Chegg After ChatGPT: Subscribers, Revenue and Layoffs, Read From the Filings

Chegg reported $766.9 million of revenue for 2022, the year ChatGPT launched. For 2025 it reported $376.9 million, a fall of 51% in three years. The Chegg ChatGPT impact is now written into four years of filings: a shrinking subscriber base, $677.2 million of impairments in a single year, four rounds of layoffs, and a lost lawsuit against Google.

The short answer. ChatGPT did not destroy Chegg in a quarter. It removed new-customer growth first, in spring 2023, then Google's AI Overviews cut the search traffic that refilled the funnel. Chegg's subscribers fell 20% between 2022 and 2024, its revenue halved by 2025, and it is now a company guiding to $38 million to $40 million a quarter.

Reader: this is written for a founder or director of a subscription software business, because the useful output is a test you can run on your own revenue, not a verdict on Chegg's stock.

The Chegg ChatGPT timeline, dated from the filings

Every row below comes from a Chegg earnings release, an 8-K, a 10-K or a court report. The dates are filing or announcement dates, not the end of the period being reported. That distinction matters, because markets reacted on the filing date.

Sources: Chegg 8-K earnings releases and restructuring 8-Ks on SEC EDGAR, the Chegg 2025 Form 10-K, and Reuters reporting of the dismissal. Share price move is the closing change reported on 2 May 2023; intraday reports ranged from about 37% to 50%. Last verified 7 November 2026.

The single most quoted line in this story is in Chegg's own Q1 2023 earnings release. Chief executive Dan Rosensweig wrote that "since March we saw a significant spike in student interest in ChatGPT". He added: "We now believe it's having an impact on our new customer growth rate".

The same release said Chegg would guide "to the current quarter only, while these conditions exist". Withdrawing full-year guidance told investors more than the quote did. A company that cannot see two quarters ahead has lost confidence in its own funnel.

Most companies attribute weak results to macro conditions, seasonality or execution. Chegg named a specific competitor's product within five months of its launch. I think that was the right call, even though the market punished it. The alternative was a slow drip of misses with no explanation, which costs more credibility over time.

Chegg subscribers decline: what the reported numbers show

Chegg reported two subscriber figures. The headline number counts unique paying subscribers across a full year. A second number counts subscribers within the fourth quarter alone. Both tell the same story, and the gap between them tells a second one.

The full-year count went from 8.2 million in 2022 to 6.6 million in 2024. That is a 20% decline across two years, according to Chegg's Q4 releases. The Q4-only count fell harder, from 5.0 million to 3.6 million, a 28% decline across the same period.

The Q4 2022 release still reported growth of 5% for the year. ChatGPT had been public for about two months when that number was filed. The damage starts in the next reported quarter.

Chegg's 2025 annual report contains no subscriber count. In the same filing, Chegg changed its revenue split from Subscription Services to two new segments, Academic Services and Chegg Skilling. The 10-K explains the 2025 fall in Academic Services as "a decrease in subscription revenue of $223.8 million primarily related to reduced traffic which led to fewer subscribers".

So the company confirms fewer subscribers without saying how many. When a subscription business stops publishing its core volume metric, treat that as information. It rarely happens when the number is going up.

Chegg revenue after ChatGPT, quarter by quarter

Revenue is the one series Chegg has never stopped reporting, which makes it the cleanest line through the whole period. The annual figures come from the Q4 releases and the 10-K.

Total net revenue was $776.3 million in 2021, $766.9 million in 2022, $716.3 million in 2023, $617.6 million in 2024 and $376.9 million in 2025. The decline accelerated each year: down 7% in 2023, 14% in 2024, then 39% in 2025. That acceleration is the most important shape in the whole data set.

2026, computed from Chegg's own numbers

Chegg reported $63.3 million for Q1 2026, $51.8 million for Q2 and $46.5 million for Q3. It guided Q4 to between $38 million and $40 million. Adding those gives a full-year range of $199.6 million to $201.6 million.

