← back to rankingSNAP · Snap Inc.
Internet Content & Information · mkt cap $9.6B · calls: Q1 FY2026 vs Q4 FY2025
72.0 conviction · conf-adj 72
conf 3/10 partial
enthusiasm:27.0 · trend:8 · quantifies:0 · impact:0 · under_radar:14 · credibility:12 · business_impact:8 · disruption:-6 · commitment:6 · confirmation:3
Enthusiasm latest 9 / prev 8 (rising)
Snap's AI thesis is operational, not rhetorical: ML ranking and automation on ads (now including LLM/VLM retrieval), generative/AI Lens creation driving engagement and Lens+ subscriptions, and on-device generative camera models—with management citing conversion lifts and ~70% of spend on AI ad automation. Credibility is strengthened by specific incremental metrics on DPA/app models and subscription ARPU from Lens+, though many figures are internal conversion lifts rather than disclosed AI revenue lines; Q1 adds product launches (AI Sponsored Snaps) and explicit AI training in infrastructure costs while Q4 emphasized ecosystem scale (17B generative Lens engagements) and internal engineering productivity (~40% AI-generated code).
GROUNDED NEXT-FY IMPACT vs CONSENSUS
Grounded on actual base — revenue $5.9B · net income $-0.5B · net margin -7.8% · diluted EPS -0.27
These are next-fiscal-year annual uplift estimates, not next-quarter numbers.
Aggregate next-FY est. rev uplift: % · next-FY EPS uplift: % · vs analysts: in-line · priced in: low · confidence: 3/10
| Claim | Figure | Arithmetic | Next-FY Rev % | Next-FY EPS % |
|---|
Spotlight watch time +11% YoY (AI personalization)
engagement · soft | 11% YoY | Engagement KPI only; no Spotlight monetization rate or revenue $ disclosed → cannot map 11% watch time to the $5,931M revenue base. Directionally supports impression/ad-load growth. | | |
AI-powered Lens submissions 400K+ Q1, >150% YoY
engagement · soft | 400,000+; +150% YoY | 400,000/2.5 = 160,000 prior-Q1 (+240,000 units). Creator-supply volume; no submission→revenue $ bridge disclosed. Unanchored. | | |
~70% of ad spend uses AI automation (Smart Audience/Budget/Placement)
engagement · soft | ~70% | Adoption/penetration breadth, not an incremental lift. The pool (0.70×$5,931M≈$4,152M) is disclosed, but no delta vs non-AI is given, so no incremental revenue is anchorable. Unanchored on impact. | | |
DPA LLM user-intent retrieval — Pixel Purchase conversions +>2%
revenue · soft | >2% | Conversion lift improves advertiser ROI and can pull spend, but sizing it (Y's 0.70×rev×2%×0.50 elasticity → 0.70%) requires an invented budget-elasticity factor not disclosed anywhere. No Pixel Purchase revenue base. Unanchored. | | |
DPA VLM similar-product retrieval — conversion lift
revenue · soft | high single-digit | Vague magnitude ('high single-digit') AND no DPA revenue base. Double-unanchored; cannot compute without inventing a digit. | | |
App Re-engagement model — purchase conversions +~2%
revenue · soft | ~2% | Same as DPA LLM row: conversion lift on a sub-segment of undisclosed size; mapping to $ needs an invented elasticity. Unanchored. (Same Q1 upgrade as the −9% CPA row — would be excluded to avoid double-count if sized.) | | |
App Re-engagement model — CPA −~9%
revenue · soft | ~9% | Advertiser cost efficiency (helps demand/yield indirectly), not Snap opex. Alternate metric for the same Q1 upgrade as the +2% conversion row. Sizing it (Y's upper-bound 3.15%) again relies on an invented spend-expansion elasticity. Unanchored. | | |
Pixel Purchase 7-0 purchases/$ +>23% YoY
revenue · soft | >23% YoY | Backward-looking Q1 ROAS/efficiency already in run-rate; no Pixel Purchase revenue $ share disclosed, so forward incremental is not quantifiable without double-counting. Unanchored. | | |
Infrastructure costs $401M Q1, +7% YoY (incl. AI model training)
cost | $401M, +7% YoY | HARD $: $401M/1.07 = $374.8M prior Q1; ΔQ1 ≈ $26.2M; annualized ≈ +$105M (1.77% of revenue) AI-investment cost INCREASE — a bottom-line DRAG, not a saving. Topline ~0. Loss-making base → EPS% meaningless (null); flag as headwind. | 0 | |
Generative AI Lens engagement (700M+ users, 17B+ engagements)
engagement · soft | 700M+; 17B+ | Cumulative engagement scale; no ARPU, ad load, or Lens revenue $ disclosed. Unanchored. | | |
Imagine Lens engagements ~2B
engagement · soft | ~2B | Engagement count since Sept launch; no monetization rate or revenue $ disclosed. Unanchored. | | |
In-app optimization revenue +89% YoY
revenue · soft | +89% YoY | Explicit growth rate but NO absolute 'in-app optimization revenue' $ disclosed anywhere → cannot compute Δ$ or % of the $5,931M total. High growth on an unknown-size line = unanchored. | | |
Spotlight reposts/shares +69% YoY (U.S.)
