in 2024 to $376.9 million in 2025, a decline of 39%. Academic Services revenue fell from $543.6 million to $308.3 million, down 43%. [9]

The important part is that we do not have to invent a theory for why. In its 2025 annual report, Chegg says students increasingly see general-purpose generative-AI products such as ChatGPT as alternatives to vertically specialized educational services such as Chegg. It also says Google’s AI Overviews increasingly answer educational questions directly in search results, keeping users on Google rather than sending them onward to Chegg. [9]

Nobody has to acquire Chegg, split it into pieces, or put it through Chapter 11. Students still get stuck on homework. They still need explanations and help the night before an exam. The demand survives. The layer that used to monetize being the place they went for that help becomes easier to bypass.

What is actually being compressed

Looking only at the company count misses it. Looking only at valuations misses it too. The thing losing mass is the industrial structure around educational capability.

Sometimes that mass is equity and debt. Sometimes it is the existence of two separate companies. Sometimes it is an intermediary layer whose economics no longer resemble those that once justified a $650 million acquisition. Sometimes several product categories become one platform. Sometimes an EdTech company becomes part of consulting or a general-purpose technology business. And sometimes the company stays exactly where it is while the learner discovers a shorter route around it.

This does not mean value disappears evenly. Duolingo ended 2025 with 52.7 million daily active users, 12.2 million paid subscribers, $1.038 billion in revenue, and $1.158 billion in bookings. [10] Compression does not prevent enormous winners. It may produce a more concentrated market in which the survivors become much denser businesses.

If you wait for funerals, you will miss most of the compression. The company can remain. The brand can remain. The login page can remain. The learning can remain. For twenty years, EdTech behaved as though every important educational function could support its own company, product category, sales organization, service layer, and valuation. That assumption is what is breaking. Not learning. Not education. The amount of industry required to deliver it.