Andrew Boone

Tony confirmed evolution from middle-of-week top-up to larger weekend stock-up baskets. Happening faster with each successive cohort. Existing cohorts increasing share of wallet.

Ravi confirmed ~2% of GOV for 2026 OpEx, noting Q4 included ROO. Investing in areas that drive scale and profitability. Goal is to generate leverage on OpEx like any other P&L line.

Ravi said unit economics improving both sequentially and YoY. Comfortable on path to breakeven from scale and product quality improvements. Focus is on scaling the business.

Tony attributed growth to compounding product improvements initiated 3 years ago. Delighting each cohort raises all cohorts. Leader in acquiring new customers across all categories.

Tony said frequency growth comes from many sources: new categories, DashPass benefits, improvements within each use case. Finding several dozen ways to drive frequency growth for geometric compounding.

Ravi cited all-time high international MAUs, Wolt+ growing faster than DashPass at same stage, gaining share across most countries. Gross profit positive with improving unit economics.

Ravi attributed S&M increase to seasonal Dasher acquisition. Said plenty of room for continued leverage from product improvements. Take rate expected to recover in H2.

Tony described DashPass as the membership program to the physical world. Vision is to maximize ways of connecting local businesses to consumers to drive maximum subscriber utility.

Asked about U.S. top-of-funnel; management said top-of-funnel improved in nearly every region and U.S. growth reaccelerated, aided by influencers and performance marketing.

Pressed on MAU/top-of-funnel deceleration; von Ahn conceded top of funnel was about flat and pointed to marketing and product fixes to reaccelerate word-of-mouth; direct.

Luis said Duolingo is both under- and over-monetized; future monetization via features and customizations, not friction. Direct.

Asked whether AI-driven efficacy is drawing advanced English learners and if the slowdown hit them; Luis said English and advanced-English learners are growing steadily, with English-learning regions like Asia growing fastest.

Said ARPU was up around 6% on plan mix shift and should stay low-single-digit positive, with FX a meaningful tailwind spread across the back-half guide.

Matt said Duolingo optimizes for platform LTV, launched Max after A/B tests showed it was LTV-positive, and sees continued incrementality.

von Ahn said experimentation velocity keeps rising, more A/B tests in 2024 than ever and way more expected in 2025, focused on driving Max subscriptions.

Management reframed the platform flywheel driving downstream Vector and commerce revenue; direct.

Management described a shift toward consumption-based pricing that scales with value; direct.

Management confirmed 90%+ opt-in and Q2 runtime data integration; direct.

Management described democratization plus AI extending the creator base; direct.

Management showed the guide implies double-digit sequential Unity Ad Network growth; direct.

Management stayed prudent given Vector's newness but said it is pacing ahead of expectations.

Management pointed to a large first-party runtime data advantage to be leveraged over time; direct.

Chris credited personalization, order quality across 16 consecutive quarters and affordability; Arpalus adds computer-vision shelf data for fulfillment efficiency.

Chris said Caper is live in more than 100 cities with over a dozen banners, now at about 20% of Wakefern stores with new pilots at Sprouts, Wegmans and Coles, framing in-store as a mid- to long-term growth lever.

Rogers prioritized discoverability across AI platforms over near-term ad monetization, framing ads and AI as a longer-term benefit.

Rogers described a combination path, on-platform innovation, Carrot Ads, ads on Caper, off-platform partners, but no specific timeline. Partial.

Directly explained shopper-pay/batching savings are reinvested into affordability (lower minimums, free pickup); no immediate AI OpEx impact committed.

Attributed momentum to superior ad technology and demand from 7,000 brands creating a virtuous cycle; more than 220 partners on Carrot Ads.

Said growing the machine is an execution not a capital question, and that giving rate back to customers is a deliberate long-term choice given financial independence.

Cited ~30% operating return on net operating assets and said the priority for cash is reinvesting in the machine, not returns yet.

Ernie confirmed the couple-hundred-dollar spread between top and bottom quartile IRC performers remains despite overall improvement. Said opportunity is clear but hard to unlock while growing 40%.

Same-day delivery is primarily technology and process investment, not cost investment. Some incremental staffing for slack capacity but not material in grand scheme.

Pointed to deepening automation across all workflows: finance verifications, registration, customer care, reconditioning, logistics. Insurance document processing as example.

Broke GPU into two halves: reconditioning/transport improvements (~$100) and tariff-related April demand (~$100). Word of mouth discussed as foundation of customer growth.

Huffman gave contribution metrics direction and argued Reddit's creator model is deliberately different from the 1% creator model of social platforms.

Wong laid out the Shopify co-marketing plan and conceded SMB onboarding into Max is not yet optimal and remains foundational work ahead.

Jen said shopping ads are competitive with tier-two peers and need more ML signal work and feed/CAPI adoption to reach tier-one.

Steve said the biggest opportunity is improving the first-open experience for new app users, not converting search visitors.

Steve tied international self-sustainability to product quality, machine translation, moderator recruitment and community discovery.

Jen said pricing is an output of supply/demand plus more valuable impressions from performance work, not something they optimize for.