engagement · soft | +69% YoY US | Sharing/virality KPI; no Spotlight ad revenue $ or U.S. revenue share disclosed. Unanchored. | | |
~40% of new code AI-generated
productivity · soft | ~40% | Eng-productivity signal, but no R&D/opex $ or FTE base disclosed → cannot convert to after-tax saving. Topline ~0. Unanchored. | | |
DPA CPA −55% (7-0) / −45% (1-0), internal GBB testing
revenue · soft | 55% / 45% | Strongest efficiency claim (−55% CPA ⇒ ~122% more actions at constant spend), but it lowers advertiser cost, not Snap's, is a cumulative internal-test run-rate, and no GBB/Pixel Purchase revenue base is disclosed to size pull-through. Unanchored. | | |
Lens+ subscription — higher ARPU / gross-margin expansion
revenue · soft | no % given | Directional only ('encouraging','higher','expansion'); no ARPU $, subscriber count, or margin % disclosed. Vague/unanchored by definition. | | |
Assumptions: Default incremental net margin = current net margin (−7.76%), tax rate 21% — both moot: no claim carries a disclosable revenue/saving $ base. Loss-making guardrail invoked: GAAP NI = −$460.5M and consensus EPS stays negative through FY26 (≈−$0.067), so any EPS-uplift % off that base is an artifact → all eps pcts and est_eps_uplift_pct = null. Disclosed-base rule applied: searched every claim for a reusable base — the ONLY anchorable dollar figure is the $401M infra COST (marked HARD, sized as a ~$105M/yr headwind, topline 0). Conversion-lift rows (+2%, ~9% CPA, etc.) describe sub-segments of undisclosed size; turning them into revenue requires an invented budget elasticity (Y assumed 0.50), which the no-invented-numbers rule forbids, so they remain unanchored. Next-FY = FY2026 (Dec). No bookings/backlog items present; no revenue claim sized.
Top line: Management lists a broad, genuinely real AI footprint — Spotlight watch time +11% YoY, ~70% of all ad spend now routed through AI automation, DPA Pixel-Purchase conversions +>2% (LLM) plus high-single-digit (VLM), App Re-engagement conversions +~2%, 7-0 purchases-per-dollar +23% YoY, in-app-optimization revenue +89% YoY, and CPA reductions of 45–55% in internal testing — but NOT ONE of these discloses the dollar revenue base of the affected ad sub-segment. Every figure is a percentage lift or engagement count on a line of unknown size, so none can be translated into a defensible incremental-revenue %. The aggregate adopter-side rev uplift is therefore null (cannot anchor without inventing a base), even though the qualitative direction clearly supports better ad yield and engagement.
Bottom line: The only hard dollar figure is a cost, and it cuts the wrong way: infra (incl. AI model training/serving) ran $401M in Q1, +7% YoY ≈ +$105M/yr of incremental spend — an AI-investment headwind, not a saving. The '40% of new code is AI-generated' productivity claim has no $ or headcount anchor, so no after-tax saving can be booked. With GAAP net income at −$460.5M (and consensus EPS still negative through FY26), the loss-making guardrail makes any EPS-uplift % meaningless → est_eps_uplift_pct = null. Net read: AI is currently a cost line being scaled into monetization, not yet a quantified earnings driver.
