RDDT · Reddit, Inc.

Internet Content & Information · mkt cap $32.0B · calls: Q1 FY2026 vs Q4 FY2025

65.0 conviction · conf-adj 63

conf 4/10 partial

enthusiasm:24.0 · trend:8 · quantifies:5 · impact:0 · under_radar:5 · credibility:12 · business_impact:8 · disruption:-6 · commitment:6 · confirmation:3

Enthusiasm latest 8 / prev 7 (rising)

Reddit's AI thesis is dual: (1) human conversation as scarce training/citation fuel and counterweight to AI summaries (licensing + traffic), and (2) ML/automation inside product (feed, search/Answers, translation) and ads (Max, DPA, signals) with growing quantified advertiser outcomes. Enthusiasm rose from thematic positioning in Q4 to operational detail in Q1 (Max live, ML hiring, licensing revenue line item), but management still does not quantify AI licensing economics or attribute core revenue growth directly to AI. Credibility is stronger on ads ML and product ML than on "$50–60M is a pimple" licensing upside, where answers stay strategic and non-numeric.

GROUNDED NEXT-FY IMPACT vs CONSENSUS

Grounded on actual base — revenue $2.2B · net income $0.5B · net margin 24.1% · diluted EPS 2.62

These are next-fiscal-year annual uplift estimates, not next-quarter numbers.

Aggregate next-FY est. rev uplift: 0.0% · next-FY EPS uplift: -2.17% · vs analysts: inline · priced in: medium (model's call-read: high; verdict above is the hard-data one used for ranking) · confidence: 4/10

Claim	Figure	Arithmetic	Next-FY Rev %	Next-FY EPS %
Reddit Max CPA reduction 17% engagement · soft	-17% CPA (beta avg)	Advertiser KPI, not Reddit revenue. CPA -17% implies conversions +~20.5% at flat spend; no disclosed Max ad-spend or revenue base to map to Reddit dollars. rev_uplift_pct=null, eps_uplift_pct=null.
Reddit Max +25% conversion outcomes engagement · soft	+25% conversions (avg)	Advertiser outcome metric overlapping Max CPA claim; no disclosed Max revenue base to convert to Reddit dollars. Unanchored.
~50% of Max advertisers using AI creative engagement · soft	~50% adoption	Adoption share only; no revenue or spend-per-advertiser disclosed. Not translatable to next-FY dollars.
Platform conversions doubled YoY engagement · soft	2x conversions YoY	Conversion volume, not revenue; CPA also improving so revenue change ne +100%. No disclosed take-rate or $/conversion. Unanchored.
DPA ROAS >90% higher YoY engagement · soft	>90% ROAS YoY	Advertiser ROAS improvement; DPA revenue share not disclosed. Cannot anchor to $2,202.5M total rev.
Search WAU +30% YoY engagement · soft	+30% search WAU	Search explicitly not yet monetized ('monetization opportunity ahead'). Engagement/option value only, no revenue base. Soft.
Content/other revenue $39M, +15% YoY revenue	$39M/qtr, +15% YoY	Prior-yr other rev $39M/1.15=$33.9M; incr $5.1M/qtr; annualized incremental ~$20.4M = 100x20.4/2,202.5 = ~0.9% of rev. Other rev includes corpus licensing to AI firms = SUPPLIER side. At ~75% incremental margin (high-margin data/services): incr NI ~$15.3M; eps_uplift_pct=100x15.3/529.7=2.88%. Not all $39M is licensing, so this is the low end.	0.9	2.88
ML cost-of-revenue pressure: CoR $56M, +52% YoY cost	$56M CoR/qtr, +52% YoY	Prior $36.8M/qtr; +$19.2M/qtr = +$76.6M annualized. After-tax @21%: -$60.5M / $529.7M NI = -11.4% EPS — but a HEADWIND and an UPPER BOUND: only a portion is ML (also volume + international). Volume-adjusting to CoR growth above the +46.3% consensus rev ramp (~5.7% excess on ~$147M prior CoR ~ $8.5M, after-tax ~$6.7M ~ -1.3%) is used in the adopter aggregate.		-11.4
Shopping ad ML ROAS >75% engagement · soft	>75% ROAS	Advertiser ROAS metric; shopping-ad revenue base not disclosed. Cannot anchor to total rev. Soft.
Weekly search users 80M (from 60M) engagement · soft	80M from 60M WAU (+33%)	Unmonetized search audience; future opportunity, no ARPU/revenue base disclosed. Soft.
Reddit Answers 1M->15M queries engagement · soft	15x query growth	Product usage on tiny base, no revenue/CPM attached; immaterial vs $2.2B rev. Soft.
CapEx $3M Q4 ('AI benefits without AI costs') other · soft	$3M CapEx	Positioning note — Reddit is a CapEx-light AI adopter, no large D&A drag. Supports the margin story but no discrete EPS uplift is quantified; annualizing to a small D&A expense would be inventing phasing. Soft.