Against 2022's $766.9 million, that is a decline of about 74%. The working is simple and you can check it: $766.9 million minus $200.6 million at the midpoint, divided by $766.9 million. Chegg has not published this number itself, so treat it as our calculation from its filings, not company guidance.

The remaining revenue now splits two ways. In Q1 2026 Academic Services earned $45.7 million and Chegg Skilling, the language and career learning unit, earned $17.6 million, up 9%. Skilling was about 28% of that quarter. Its revenue held roughly flat while the academic side kept shrinking, which is why the company is now pitched as a skilling business.

The write-downs: $677 million in two quarters

An impairment is an accounting charge that cuts the book value of an asset. It is booked when the cash the asset is expected to earn falls below what the balance sheet says it is worth. For goodwill and acquired intangibles, it is the moment management concedes that past purchases will not earn what was paid for them.

Chegg booked $481.5 million of impairment in Q2 2024 and $195.7 million in Q3 2024. The 10-K describes the full-year total as covering "impairments of goodwill, intangible assets, and other related property and equipment".

The $677.2 million total equals about 110% of Chegg's 2024 revenue. Net loss for 2024 was $837.1 million, against net income of $18.2 million in 2023. By the end of 2025 the 10-K reported an accumulated deficit of $992.9 million.

The write-downs landed in mid 2024, more than a year after the ChatGPT warning. Accounting moves slowly by design, because impairment tests compare forecasts against carrying values once a trigger appears. For a reader, the lesson is that the balance sheet confirms damage the subscriber line already showed. It is a lagging indicator, never an early one.

Chegg layoffs: four rounds in 17 months

Chegg cut staff four times between June 2024 and October 2025. Each round was announced through an 8-K or an earnings release with a stated headcount.

The June 2024 plan affected about 441 employees, or 23% of the workforce, with expected annual savings of $40 million to $50 million. The November 2024 plan affected 319 employees, about 21%. The May 2025 plan affected 248 employees, about 22%. The October 2025 announcement affected 388 employees, about 45% of those remaining.

Chegg's 2025 10-K states that it ended the year with 595 employees. It also says the two 2025 plans together removed about 640 roles, or 56% of the workforce at the time. Restructuring charges were $24.6 million in 2024 and $51.5 million in 2025.

The October announcement did three things at once. It ended the strategic review begun in February, concluding that Chegg would stay an independent public company. It brought Dan Rosensweig back as chief executive, replacing Nathan Schultz. And it attributed the decline to "the new realities of AI and reduced traffic from Google to content publishers."

Notice the order of blame in that sentence. In 2023 the named cause was ChatGPT. By 2025 it was Google. Both can be true, and the distinction matters for the checklist below. If you want the wider context on how AI is being cited in layoff announcements, we looked at that in the piece separating AI layoffs from the overhiring correction.

The Chegg Google lawsuit over AI Overviews

On 24 February 2025 Chegg sued Google and Alphabet in the US District Court for the District of Columbia. The complaint brought federal antitrust and unjust enrichment claims over AI Overviews, the AI-generated summaries Google places above search results. It was filed the same day as full-year 2024 results and the strategic review.

Chegg's 10-K set out the theory plainly. Its risk factors say AI Overviews keep "users on Google search results instead of leading them to Chegg's site". They add that Chegg expects "Google to continue its shift from being a search origination point to the destination".

Google moved to dismiss the amended complaint on 25 July 2025. Judge Amit Mehta granted dismissal on 30 September 2026, in a single opinion that also covered a similar case brought by Penske Media, Reuters reported. The judge wrote that the antitrust claims "fail to get out of the starting gate," and that "an expectation is not an agreement."

The ruling was at the pleading stage. It did not find that AI Overviews cause no harm to publishers. It found that Chegg had not stated an antitrust claim the court could hear. As of this post we have found no appeal on record, and Chegg's Q3 2026 release mentions the litigation only in general terms.

