Duolingo Q2 2026: users up 23%, bookings up 8% — that gap is the case

Duolingo’s second quarter contains three growth rates. They are all correct, they all describe the same company, and they are very far apart.

Measure	Q2 2026 growth
Daily active users	+23% — accelerating
Revenue	+18%
Bookings	+8% reported, +6% constant currency

Most coverage led with the first number and the earnings beat. Here is why the third number is the one that matters, and why it is not a disaster either.

Did Duolingo beat its own guidance in Q2 2026?

Yes. Revenue of $298.5 million came in above the $295.5 million Duolingo guided to in its May 4, 2026 8-K (the Q1 shareholder letter, Exhibit 99.2), bookings of $289.1 million topped the $283.5 million guide, and adjusted EBITDA of $77.3 million was above the $71.0 million guide. Duolingo guides to single figures rather than ranges, and that letter guided Q2 bookings growth of 5.8% (4.1% in constant currency) against the 8% (6% constant currency) delivered. Year over year, revenue rose 18% and bookings 8%, while GAAP diluted EPS was $0.66 against $0.91 a year earlier.

Bookings came in below revenue

Revenue was $298.5 million. Bookings were $289.1 million. Bookings were lower than revenue.

In a subscription business those two lines mean different things and the ordering between them is a forecast. Bookings are the cash customers commit when they buy a subscription (plus advertising, Duolingo English Test and in-app purchases). Revenue recognises subscription cash month by month over the life of the subscription — most of which are twelve months, the letter says. So today’s revenue is largely the echo of bookings made in prior quarters, and today’s bookings are the raw material of future revenue.

When bookings run below revenue, the pool of already-paid-for subscription value — deferred revenue — is drawn down rather than built up. The balance sheet shows exactly that: deferred revenue fell from $513.3 million at March 31 to $505.1 million at June 30, a decline of about $8.2 million (our arithmetic). A year earlier the ordering was the other way round: Q2 2025 bookings of $268.0 million exceeded revenue of $252.3 million.

Two pieces of context keep this from being read as a structural break. First, it is partly seasonal: in Q1 bookings of $308.5 million ran above revenue of $292.0 million, and the letter says Q2 “bookings seasonality was within our expectations”. Second, management attributes the slower year-over-year bookings growth to a tough comparison — the initial rollout of Energy, a price increase and advertising outperformance in Q2 2025.

What does not go away is the growth-rate gap. An 18% revenue line and an 8% bookings line cannot coexist indefinitely; revenue growth has to converge toward bookings growth. Management’s own guidance shows that convergence under way. For the third quarter Duolingo guided bookings of $307 million, up 8.9%, and revenue of $302 million, up 11.1% — bookings back above revenue, and the gap between the two growth rates narrowing from about ten points to about two (our arithmetic). Carry the full-year guide through and the fourth quarter implies bookings of roughly $380 million, up about 13%, against revenue of roughly $315 million, up about 11% (our arithmetic, from the FY guide of $1,285 million bookings and $1,207 million revenue less Q1, Q2 and the Q3 guide, against Q4 2025 bookings of $336.8 million and revenue of $282.9 million). On the company’s own numbers, bookings growth overtakes revenue growth by year-end — provided the guided bookings acceleration actually arrives.

An investor reading only the revenue line is reading a number whose deceleration is already visible in a different line on the same page.

This was a deliberate trade, not a stumble

The important context, and the reason the bookings gap is a trade rather than a stumble: Duolingo chose this.

The shareholder letter describes a “2026 strategy of prioritizing user growth and teaching better”, and says the company has concentrated monetisation work on “ways to increase revenue without adding friction to free users”. CEO Luis von Ahn framed the quarter as confirmation: “The results reinforce our strategy to improve the product and prioritize user growth.” The stated ambition is to teach a billion people.

Seen that way the three growth rates are not in conflict — they are the input and the output of one decision:

• The intended output: DAUs of 58.7 million, up 23%, accelerating two percentage points from Q1’s 21%, and “ahead of our expectations”, the letter says. Management attributes the acceleration to product changes, marketing and a one-time Streak Revival event in June, and says that because two of the three are permanent it expects DAU growth to stay above 20% for the rest of the year. In a business where the free funnel feeds every future paying cohort, reaccelerating user growth after a slowdown — DAU growth was 49% in Q1 2025 and 21% in Q1 2026 — is a genuine achievement.
• The intended cost: bookings growth of 6% in constant currency. A more generous free tier converts fewer of those users today.

The conversion question is visible in the filed numbers. Paid subscribers grew 17% to 12.7 million, against DAU growth of 23% and subscription bookings growth of 10% to $250.3 million. Users are growing faster than payers, and payers faster than the cash they commit.

So the question this report poses is not “did Duolingo execute?” It is a harder one: is a user acquired on a deliberately more generous free product worth as much as one acquired before?