Assumptions: Next fiscal year = FY2026 (consensus date 2026-12-31); base revenue $2,202.5M, base NI $529.7M, base EPS comparable to profitable GAAP. Tax 21%. Content-licensing incremental = full Q1 YoY delta annualized (~$20.4M) at ~75% incremental net margin (high-margin data/services); kept on supplier side, excluded from the adopter aggregate. ML CoR drag computed gross/annualized after-tax as an upper bound (ML is only part of the +52% CoR growth, which also reflects volume + international); a volume-adjusted excess (CoR growth above the +46.3% consensus rev ramp) feeds the adopter aggregate. All ad-performance claims (Max CPA/conversions, DPA/shopping ROAS) are advertiser-side ratios with no disclosed Max/DPA revenue base, so not convertible to Reddit dollars without inventing a base — soft. Search/Answers treated as unmonetized engagement. No double-count of duplicate Q1/Q4 Max KPIs.

Top line: Reddit's AI story is overwhelmingly an ADOPTER story: ML-driven ad optimization (Reddit Max -17% CPA, +25% conversions; conversions delivered 2x YoY; DPA ROAS +90%; shopping ROAS +75%) plus a fast-growing but still-unmonetized search/Answers surface (search WAU +30%, 80M weekly searchers, Answers 1M->15M). Every one is disclosed as an advertiser-performance ratio or an engagement count, NOT incremental Reddit revenue, and none carries a disclosed Max/DPA revenue base — so a defensible incremental top-line % cannot be anchored without inventing the base. These gains are real and large but already embedded in the consensus ~+46% FY26 revenue ramp, so adopter incremental-vs-consensus est_rev_uplift_pct ~ 0. The only HARD anchored revenue figure is supplier-side: content/data licensing (other rev $39M/qtr, +15% YoY) -> ~+$20M annualized = ~0.9% of revenue (supplier_rev_uplift_pct 0.9).

Bottom line: The only quantified bottom-line item is a HEADWIND, not an uplift: cost of revenue $56M/qtr, +52% YoY (+$19.2M/qtr, ~+$76.6M annualized) from volume + ML usage + international. After-tax @21% that is -$60.5M = -11.4% of NI as a gross upper bound; since ML is only part of the +52% and most CoR scales with the +46% revenue ramp, the AI-attributable adopter drag is far smaller — volume-adjusting the excess gives roughly -1 to -2% EPS. Net adopter est_eps_uplift_pct ~ -2.17% (modest CoR/D&A headwind), partly offset by CapEx-light positioning ($3M/qtr). Supplier-side licensing adds a high-margin ~+2.9% EPS at ~75% incremental margin, but that sits outside the adopter aggregate. Confidence is low — magnitudes depend on the ML share of CoR, which is undisclosed.

Consensus already embeds the AI ad ramp aggressively: FY2026 revenue $3,221.9M = +46.3% vs FY2025 actual $2,202.5M; NI $1,018.4M = +92.3%; EPS $4.93 = +88.3%. The Max/ML/DPA performance gains and search monetization optionality are the engine of that +46% — and management quantified none of them above what consensus assumes. The single hard supplier figure (licensing +0.9% of revenue) is immaterial against a +46% bar. So the claims justify but do not exceed the trajectory already in the tape; the interesting upside (search monetization) is real but unquantified and forward.

MODEL CONSENSUS (impact)

partial

Both agree all ad/search KPIs are unanchored; reconciled tax treatment and supplier/adopter splits, kept the conservative low-confidence net EPS headwind.