[EPS uplift n/m (loss-making base)] Consensus already embeds the AI story: FY26 revenue $6,699.3M vs FY25 $5,931.4M = +$767.9M, +12.94% growth, with the net loss narrowing from −$460.5M (GAAP) / −$565.6M (consensus) to −$129.9M in FY26 and turning positive (+$211.3M, EPS +$0.123) by FY27. The engagement and ad-performance gains management cites are exactly the mechanism behind that +13% trajectory and margin recovery — and no claim sizes an incremental AI contribution that lands clearly ABOVE that path (none is sized at all). With every uplift unanchored and the one hard number being a cost increase, there is no demonstrable upside to consensus; the AI initiatives look fully reflected in the existing forward curve.
MODEL CONSENSUS (impact)
partial
Preferred X's conservative no-invented-numbers stance over Y's elasticity-based revenue sizing; kept Y's $401M infra arithmetic and prior-period unit calcs. All EPS null (loss-making base).
Conflicts reconciled
- rev_uplift_pct(DPA Pixel +2%): X=null vs Y=0.7 -> used null because Y's 0.50 budget-elasticity factor is invented, not disclosed
- rev_uplift_pct(App Re-engagement +2%): X=null vs Y=0.7 -> used null (same invented-elasticity issue)
- rev_uplift_pct(App Re-engagement CPA -9%): X=null vs Y=3.15 -> used null (invented spend-expansion elasticity)
- soft(Pixel 7-0 +23% & DPA CPA -55/-45): X=true vs Y=false -> used true because no revenue base anchors the impact
- rev_uplift_pct(Infra cost): X=null vs Y=0 -> used 0 because it is a cost claim (topline ~0)
- est_rev_uplift_pct: X≈null vs Y≈0.7 -> used null since the only sized rev rows rely on invented elasticity
| Field | Opus 4.8 | GPT-5.5 |
|---|
| Rev uplift % | – | – |
| EPS uplift % | – | – |
| Priced in | high | – |
| vs analysts | inline | – |
| Confidence | 3 | – |
| Top line | Management lists a broad, genuinely real AI footprint — Spotlight watch time +11% YoY, ~70% of all ad spend now routed through AI automation, DPA Pixel-Purchase conversions +>2% (LLM) plus high-single-digit (VLM), App Re-engagement conversions +~2%, 7-0 purchases-per-dollar +23% YoY, in-app-optimization revenue +89% YoY, and CPA reductions of 45–55% in internal testing — but NOT ONE of these discloses the dollar revenue base of the affected ad sub-segment. Every figure is a percentage lift or engagement count on a line of unknown size, so none can be translated into a defensible incremental-revenue %. The aggregate adopter-side rev uplift is therefore null (cannot anchor without inventing a base), even though the qualitative direction clearly supports better ad yield and engagement. | – |
| Bottom line | The only hard dollar figure is a cost, and it cuts the wrong way: infra (incl. AI model training/serving) ran $401M in Q1, +7% YoY ≈ +$105M/yr of incremental spend — an AI-investment headwind, not a saving. The '40% of new code is AI-generated' productivity claim has no $ or headcount anchor, so no after-tax saving can be booked. With GAAP net income at −$460.5M (and consensus EPS still negative through FY26), the loss-making guardrail makes any EPS-uplift % meaningless → est_eps_uplift_pct = null. Net read: AI is currently a cost line being scaled into monetization, not yet a quantified earnings driver. | – |
| Reasoning | Consensus already embeds the AI story: FY26 revenue $6,699.3M vs FY25 $5,931.4M = +$767.9M, +12.94% growth, with the net loss narrowing from −$460.5M (GAAP) / −$565.6M (consensus) to −$129.9M in FY26 and turning positive (+$211.3M, EPS +$0.123) by FY27. The engagement and ad-performance gains management cites are exactly the mechanism behind that +13% trajectory and margin recovery — and no claim sizes an incremental AI contribution that lands clearly ABOVE that path (none is sized at all). With every uplift unanchored and the one hard number being a cost increase, there is no demonstrable upside to consensus; the AI initiatives look fully reflected in the existing forward curve. | – |
Rows highlighted where the two models disagreed.
QUANTIFICATIONS
Spotlight total watch time (AI-driven personalization contribution): 11% year-over-year increase (Q1 FY2026, topline)
“continued investment in AI-driven personalization, resulted in higher engagement with total time spent watching Spotlight increasing 11% year-over-year.”
AI-powered Lens submissions: 400,000+ in Q1; more than 150% YoY (Q1 FY2026, topline)
“AI-powered Lens creation is transforming our AR ecosystem, with more than 400,000 lenses submitted in Q1, increasing more than 150% year-over-year.”