The bull answer is that the funnel is being widened now and monetised later at greater scale — the letter points to longer free trials, which management expects to raise both revenue and DAUs, and to Video Call, the AI conversation feature, now reaching most new Super Duolingo subscribers with existing Super subscribers to follow later this year. The bear answer is that 23% more daily users producing 6% more constant-currency bookings means revenue per user is falling, and a free-user base that never converts is a cost centre with a nice engagement chart.

Nothing in this quarter resolves that. It is a two-to-four-quarter question.

The margin story is genuinely good, and partly offsets

Easy to miss: Duolingo raised the profitability half of its guidance.

	Q2 2026	Guide
Gross margin	72.6% (vs ~71.0% expected by the company)	~71.0% in Q3, ~71.6% for FY2026
Adjusted EBITDA	$77.3M, 25.9% margin	—
Q3 adjusted EBITDA	—	~$76.0M, 25.2% margin
FY2026 adjusted EBITDA	—	~$320M, 26.5% margin — raised from ~$310M and ~25.7%

The letter credits gross margin to “our measured pace of AI-powered feature expansion (such as Video Call), as well as AI cost efficiencies”, and says the full-year margin outlook is “better than the trajectory we outlined on our Q1 call based on AI cost trends”. The same technology that is costing the company monetisation in the short run is getting cheaper to run. Those two forces partly cancel, which is why full-year profit guidance went up while the topline was reaffirmed.

One caution the headline hides. The within-year margin path is downward: 28.6% in Q1, 25.9% in Q2, 25.2% guided for Q3. Reaching a 26.5% full-year figure therefore requires a fourth quarter better than the third. Working from the company’s own numbers — $83.4 million of Q1 adjusted EBITDA, $77.3 million in Q2 and about $76 million guided for Q3, against a full-year target of about $320 million — the implied fourth quarter is around $83 million, on implied revenue near $315 million, or roughly a 26.5% margin (our arithmetic). That is achievable — the Q1 letter already said margin would be “flat to slightly down in Q3 before rising in Q4” — but the full-year raise depends on it.

The dilution, and the buyback against it

Duolingo told shareholders to expect dilution from equity grants of 3.5–4.0% this year, before any share repurchases, and stock-based compensation of around 15.0% of revenue.

That belongs in any per-share model. Before buybacks, if bookings grow 8% and the share count grows 4%, bookings per share grow about 4% (our arithmetic). Stock-based compensation is also excluded from adjusted EBITDA, so the 26.5% margin figure is calculated before the cost of the shares being issued.

The offset is disclosed in the same letter. Under a $400 million repurchase programme announced in February, Duolingo bought back about $44.4 million of stock, roughly 432 thousand shares, in Q2, and about $71.9 million, roughly 708 thousand shares, through August 1 — which the company says “represents nearly all of 2024 and 2025 dilution”. Against 50.7 million fully diluted shares, that is about 1.4% of the share base so far this year (our arithmetic). Net dilution for 2026 will depend on how much more of the programme is used.

The targets on the record

Target	What management said
Q3 2026 revenue	approximately $302 million, +11.1%
Q3 2026 bookings	approximately $307 million, +8.9%
Q3 2026 adjusted EBITDA	approximately $76.0 million, a 25.2% margin
FY2026 bookings / revenue	$1,285 million (+10.9%) / $1,207 million (+16.3%) — “reaffirming our topline guidance”
FY2026 adjusted EBITDA	approximately $320 million, a 26.5% margin — raised “by nearly one point”
DAU growth	expected to stay above 20% for the rest of the year
Dilution	3.5–4.0% from equity grants, before buybacks
Long-term ambition	teach one billion people

The Q3 constant-currency growth rates are the same 8.9% and 11.1%; for the full year, constant-currency growth is guided at 9.4% for bookings and 14.9% for revenue. The full-year point estimates moved up slightly from the Q1 letter’s 10.5% bookings and 16.1% revenue growth, within the 10–12% and 15–18% targets the company says it continues to manage to.

The most informative pairing there is the profitability raise against the reaffirmed topline and the above-20% user commitment. That is a management team telling you exactly which variable it is optimising and which one it is willing to let sit still.

The bear case, stated fairly

1. Bookings below revenue. Deferred revenue fell about $8 million in the quarter, and reported revenue growth must converge toward the bookings rate.
2. Constant-currency bookings growth of 6% is a very different business from the one the headline growth rate implies. Part of the 8% reported figure is currency, not demand.
3. Guided convergence. Q3 bookings +8.9% against revenue +11.1% — the gap narrows, and the direction of travel for reported revenue growth is down.
4. Monetisation per user is falling. 23% more daily users, 17% more paid subscribers and 6% more constant-currency bookings is a deliberate trade, but it is still a trade.
5. The margin raise leans on Q4. Q3 is guided below Q2, which is below Q1.
6. 3.5–4.0% gross dilution sits between company growth and per-share growth, partly offset by buybacks.

Against that, the bull case is not complicated: DAU growth reaccelerated, gross margin came in ahead of the company’s own expectation, full-year profitability was raised, bookings are guided back above revenue in Q3, and the quarter beat the company’s own guidance on revenue, bookings and adjusted EBITDA.