Conflicts reconciled

content eps_uplift_pct: X=2.88 vs Y=null -> used 2.88 because the method requires flowing incremental supplier revenue to EPS; X applied an explicit ~75% licensing margin
ML CoR eps_uplift_pct: X=-14.47 (no tax) vs Y=-11.4 (after-tax) -> used -11.4 because the method applies (1-tax) and it is flagged an upper bound
CapEx $3M: X=cost, eps -0.57 vs Y=soft/null -> used soft/null because it is a CapEx-light positioning note with no discrete uplift quantified
est_rev_uplift_pct: X=0 vs Y=null -> used 0 because adopter ad gains are real but already embedded in the consensus +46% ramp
est_eps_uplift_pct: X=-2.17 vs Y=null -> used -2.17 (X's volume-adjusted computed net headwind) as the only computed aggregate, with low confidence
supplier_rev_uplift_pct: X=0.924 vs Y=0.9 -> used 0.9 (avg 0.912 rounded)

Field	Opus 4.8	GPT-5.5
Rev uplift %	–	–
EPS uplift %	–	–
Priced in	high	–
vs analysts	inline	–
Confidence	4	–
Top line	Reddit's AI story is overwhelmingly an ADOPTER story — ML-driven ad optimization (Reddit Max: -17% CPA, +25% conversions; conversions delivered 2x YoY; DPA ROAS +90%; shopping ROAS +75%) plus a fast-growing but still-unmonetized search/Answers surface (search WAU +30%, 80M weekly searchers, Answers 1M->15M queries). Every one of these is disclosed as an advertiser-performance ratio or an engagement count, NOT as incremental Reddit revenue, and none carries a disclosed Max/DPA revenue base — so a defensible top-line % cannot be computed without inventing the base; adopter aggregate est_rev_uplift_pct = null. The only hard, anchored revenue figure is supplier-side: content/data licensing (other rev $39M/qtr, +15% YoY) -> ~+$20-23M annualized incremental = ~0.9-1.0% of revenue (supplier_rev_uplift_pct 0.9). The realized adopter impact is real and large but is already embedded in the consensus +46% FY26 revenue ramp.	–
Bottom line	The only quantified bottom-line item is a HEADWIND, not an uplift: cost of revenue $56M/qtr, +52% YoY = +$19.2M/qtr (~+$76.6M annualized) from volume + ML usage + international, which after-tax (21%) is -$60.5M = -11.4% of $529.7M NI as an UPPER bound (ML is only a slice of that growth). This is partly offset by Reddit's CapEx-light posture ($3M/qtr) — it consumes AI benefits without the buildout cost, protecting D&A/margins. Net: AI is margin-accretive operationally but management disclosed no quantified bottom-line SAVING, only a partial cost drag, so est_eps_uplift_pct = null (no positive adopter EPS uplift is anchorable; the quantifiable number is a cost headwind).	–
Reasoning	Consensus already embeds the AI ad ramp aggressively: FY2026 revenue $3,221.9M = +46.3% vs FY2025 actual $2,202.5M; NI $1,018.4M = +92.3%; EPS $4.93 = +88.3%. The Max/ML/DPA performance gains and search monetization optionality are the engine of that +46% — and management quantified none of them above what consensus assumes. The single hard supplier figure (licensing +0.9% of revenue) is immaterial against a +46% bar. So the claims justify but do not exceed the trajectory already in the tape; the interesting upside (search monetization) is real but unquantified and forward.	–

Rows highlighted where the two models disagreed.

QUANTIFICATIONS

Reddit Max CPA reduction: 17% (Q1 FY2026 (beta campaigns, on average), topline)
“On average, advertisers are seeing a 17% reduction in cost per action and 25% more conversion outcomes when running Max campaigns.”

Reddit Max conversion outcomes lift: 25% (Q1 FY2026, topline)
“On average, advertisers are seeing a 17% reduction in cost per action and 25% more conversion outcomes when running Max campaigns.”

Max advertisers using AI-powered creative: ~50% (Q1 FY2026, topline)
“about 50% of Max campaign advertisers using AI-powered creative features to unlock even stronger performance.”