Ad spend using AI automation (Smart Audience/Budget/Placement): Nearly 70% (Q1 FY2026, topline)
“Nearly 70% of advertising spend now use at least one of our AI-powered automation solutions including Smart Audience, Smart Budget or Smart Placement”
DPA Pixel Purchase conversion lift from LLM user-intent retrieval: more than 2% (Q1 FY2026 launch, topline)
“LLM based user intent understanding for Dynamic Product Ads retrieval, which improved Pixel Purchase conversions by more than 2%”
DPA conversion lift from VLM similar-product retrieval: additional high single-digit lift (Q1 FY2026 launch, topline)
“multimodal similar product retrieval using a vision language model fine-tuned on Snap data, which delivered an additional high single-digit lift in DPA purchase conversions.”
App Re-engagement model — purchase conversions: approximately 2% increase (Q1 FY2026 upgrade, topline)
“increasing purchase conversions by approximately 2%, while improving CPA by nearly 9%.”
App Re-engagement model — CPA: nearly 9% improvement (Q1 FY2026 upgrade, topline)
“improving CPA by nearly 9%.”
Pixel Purchase efficiency (7-0 purchases per dollar): more than 23% year-over-year growth (Q1 FY2026, topline)
“Across Pixel Purchase campaigns, 7-0 purchases generated per dollar of revenue grew more than 23% year-over-year”
Infrastructure costs (includes AI model training): $401 million in Q1, up 7% YoY (Q1 FY2026, bottomline)
“Total infrastructure costs were $401 million in Q1, up 7% year-over-year, driven primarily by community growth, strategic investments in AI model training and monetization serving costs.”
Generative AI Lens engagement: 700M+ Snapchatters; 17B+ engagements (Q4 FY2025 (cumulative/ongoing), topline)
“More than 700 million Snapchatters have engaged with generative AI Lenses more than 17 billion times”
Imagine Lens engagements: nearly 2 billion (Since September launch through Q4 FY2025, topline)
“Our Imagine Lens launched in September has already been engaged with nearly 2 billion times”
DPA CPA reduction (Pixel Purchase GBBs, internal testing): 55% (7-0); 45% (1-0) (cumulative over past year through Q4 FY2025, topline)
“targeted ranking, format and delivery improvements delivered a 55% reduction in cost per action for 7-0 conversions and 45% reduction in cost per action for 1-0 conversions amongst all Pixel Purchase GBBs, based on cumulative internal testing over the past year.”
App in-app optimization revenue: 89% year-over-year growth (Q4 FY2025, topline)
“Revenue from in-app optimizations grew 89% year-over-year supported by advances in foundational app models”
Spotlight reposts/shares (ranking/trend models): 69% year-over-year increase in the U.S. (Q4 FY2025, topline)
“the number of Spotlight reposts and shares increased 69% year-over-year in the U.S. reflecting our ability to surface timely content at scale.”
New code AI-generated (internal productivity): approximately 40% (Q4 FY2025 (current at time of call), bottomline)
“something like 40% of new code at Snap is AI generated.”
Lens+ subscription ARPU / gross margin: higher ARPU; gross margin expansion (no % given) (Q1 FY2026 early traction, both)
“Early traction has been encouraging with Lens+ contributing to higher subscription ARPU and gross margin expansion.”
PAST (realized)
- Q4 FY2025: More than 700 million Snapchatters engaged with generative AI Lenses more than 17 billion times; Imagine Lens engaged nearly 2 billion times.
- Q4 FY2025: DPA targeted ranking/format/delivery delivered 55% CPA reduction for 7-0 conversions and 45% for 1-0 conversions on Pixel Purchase GBBs (cumulative internal testing over past year).
- Q4 FY2025: App in-app optimization revenue grew 89% YoY supported by advances in foundational app models.
- Q4 FY2025: Q4 ranking and trend detection model enhancements contributed to Spotlight reposts/shares up 69% YoY in the U.S.
- Q1 FY2026: LLM user-intent DPA retrieval improved Pixel Purchase conversions by more than 2%; VLM similar-product retrieval added a high single-digit DPA conversion lift.
- Q1 FY2026: App Re-engagement model upgrade increased purchase conversions ~2% and improved CPA nearly 9%.
- Q1 FY2026: Nearly 70% of ad spend uses at least one AI-powered automation solution (Smart Audience, Smart Budget, or Smart Placement).