My read is blunt. Litigation is not a distribution strategy. If your acquisition channel belongs to a company that can change its product without your consent, the law is unlikely to give it back. The same dynamic is measured from the publisher side in our look at the metrics that replace clicks in zero-click search.

SaaS disrupted by AI: the five-point exposure checklist

Chegg is useful because the failure modes are visible in its filings. Each point below maps to something Chegg disclosed. Score your own business one point for each that applies. I would treat three or more as a board-level risk, not a product roadmap item.

Tests one and four: answers and archives

Chegg's product was a large library of explained answers plus on-demand expert help. That is precisely the shape of work a general chatbot does at zero marginal price to the student. A workflow product holds state, permissions, history and integrations. An answer product holds none of those, so switching costs are close to zero. The same argument for why workflow beats data as a defence is made in the post comparing data moats and workflow moats.

Test two: the borrowed channel

Chegg's homework pages ranked well on Google for years. That made acquisition cheap and also made it rented. When Google began answering the query itself, the funnel narrowed without any change in Chegg's product. If you rely on organic search or a marketplace listing for most new customers, you carry the same exposure at a smaller scale.

Test three: episodic demand

The gap between Chegg's full-year and Q4 subscriber counts is large every year. In 2024 roughly 6.6 million people paid at some point, but only 3.6 million in the final quarter. A business that must re-acquire much of its base each term is far more sensitive to a funnel shock than one with annual contracts. Seat-based software faces a different version of this, examined in our seat exposure scorecard for SaaS vendors.

Test five: the fixed cost base

Chegg entered 2023 with a cost structure built for 8 million subscribers and growing. Cuts came in four rounds, each after the revenue had already moved. In my view the October 2025 cut of 45% was the one that should have come a year earlier. Smaller, repeated cuts carry a morale cost that one decisive cut avoids.

Where this argument is weakest

A clean story about ChatGPT killing Chegg is tempting. The filings support most of it, but not all of it.

The decline started before ChatGPT

Revenue had already slipped from $776.3 million in 2021 to $766.9 million in 2022, and almost all of 2022 predates ChatGPT's release. Pandemic-era demand for online study help was unwinding. Some of the 2023 fall would likely have happened anyway. The filings do not let anyone separate the two effects with precision.

The company's own attribution is self-interested

Blaming ChatGPT in 2023 and Google in 2025 positioned Chegg as a victim of external shocks rather than of its own product choices. Both claims are plausible. Neither has been tested by an independent party. The one external test, the lawsuit, was dismissed without reaching the question of harm.

The best metric went dark

Without subscriber counts after 2024, any claim about 2025 and 2026 subscribers is inference from revenue. Revenue also moves with price, mix and the Skilling unit, so it is an imperfect proxy. We have chosen not to estimate a 2025 subscriber number because no filing supports one.

One company is not a category

Chegg is the clearest public example, not a representative sample. Many software businesses share one or two of the five traits and are growing. The checklist identifies exposure. It does not predict an outcome, and it should not be read as a forecast for any company other than the one whose filings produced it.

Frequently asked questions

How to run the test on your own business this week

Take your last 12 months of new customers and tag each one by acquisition channel. Then mark which channels you control outright and which belong to a search engine, marketplace or platform. If more than half of new revenue came through someone else's surface, that is your test-two score, and it is the cheapest of the five to measure.

Next, ask one honest question in your next leadership meeting. If a general-purpose model could produce 80% of what our customer pays us for, what would make them stay? Write the answer down. If it names your content or your speed, rather than your workflow or your data about the customer, you have found the Chegg exposure early, while there is still time to act on it. For a wider frame on how markets are pricing this risk across software, see our breakdown of the SaaSpocalypse sell-off, and for the pricing side, the analysis of seat compression in SaaS pricing.

Related reading