Platform conversions delivered (ML/optimization): doubled YoY (Q1 FY2026 vs Q1 prior year, topline)
“In Q1, we doubled the number of conversions delivered for advertisers across the platform versus last year”

DPA ROAS improvement: >90% YoY (recent DPA investments, topline)
“Recent investments in DPA delivered more than 90% higher ROAS year-over-year on average”

Search WAU growth: 30% YoY (Q1 FY2026, topline)
“Search WAUqs are up 30% year-over-year.”

Content licensing / other revenue (includes licensing): $39 million, +15% YoY (Q1 FY2026, topline)
“Other revenue, which included revenue from our Content Licensing business reached $39 million, up 15% year-over-year.”

ML-related cost of revenue pressure: CoR $56M, +52% YoY (volume, ML usage, international speed/reliability) (Q1 FY2026, bottomline)
“Those cost of revenue increases reflect volume growth and users and ads served more ML usage and more international investments in speed and reliability.”

Reddit Max CPA reduction (pre-launch testing): 17% (Q4 FY2025 testing, topline)
“In testing, Max campaigns delivered an average 17% CPA reduction and a 27% conversion volume validating it as a performance driver for our partners.”

Reddit Max conversion volume lift (testing): 27% (Q4 FY2025 testing, topline)
“In testing, Max campaigns delivered an average 17% CPA reduction and a 27% conversion volume validating it as a performance driver for our partners.”

Shopping ad ML ROAS improvement: over 75% (since prior year, Q4 FY2025, topline)
“enhancements to our shopping ad ML models delivered over 75% improvement in advertiser ROAS.”

Weekly search users: 80 million (from 60 million) (Q4 FY2025 vs year ago, topline)
“with over 80 million people searching directly on Reddit, Inc. every week in Q4, up from 60 million just a year ago.”

Reddit Answers query growth: 1 million to 15 million queries (last year (Q4 call), topline)
“Red answers queries, I believe we're up from 1 million to 15 million queries over the last year.”

CapEx / AI cost positioning: CapEx $3 million Q4 (Q4 FY2025, bottomline)
“Continue to benefit from AI in many ways without the AI costs. CapEx was $3 million.”

PAST (realized)

Q1: Seventh consecutive quarter 60%+ revenue growth; machine translation carried weight last couple of quarters in 30 languages with lower cost (Huffman).
Q1: Reddit Max launched to beta early Q1 with adoption and performance outcomes (Wong).
Q1: Doubled conversions delivered for advertisers vs last year; DPA >90% ROAS YoY (Wong).
Q1: Onboarding experiments ramped, search WAU +30% YoY, bot verification shipped Q1 (Huffman).
Q4: 2025 revenue $2.2B +69%; unified search with Reddit Answers; 80M weekly searchers vs 60M prior year (Huffman).
Q4: Reddit Max testing showed 17% CPA reduction and 27% conversion volume lift (Wong).
Q4: Shopping ad ML models delivered 75%+ ROAS improvement since prior year (Wong).
Q4: Lower-funnel conversion revenue doubled YoY in Q4 (Wong).
Q4: Verified profiles launched Q4; Reddit Answers in five new languages (Huffman).
Q4: Reddit cited as #1 source in AI answers; Google/OpenAI partnerships described as healthy (Huffman).

CURRENT (now)

Q1: Integrating automation and AI into ad stack; ~50% of Max advertisers use AI-powered creative (Wong).
Q1: Hiring ML engineering; higher ML usage in cost of revenue (Vollero).
Q1: Users validate LLM outputs on Reddit; licensing/content in other revenue $39M (Wong/Vollero).
Q1: Feed ML investment, Passkeys/bot labeling work underway (Huffman).
Q1: Reddit Answers with agentic behavior and product catalog integration in search (Huffman).
Q4: Public beta Reddit Max at CES; piloting agentic search results (Huffman/Wong).
Q4: LLM-assisted onboarding triangulation; ML cold-start feed investment (Huffman).
Q4: Using AI/ML for feed personalization from app open (Huffman).
Q4: Search incremental to engagement; search ads not monetized yet (Wong).