- Q1 FY2026: AI-driven personalization contributed to Spotlight total watch time up 11% YoY.
- Q1 FY2026: AI-powered Lens creation drove 400,000+ lenses submitted in Q1, up more than 150% YoY.
- Q1 FY2026: Lens+ contributed to higher subscription ARPU and gross margin expansion.
- Q4 FY2025: ~40% of new code at Snap is AI generated; trust/safety and customer service workflows automated.
CURRENT (now)
- Q1 FY2026: Scaling AI Sponsored Snaps and working with partners like Experian to evolve the product.
- Q1 FY2026: Continuing strategic investments in AI model training and monetization serving costs (infrastructure).
- Q1 FY2026: Lens+ emerging as a subscription tier anchored on AI creative/editing tools in the camera.
- Q4 FY2025: Early testing of smart ads that automatically assemble and iterate creative elements.
- Q4 FY2025: Investing in AI agents to accelerate SMB activation via automated recommendations and onboarding optimizations.
- Q4 FY2025: Embedding AI across the ad platform from creative development through delivery and optimization.
- Q4 FY2025: Running proprietary generative AI camera models on device at scale.
FORWARD (guidance)
- Q1 FY2026: Commercial launch of Specs later in 2026 with AI-powered assistance and agent oversight use cases (Agent Center Lens on developer Specs).
- Q1 FY2026: Lens+ and additional AI-powered premium experiences expected to drive subscription ARPU higher over time.
- Q1 FY2026: Road map for AI Sponsored Snaps: careful expansion of capability, improving demand/yield, more direct response features.
- Q4 FY2025: Specs public launch in 2026 as step in human-centered computing and AR platform evolution.
- Q4 FY2025: Plan to continue heavy ML/AI investment for ads and content personalization while calibrating broader infrastructure to monetization.
- Q4 FY2025: AI coding/agents seen as force multiplier for new apps and faster product invention using Snap distribution and friend graph.
TRACK RECORD — PROMISE vs DELIVERY
74/100 track record mixed 6 calls reviewed
Snap rarely sets hard numeric AI KPIs with deadlines; most judgeable items are milestone timelines (on-device models, faster ML refreshes, Specs 2026) rather than revenue or efficiency targets. Within that narrow set, on-device generative deployment and ranking-speed work show credible follow-through, while near-real-time ranking and the 1B-user goal are only partial and the 2026 Specs launch remains unproven in these calls.
Bring compact on-device generative image model (379M params, ~1.4s on iPhone) into production in coming quarters — promised Q4 FY2024
delivered By Q4 FY2025 management cited proprietary on-device generative camera models running efficiently at scale; Imagine Lens and broad Gen AI lens engagement followed in 2025.
Reach near-real-time content-ranking model refreshes within several quarters — promised Q1 FY2025
partial Q3 FY2025 reported training cycles cut from days to ~2 hours and fresher Spotlight, but management still framed this as progress toward—not achievement of—near-real-time refreshes.
Publicly launch first fully standalone lightweight Specs AR glasses in 2026 — promised Q2 FY2025
too-early Through Q1 FY2026 calls Specs remained pre-launch with consumer debut still planned for 2026 (AWE June 2026 teased); no evidence of a completed public launch in this transcript set.
Begin Q1 testing of Smart Budget Optimization (AI automated budget allocation across assets) — promised Q4 FY2024
partial Q2 FY2025 referenced alpha testing of smart budget alongside Smart Campaign/smart bidding automation; not framed as full GA rollout.
Double pace of fresher ML content-ranking models integrating new trends/signals — promised Q1 FY2025
delivered Same call reported implementation and Spotlight views on posts <24h old doubling YoY; later quarters cited larger models and much faster training cycles.
1 billion global monthly active users (repeated community goal tied to AI/ML investment narrative) — promised Q1 FY2025
partial MAU rose to ~956M by Q1 FY2026 but management pivoted to profitable growth and warned of engagement/DAU pressure; 1B not reached in period covered.
