Edtech After the Boom: Byju's, Chegg, Course Hero, Unacademy

Why edtech shows up so often in the graveyard data

Of the 10,525 companies assessed in WillItSurvive's July 2026 assessment, the blind backtest arm includes 116 dead companies, and edtech contributes a striking cluster to that sample: Byju's and Udacity sit in the dead arm, while Chegg, Course Hero, and Unacademy were originally seeded as healthy comparison peers and were flagged negative anyway. That pattern, three edtech "survivors" flagged as unsafe by blind judges, is the core of this story.

The boom itself is well documented outside our assessment. Byju's was widely reported as India's most valuable startup, with a peak valuation of roughly $22 billion in 2022. Chegg's pandemic-era subscription growth made it a stock-market favorite before its CEO publicly acknowledged, on the May 2023 earnings call, that ChatGPT was hurting new-customer growth, a disclosure that sent the stock down roughly 48% in a single session, per contemporaneous coverage. Course Hero's parent Learneo announced layoffs of about 15% of staff in mid-2023, per press reports at the time. Byju's was admitted into insolvency proceedings by India's National Company Law Tribunal in July 2024 over an unpaid sponsorship debt to the Indian cricket board, a filing covered by Reuters and the Economic Times. These are external, dated, verifiable events; our assessment adds the question of whether the trouble was visible in the fundamentals before the headlines.

How the backtest actually works

The blind backtest is not a hindsight exercise. Each dead company was judged only on facts knowable 18 months before its terminal event, with company identity masked and context time-boxed to the cutoff date. Matched surviving peers were judged at the same cutoff, same masking, with two independent replicates on the survivor arm. An identifiability probe then tried to name each masked company so recognized and unrecognized records could be analyzed separately.

On the survivor side, only 17% of 58 matched surviving peers were flagged negative in the first replicate, and replicate agreement on the survivors was 49 of 53, or 92% (July 2026 backtest). We never say we predicted deaths; the supported claim is that, given only pre-collapse facts, the assessment separated failing companies from matched survivors and never once said safe-to-join on a company that was dead within 18 months.

The edtech names, one by one

Course Hero. Same mirror structure. Course Hero appears in the dead arm with a September 2024 cutoff, paired against Chegg, and in the survivor false-negative list with a January 2025 cutoff, where both replicates said not called safe. When Chegg and Course Hero were each seeded as the other's healthy peer, blind judges rejected both as safe picks in both replicates (July 2026 backtest). The Learneo layoffs of 2023, publicly reported, were knowable at both cutoffs.

Unacademy. Unacademy was seeded as Byju's healthy twin with a January 2023 cutoff. Both blind replicates flagged Unacademy not called safe, and it landed on the survivor false-negative list alongside the other mis-seeded twins (July 2026 backtest). Unacademy's repeated layoff rounds across 2022 through 2024 were widely covered in Indian tech press and were partly visible at the cutoff.

Udacity. Udacity rounds out the edtech graveyard presence: it sits in the dead arm with a September 2022 cutoff, paired against Coursera, with edX, Pluralsight, and Skillsoft as alternate candidates.

What the mis-seeding episode says about the category

Of the 10 mis-seeded "survivors" flagged negative in the July 2026 blind backtest, three were edtech companies: Chegg, Course Hero, and Unacademy. No other single category contributed that many. The seeding process picked what looked like healthy comparison peers; in consumer edtech, the pool of genuinely healthy peers was thin enough that the seeds themselves turned out to be troubled. Duolingo, which appears in the assessment as the surviving peer for language-learning app Busuu (November 2021 cutoff), is a reminder that the category is not uniformly stressed; the pressure concentrated in homework-help and test-prep models exposed to generative AI and to cash-burn-funded growth.

One caveat we publish ourselves: on the recognized dead subset (n=99), the negative rate was 90%, and on the unrecognized "clean" dead subset (n=17) it was 88% with zero safe calls, so recognition-independence is claimed for the dead arm; the clean survivor subset (n=5) is too small to make the same claim there (July 2026 backtest).

Questions this data answers

Did the data show trouble at Byju's before the insolvency filing? Byju's was judged in the blind backtest at a January 2023 cutoff, eighteen months before the July 2024 NCLT insolvency admission, on masked, time-boxed facts. It sits in the dead arm, where 90% of 116 companies drew negative verdicts and none drew a safe call (July 2026 backtest). The verdict rested on what was publicly knowable in January 2023, including delayed audited financials.

Why do Chegg and Course Hero appear as both dead and survivor in the data? Each was seeded as the other's healthy comparison peer, Chegg at a September 2024 cutoff and Course Hero at January 2025. Blind judges returned not called safe on both, in both replicates, and both landed on the mis-seeded-twin list of 10 companies later confirmed as genuinely failed or gutted (July 2026 backtest).

Was Unacademy assessed as healthy? No. Seeded as Byju's surviving twin at a January 2023 cutoff, Unacademy drew not called safe from both blind replicates and is listed among the survivor false-negatives that turned out to be mis-selected peers (July 2026 backtest).

Is all of edtech in trouble in this assessment? No. The stress concentrated in homework-help and cash-burn models; Duolingo, for example, appears as the surviving peer for Busuu in the blind-pair set. Coursera appears as the surviving peer for Udacity. The assessment treats the category company by company.

How reliable are these blind verdicts? Replicate agreement on the survivor arm was 49 of 53, or 92%, and the dead arm produced zero safe-to-join calls across all 116 companies (July 2026 backtest).

What to watch, dated

As of July 2026: the resolution of Byju's insolvency proceedings before the NCLT; Chegg's subscriber counts in its next quarterly filings; Learneo's disclosures about Course Hero traffic; and any new funding or audited results from Unacademy. Check these yourself; we never declare an outcome.

Assessments reflect evidence available July 2026. Companies change; treat watch-items as things to verify, not conclusions.