FORWARD (guidance)

Q1: 2026 priorities — top of funnel, retention, speed; feed as major driver via ML (Huffman).
Q1: Move new advertisers directly into Reddit Max; Max in API for partners (Wong).
Q1: Licensing renewals/partner expansion discussed qualitatively; no new deal terms (Huffman).
Q1: More bot verification, Passkeys, post-creation improvements, AI spam protection for new users (Huffman).
Q1: Search monetization opportunity ahead; international DPA/catalog expansion (Wong/Huffman).
Q4: 2026 — faster, more relevant Reddit; expand Reddit Max automation using 24B posts/comments as signals (Huffman/Wong).
Q4: Scale brand auto-bidding, auto targeting; deepen video and measurement (Wong).
Q4: New advertiser onboarding via Reddit Max quarters out after conversion focus (Wong).
Q4: Handle more queries with Answers over time (Huffman).

TRACK RECORD — PROMISE vs DELIVERY

76/100 track record mixed 6 calls reviewed

Reddit rarely gives forward-looking, dated numeric AI KPIs; management mostly reports ML/genAI results after the fact. The few testable commitments (ad-review automation, ad-format impression mix, Reddit Max efficiency claims) were largely met, with a small slip on Max conversion uplift and the 100M U.S. DAU goal still unjudged.

Ads-in-comments placement up to mid-single-digits % of impressions in the near term — promised Q4 FY2024
delivered Q1 FY2025 reported ~6% of impressions from the format, above the mid-single-digit target.

LLM ad review: +70% automated review volume and review time cut from 30 minutes to 1 minute — promised Q4 FY2024
delivered Announced as live at launch; later calls keep citing LLM/automated ad review without walking back the capability, though they do not re-report the 70%/1-minute figures.

Reddit Max beta: ~17% average CPA reduction and ~27% higher conversion volume vs. non-Max campaigns — promised Q4 FY2025
partial Q1 FY2026 after broader beta launch cited the same 17% CPA reduction but ~25% more conversion outcomes (vs. 27% at CES).

Reach 100 million daily active users in the U.S. — promised Q1 FY2026
too-early Stated as an active strategic goal with no deadline; company reported ~121M global DAUq in Q4 FY2025 but did not report U.S. daily users at that level in the transcript set.

Scale interactive developer-platform ads more broadly next year — promised Q3 FY2025
too-early Q4 FY2025 launched interactive ads to beta (e.g., Paramount); Q1 FY2026 did not report scaled adoption metrics against a numeric target.

Gen-AI sales tooling to 10x insights reports used on campaigns — promised Q1 FY2025
delivered Reported as achieved in the same quarter; no earlier call had set a dated 10x target to score as a forward promise.

PRICED-IN (REFINED)

MEDIUM

Est. revisions flat · Fwd P/E 71.5 · EV/Sales 12.3x

AI claim maps to Advertising, Other Revenue

Rating counts have been stable since Feb 2026 (6 strong buy / 15 buy / ~10 hold) with only a modest Jan step-up, while price targets have been cut (last-year avg $230 vs last-quarter $198 vs $166 spot), so revision momentum is not rising. Valuation is rich (71.5x next-FY EPS, ~12.3x EV/Sales) and consensus already embeds aggressive AI-linked growth (FY26 revenue ~$3.2B, EPS ~$4.93 vs ~$2.31 in FY25). AI upside most plausibly flows into Advertising and Other Revenue, not the whole P&L equally. Rich multiples plus baked-in growth argue much of the thesis is in the price, but flat/falling target momentum prevents a full "high" (rising estimates + rich) call.

COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS

7Q4 FY20247Q1 FY20259Q2 FY20258Q3 FY20258Q4 FY20258Q1 FY2026

AI enthusiasm across 6 calls — trend ↗ rising

From beta Answers and translation to scaled search, ad ML, Reddit Max, and data-for-AI positioning.

RECENT AI CATALYSTS & NEWS

2026-05-29 · Reddit: The Most Human Place On The Internet
2026-05-28 · Meta's AI Push Seeks To Outpace Reddit And Other Competitors
2026-05-24 · Reddit: Back In The Game
2026-05-21 · I'm Buying Reddit Despite The Google Headwind. Here's Why
2026-05-20 · Reddit Benefits From Solid ARPU Growth: A Sign for More Upside?
2026-05-15 · RDDT vs. AMZN: Which Digital Advertising Stock Is a Better Buy?