PRICED-IN (REFINED)
LOW (room left)Est. revisions flat · Fwd P/E -17.2 · EV/Sales 2.1x
AI claim maps to Advertising Revenue, Other Revenue
Rating migration is muted (buys ~8–11, holds ~31) with a June dip, and price targets fell recently (lastMonth 7.5 vs lastQuarter 8.25), so revision momentum is not rising. At ~2.1x EV/Sales and negative forward P/E, the stock is not trading at a premium that would embed an AI rerating. AI upside would mainly flow through Advertising Revenue (and secondarily Other Revenue); consensus already models revenue/EPS growth to 2027 but without upward revisions, leaving room if AI improves ad yield faster than estimates.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
5Q4 FY20247Q1 FY20257Q2 FY20258Q3 FY20258Q4 FY20259Q1 FY2026
AI enthusiasm across 6 calls — trend ↗ rising
From sparse ML-infra mentions to gen-AI lenses and quantified ad ML lifts tied to revenue.
RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY
8/10 qualitative impact material near-term · mixed evidence
Where AI matters: advertising yield & AR/camera engagement
~70% of ad spend already runs through AI automation with disclosed conversion/CPA lifts on DPA and app campaigns, and gen-AI lenses are scaling creator supply and Lens+—but every uplift is a sub-segment % with no AI revenue line or dollar bridge to the ~$5.2B ad base.
Caveats: Conversion and engagement lifts are unanchored to incremental revenue on a still-loss-making P&L; AI model training/serving is an explicit infra cost headwind (~$401M Q1) with no quantified opex savings; Attention and discovery could shift to AI-first interfaces faster than Specs/Lens+ monetization matures
AI DISRUPTION / CANNIBALIZATION RISK two-sided · 4/10
AI-native chat, agents, and commoditized generative media can divert time and ad attention away from friend-graph social, yet Snap's near-term economics are still defended by AI-improved ad ROI and on-platform creation tools rather than automation of what it sells.
OPTIONS / MARKET STRUCTURE
option liquidity: good
proxy inputs — dollar-ADV $282M · beta 1.055 · px $5.67
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 8/10 committed.
INSIDERS selling 34 open-market sell(s) vs 0 buy(s) — net distribution
INSTITUTIONS (13F) adding as of 2026-03-31: 104 new / 117 closed positions; 304 increased / 134 reduced; institutional ownership -7.11pp; -16 net 13F holders
MGMT LANGUAGE 8/10 committed Firm Q1 AI ad and Lens metrics; softer on newer Chat AI formats and eyewear.
commit “Nearly 70% of advertising spend now use at least one of our AI-powered automation solutions”
commit “which improved Pixel Purchase conversions by more than 2%”
commit “AI-powered Lens creation is transforming our AR ecosystem, with more than 400,000 lenses submitted in Q1”
VERBATIM AI QUOTES
“These efforts, combined with continued investment in AI-driven personalization, resulted in higher engagement with total time spent watching Spotlight increasing 11% year-over-year.”
— Evan Spiegel, Q1 FY2026
“AI-powered Lens creation is transforming our AR ecosystem, with more than 400,000 lenses submitted in Q1, increasing more than 150% year-over-year.”
— Evan Spiegel, Q1 FY2026
“These gains are being driven by continued progress in AI, ranking, retrieval, and automation across the ad platform.”
— Evan Spiegel, Q1 FY2026
“Nearly 70% of advertising spend now use at least one of our AI-powered automation solutions including Smart Audience, Smart Budget or Smart Placement, which gives us confidence that these improvements are benefiting a broad share of the business.”
— Evan Spiegel, Q1 FY2026
“In Q1, we launched LLM based user intent understanding for Dynamic Product Ads retrieval, which improved Pixel Purchase conversions by more than 2%, and multimodal similar product retrieval using a vision language model fine-tuned on Snap data, which delivered an additional high single-digit lift in DPA purchase conversions.”
— Evan Spiegel, Q1 FY2026
“We also upgraded our App Re-engagement model with stronger foundational user beddings and a new multi-task architecture, increasing purchase conversions by approximately 2%, while improving CPA by nearly 9%.”
— Evan Spiegel, Q1 FY2026
“Building on this momentum, we introduced AI Sponsored Snaps, a new format that enables brands to engage Snapchatters through interactive, AI-powered conversations in Chat and extends our strategy of delivering more personalized, high-intent advertising experiences.”
— Evan Spiegel, Q1 FY2026
“including Lens+, AI-powered Lens interactions or deepening user engagement and increasingly serving as a natural discovery layer for premium AI-powered experiences.”
— Evan Spiegel, Q1 FY2026
“new experiences that expand what is possible on Specs across learning, gaming, utility and AI-powered assistance.”