What to check at the Q3 report

• Bookings growth against the 8.9% guide. This is the number. Everything else is downstream of it. An acceleration here reverses the whole thesis; a miss confirms it.
• Whether bookings exceed revenue, as the Q3 guide ($307 million against $302 million) implies. The quarter deferred revenue starts rebuilding is the quarter the strategy begins to pay.
• DAU growth against the “above 20%” commitment, and whether the Q2 acceleration — partly a one-time Streak Revival event — was one quarter or a trend.
• Paid-subscriber growth against DAU growth. 17% against 23% this quarter. The ratio between them is the conversion question, stated directly.
• Adjusted EBITDA against ~$76 million, and whether the implied Q4 step-up to roughly $83 million still looks reachable.
• Gross margin against ~71.0%. Continued outperformance means the AI cost tailwind is real and durable; a stall means the margin half of the story weakens at the same time as the bookings half.

The read

Duolingo delivered exactly what it said it would: more users, faster than last quarter, at the cost of monetising them more slowly. The beat against its own guidance and the DAU acceleration are real. So is the fact that the cash customers actually committed grew 6% in constant currency, and that revenue — the number most people look at — is a lagging record of decisions made when growth was faster.

The story is not the 0.9% gap between the Q3 revenue guide and analyst estimates that StockStory reported. It is that the leading indicator (bookings) and the reported indicator (revenue) have separated — and that the company’s own guidance says they should come back together by year-end.

For anyone holding DUOL, that makes the tracking job unusually clear. Watch bookings against the 8.9% guide, and watch whether a widening free funnel starts converting into paid subscribers. The company has told you it is buying users now to sell subscriptions later. The bookings line is the receipt.

---

Sources and method. Company figures are from Duolingo’s Q2 2026 shareholder letter and results press release, filed as Exhibits 99.2 and 99.1 to the Form 8-K of 2026-08-05 (accession 0001628280-26-053299), read directly from SEC EDGAR: revenue of $298.5 million and its 18% growth, bookings of $289.1 million at +8% reported and +6% constant currency, subscription bookings of $250.3 million, 58.7 million DAUs at +23% and the two-point sequential acceleration, 12.7 million paid subscribers at +17%, gross profit of $216.7 million and the 72.6% gross margin, operating income of $33.9 million, GAAP diluted EPS of $0.66 against $0.91, adjusted EBITDA of $77.3 million at a 25.9% margin, deferred revenue of $505.1 million at June 30, the share repurchases and diluted share count, all Q3 and full-year guidance including the constant-currency growth rates, the “above 20%” DAU expectation, the 3.5–4.0% dilution expectation and the von Ahn quotation (Exhibit 99.1). The Q2 guidance the quarter is measured against ($295.5 million revenue, $283.5 million bookings, $71.0 million adjusted EBITDA, 5.8% bookings growth), the prior full-year guide (about $310 million adjusted EBITDA, 25.7% margin) and the Q1 comparatives (revenue of $292.0 million, bookings of $308.5 million, 21% DAU growth, $83.4 million adjusted EBITDA at a 28.6% margin, deferred revenue of $513.3 million at March 31) are from the Q1 2026 shareholder letter, Exhibit 99.2 to the Form 8-K of 2026-05-04 (accession 0001628280-26-029790). Q4 2025 revenue of $282.9 million and bookings of $336.8 million are from the Q4 2025 shareholder letter, Exhibit 99.2 to the Form 8-K of 2026-02-26 (accession 0001628280-26-012246). Duolingo’s investor-relations copies of the Q2 letter and release are also linked above.

Not company figures, and attributed accordingly: the analyst-estimate comparison is StockStory’s, which reported that the Q3 revenue guide of $302 million came in 0.9% below analysts’ estimates. Whether the quarter beat or missed its own targets is judged only against the company’s own previously issued guidance (Form 8-K, Exhibit 99.2, filed 2026-05-04) and against the year-ago quarter; the third-party comparison is reported separately and plays no part in that verdict. This piece makes no claim about the share-price reaction.

Our arithmetic, labelled as such: the $8.2 million decline in deferred revenue; the narrowing of the bookings-versus-revenue growth gap from about ten points to about two; the implied fourth-quarter bookings of roughly $380 million (+13%), revenue of roughly $315 million (+11%) and adjusted EBITDA of roughly $83 million (about a 26.5% margin), each derived from the company’s full-year guidance less Q1, Q2 and the Q3 guide; the per-share illustration; and the buyback as about 1.4% of the diluted share base. The interpretation — that revenue growth must converge toward bookings growth, and that the report is best read as one deliberate trade rather than a miss — is our analysis and is not a forecast.

Related: AppLovin Q2 2026 and Spotify Q2 2026, two more consumer-tech reports where guidance, not the quarter’s results, carried the analysis, and Shopify Q2 2026, a contrasting case.

Tradune shows what you actually hold across every account, and whether a position has beaten a plain index fund since you bought it. Educational analysis, not investment advice.