BUSINESS IMPACT - QUALITATIVE MATERIALITY

8/10 qualitative impact material medium-term · mixed evidence

Where AI matters: ads ML optimization + feed/search product

ML is operationally central to Reddit Max/DPA and the feed (disclosed -17% CPA, 2x conversions, search WAU +30%) and likely supports the ad ramp, but disclosed upside is mostly advertiser KPIs not incremental Reddit revenue above ~$3.2B FY26 consensus, while hard licensing/other rev (~$39M/qtr) remains tiny.

Caveats: Licensing ~6% of revenue with no disclosed renewal uplift—'oil' narrative may outrun economics; Search/Answers monetization still forward while AI answer engines compete for intent; ML usage is lifting cost of revenue (+52% YoY) with undisclosed AI-attributable share; FY26 ~+46% revenue consensus likely already prices Max/ML ad gains—limited incremental vs tape

AI DISRUPTION / CANNIBALIZATION RISK two-sided · 5/10

LLM summaries and AI search (e.g. Google) can divert discovery/Q&A traffic and commoditize shallow queries, yet the durable moat is authenticated human discourse that models cannot replicate—partially offset by validation traffic and data-licensing supplier economics.

OPTIONS / MARKET STRUCTURE

option liquidity: good

ATM IV

—

TYPICAL BID-ASK

—

OPEN INTEREST

—

proxy inputs — dollar-ADV $837M · beta 1.851 · px $165.50

source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.

CONFIRMATION — INSIDERS · 13F · LANGUAGE

Mixed — insiders selling, institutions adding, management language 7/10 committed.

INSIDERS selling 63 open-market sell(s) vs 0 buy(s) — net distribution

INSTITUTIONS (13F) adding as of 2026-03-31: 139 new / 226 closed positions; 459 increased / 258 reduced; institutional ownership -3.25pp; -83 net 13F holders

MGMT LANGUAGE 7/10 committed Ad-stack AI/automation uses firm metrics and delivery language; macro AI positioning stays conditional and aspirational.

commit “On average, advertisers are seeing a 17% reduction in cost per action and 25% more conversion outcomes when running Max campaigns.”

commit “In Q1, we doubled the number of conversions delivered for advertisers across the platform versus last year”

commit “we've added critical machine learning talent to build the capabilities required for today's Internet”

VERBATIM AI QUOTES

“Another reason Reddit stands out today is our position in the AI landscape. Reddit is built on more than 2 decades of human conversation. Over 25 billion posts and comments and every month, our communities generate the equivalent of Wikipedia's entire content library in new content. As AI becomes more prevalent, people increasingly seek out real human perspectives and in turn, AI models rely on these perspectives to train and power their products. Scarce assets tend to become more valuable over time and authentic human conversation at scale is becoming increasingly rare. Reddit's conversations are like oil for the modern Internet, a foundational resources powering the next generation of technology.”
— Steven Huffman, Q1 FY2026

“we've added critical machine learning talent to build the capabilities required for today's Internet”
— Steven Huffman, Q1 FY2026

“continued success with machine translation”
— Steven Huffman, Q1 FY2026

“While people are using AI summarized information, people are also increasingly wanting to see and incorporate a breadth of perspectives from other people into their decision-making. The value of authentic human perspective is increasing as more information is generated and summarized by models.”
— Jennifer Wong, Q1 FY2026

“people are coming to Reddit to validate what they read and hear elsewhere, including the responses they get from LLMs.”
— Jennifer Wong, Q1 FY2026

“Our strategy is to integrate more automation and AI into our ad stack to enable faster adoption of new features, make sure advertisers are set up to get the best performance from Reddit ads and increase the productivity and impact of our sales team.”
— Jennifer Wong, Q1 FY2026

“On average, advertisers are seeing a 17% reduction in cost per action and 25% more conversion outcomes when running Max campaigns. Advertisers are also increasingly adopting AI in their campaign setups with about 50% of Max campaign advertisers using AI-powered creative features to unlock even stronger performance.”
— Jennifer Wong, Q1 FY2026

“our investments in the ad stack, including machine learning for signal optimization and ad formats”
— Jennifer Wong, Q1 FY2026

“In Q1, we doubled the number of conversions delivered for advertisers across the platform versus last year”
— Jennifer Wong, Q1 FY2026

“Recent investments in DPA delivered more than 90% higher ROAS year-over-year on average”
— Jennifer Wong, Q1 FY2026

“receive AI-powered community recommendations on where to share them”
— Jennifer Wong, Q1 FY2026

“Those cost of revenue increases reflect volume growth and users and ads served more ML usage and more international investments in speed and reliability.”
— Andrew Vollero, Q1 FY2026

“We're selectively hiring talent in key revenue and consumer functions like sales, ad tech and ML engineering.”
— Andrew Vollero, Q1 FY2026

“hence, the kind of focus on ML talent for us right now”
— Steven Huffman, Q1 FY2026

“Search WAUqs are up 30% year-over-year.”
— Steven Huffman, Q1 FY2026

“carrying most of the weight the last couple of quarters has been machine translation were translated in the 30 languages today. We've been able to lower the cost there”
— Steven Huffman, Q1 FY2026

“We have important partnerships with both Google and OpenAI. Those are very meaningful to us.”
— Steven Huffman, Q1 FY2026

“at the end of the day, there is no artificial intelligence without actual intelligence, and that comes from Reddit.”
— Steven Huffman, Q1 FY2026

“Reddit has been for a while and continues to be the most cited source in AI citations across all platforms.”
— Steven Huffman, Q1 FY2026

“post guidance, which is an LLM that basically helps the user navigate the rules of Reddit have been a big driver there.”
— Steven Huffman, Q1 FY2026

“working our way out of age and Karma limits with better AI-powered spam protection”
— Steven Huffman, Q1 FY2026

“bringing in new users, advancing our own AI technology. So things like the machine translation, the LLM powered onboarding, all of the safety things”
— Steven Huffman, Q1 FY2026

“Reddit answers, you can see it better integrated into the product. It itself has more agentic behavior behind the scenes.”
— Steven Huffman, Q1 FY2026

“In a world flooded with AI swap, people are seeking real community, lived experience, and trusted opinions.”
— Steven Huffman, Q4 FY2025

“But in the age of AI, you can't easily distinguish a real person's thoughts or recommendations from a bot that trust erodes.”
— Steven Huffman, Q4 FY2025

“we made significant progress in unifying our core search with Reddit Answers, our AI-powered search feature.”
— Steven Huffman, Q4 FY2025

“with over 80 million people searching directly on Reddit, Inc. every week in Q4, up from 60 million just a year ago.”
— Steven Huffman, Q4 FY2025

“we are piloting dynamic agentic search results that include media beyond text.”
— Steven Huffman, Q4 FY2025

“improving feed relevancy. Using AI and machine learning to make Reddit, Inc. feel more personalized and useful from the second you open the app.”
— Steven Huffman, Q4 FY2025

“As we see the benefits of our investments in ML and new ad formats like shopping ads, start to pay off.”
— Jennifer Wong, Q4 FY2025

“Reddit Max campaign, our AI-powered campaign platform that uses Reddit, Inc. community intelligence to optimize performance for middle and lower funnel objectives.”
— Jennifer Wong, Q4 FY2025

“In testing, Max campaigns delivered an average 17% CPA reduction and a 27% conversion volume validating it as a performance driver for our partners.”
— Jennifer Wong, Q4 FY2025

“enhancements to our shopping ad ML models delivered over 75% improvement in advertiser ROAS.”
— Jennifer Wong, Q4 FY2025

“AI-powered insights from Reddit, Inc. community intelligence can accelerate how brands turn community conversation into actionable media strategies.”
— Jennifer Wong, Q4 FY2025

“Continue to benefit from AI in many ways without the AI costs.”
— Andrew Vollero, Q4 FY2025

“Streamlining that process does improve retention. Think we've got some interesting things about using LLMs to how to help triangulate users' interest.”
— Steven Huffman, Q4 FY2025

“Bringing users into the feed faster requires the feed to be better. And so we'll be making pretty good investment into ML to improve that kind of cold start feed for new users.”
— Steven Huffman, Q4 FY2025

“Reddit, Inc. per profound is the number one cited source in AI answers.”
— Steven Huffman, Q4 FY2025

“The conversation is shifting from know, a purely business deal to you know, more of a product partnership.”
— Steven Huffman, Q4 FY2025

“Reddit, Inc. is the point of trusted recommendations. It's where the human who actually has to deploy resources and make decisions is actually searching for what it is that they're interested in buying.”
— Jennifer Wong, Q4 FY2025

“that position is very, very important for all marketers, I think, for any business.”
— Jennifer Wong, Q4 FY2025

ANALYST QUESTIONS ON AI

Q (Q1 FY2026, George Josh Beck): Double-click on the ML talent — short list of projects the team is working on, what was most important this year?
A: Steve Huffman: ML is "basically everything" for the feed — more signals, better weighting, faster model-to-production; "entire stack" and processes; talent from platforms with billions of users; goal 50M to 100M U.S. users via feed improving retention and frequency.

Q (Q1 FY2026, Jason Helfstein): Discussions around AI and third-party agents leveraging your data — how potentially you can get credit for DAU generated on third parties, part of larger AI licensing discussions?
A: Steven Huffman: DAU is top priority; on AI/third-party agents — "ecosystem we live in," meaningful Google and OpenAI partnerships, mutual value, look for top-of-funnel opportunities, "nothing new to share on those specific relationships at this time."

Q (Q1 FY2026, Richard Greenfield): If you're the oil powering the modern Internet, $50 million to $60 million a year from Google and OpenAI seems like a pimple — how are conversations changing heading into 2027 renewals? Is there value to an exclusive deal?
A: Steven Huffman: No timeline on 100M DAU; data valuable to partners; "no artificial intelligence without actual intelligence, and that comes from Reddit"; more AI/sanitized internet increases craving for human information; on exclusive deal: "No comment on that, Rich."

Q (Q1 FY2026, Justin Post): Update on how generative AI engines are using your data — increasing since deals started? Google algorithm changes in April — impacts on usage/retention?
A: Steven Huffman: Reddit "most cited source in AI citations"; Reddit among top Google searches; people validate AI with human context; Google algo changes are business as usual, "nothing further to comment on" for the quarter.

Q (Q1 FY2026, Benjamin Black): Data licensing — what criteria beyond dollars to go from 2 partners to 3–4?
A: Steven Huffman: Citations, mind share, multifaceted partnerships; beyond dollars — bring users in, advance own AI (machine translation, LLM-powered onboarding, safety); relationships meaningful beyond dev/business deal.

Q (Q1 FY2026, Naved Khan): Rollout of AnswerPlus search in the U.S. — session time or other benefits?
A: Steven Huffman: Search DAU/WAU/queries up meaningfully YoY; retention and DAU driver; Answers more agentic (compare two things); integrating product search catalog; "contributor to basically all of the things we care about."

Q (Q4 FY2025, Benjamin Black): AI-generated content on the platform — additive or not? Positioning for agentic commerce impact on users, contribution rates, ads?
A: Steven Huffman: Machine translation is AI-generated; human behind prompts vs full bots — latter not wanted; transparency and labeling; Reddit for humans. Jennifer Wong: Reddit well positioned — trusted recommendations before commoditized execution layer; marketers need the deciding human; present on Reddit and in LLM partnerships.

Q (Q4 FY2025, Justin Post): AI deals with Google and OpenAI — how using data, growing in importance? Bot expectations — user impact or revenue?
A: Steven Huffman: Reddit #1 cited in AI answers; healthy relationships; shifting to product partnership — mutual best products, bring users into Reddit communities; bots removed before sharing data; agentic comment-writing; labeled bots OK; spam evolution.

Q (Q4 FY2025, Richard Greenfield): Magic wand for Google/OpenAI citations — how should they look to drive people into Reddit conversational content?
A: Steven Huffman: Wants users aware they can go to relevant communities (e.g., audiophile for speakers); lots of product movement; close connection, healthy relationships, expect continued evolution together.

Q (Q4 FY2025, John Colantuoni): Reddit Answers/search — impact on monetization near/long term?
A: Jennifer Wong: Search not monetized yet but huge market; search behavior "incremental and additive"; often shopping/high-intent; excited, opportunity ahead.

Q (Q4 FY2025, Andrew Boone): Generative AI tools — what experience are you trying to build for people coming to Reddit looking for things?
A: Jennifer Wong (DPA portion) and context from Steve on search: unified search/Answers, growing queries, more queries handled with answers for flexibility. (Steve on search integration covered in Ron Josey answer: Answers queries 1M to 15M, 60–80M overall search.)