— Evan Spiegel, Q1 FY2026
“driven primarily by community growth, strategic investments in AI model training and monetization serving costs.”
— Derek Andersen, Q1 FY2026
“The second will be making it easier and more performant for advertisers to connect with Snapchatters by leveraging AI tooling and capabilities end-to-end through our ad platform, including creative development, campaign setup and performance optimization.”
— Evan Spiegel, Q4 FY2025
“We're enhancing our camera with AI-powered capabilities that make creation more intuitive, dynamic and social. Recent breakthroughs in our proprietary models allow us to deliver high-quality generative AI camera experiences efficiently at scale by running our models on device.”
— Evan Spiegel, Q4 FY2025
“More than 700 million Snapchatters have engaged with generative AI Lenses more than 17 billion times, often discovering and sharing these lenses through conversations with friends and family.”
— Evan Spiegel, Q4 FY2025
“Our Imagine Lens launched in September has already been engaged with nearly 2 billion times, highlighting strong early traction and repeat usage.”
— Evan Spiegel, Q4 FY2025
“In Q4, enhancements to our ranking and trend detection models contributed to improved content freshness and engagement.”
— Evan Spiegel, Q4 FY2025
“We also began early testing of smart ads, which automatically assemble and iterate creative elements to identify the highest performing combinations.”
— Evan Spiegel, Q4 FY2025
“In addition, we are investing in AI agents designed to accelerate SMB activation through automated recommendations and onboarding optimizations that reduce decision friction and improved performance.”
— Evan Spiegel, Q4 FY2025
“Obviously, I think now, something like 40% of new code at Snap is AI generated.”
— Evan Spiegel, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Maggie (for Doug Anmuth, JPMorgan)): I was wondering if you can just talk a bit more about the broader opportunity you see with AI sponsored Snaps and sort of what you're hearing from advertisers in terms of overall interest and something like this?
A: Advertiser feedback has been really positive. I think mostly because sponsored Snaps are showing that chat can be monetized in a way that's really native to Snapchat. And rents are loving this combination of very high reach and high attention especially at such a high frequency behavior on Snapchat throughout the day. So I think what's important for us strategically is that that's not just another inventory pool. It really gives us a differentiated environment where brands can engage users in more direct and more personal way. And AI sponsored Snaps are really an extension of that because they can make those interactions and those conversations more useful and more relevant over time. So I think just looking forward, the road map is really about just careful expansion of capability improving demand in yield, obviously, it sponsored Snaps, adding more direct response features and then continuing to work with new partners like Experian to evolve the AI Sponsored Snap product.
Q (Q1 FY2026, Eric Sheridan (Goldman Sachs)): how should we be thinking about agentic AI as sort of an interface for Specs over the maybe in for long term as well?
A: the way that people are using their computers is changing really dramatically, and I think that, that's going to be evident in the adoption of wearables and the adoption of Specs over time because people are going to spend less time hunched over their computers or their phones, typing away on keyboards and spend more time supervising agents who are doing that work. on their behalf. So we actually had a team member who with AI built out a pretty cool Lens called Agent Center, where you can oversee and manage our agents through Specs with the current developer version of the glasses, which is a pretty cool way to stay on top of what your agents are getting done for you without caring your laptop around. So I think a lot of opportunity there and the way that people are using computing is changing so fundamentally in so many ways at this moment.
Q (Q1 FY2026, Michael Nathanson (MoffettNathanson)): on SnapChat+, it's impressive what you guys are doing. Is there any color you could give us on who these users are where they're coming from and kind of just the pricing dynamics...
A: we've seen some strong momentum with Lens+, which is really anchored around our AI creative and editing tools in the camera. And that, I think, could be a big opportunity for us long term. It's obviously a higher price offering.
Q (Q1 FY2026, Nitin Bansal (Bank of America)): You mentioned Memories and Lens+ are like the key drivers behind the recent strength. Can you talk about like how sustainable do you believe the current growth trajectory is over the medium term?
A: Lens+, I already mentioned, but I think it's a really meaningful opportunity just given the excitement and momentum we see around AI creative tools and AI editing for images and videos. So we certainly think that, that could help drive ARPU higher over time.
Q (Q4 FY2025, Justin Patterson (KeyBanc)): I wanted to talk about agentive coding. We've seen more companies see meaningful improvements in engineering productivity from these tools. How is this thing deployed at Snap today? And how should we think about potential benefits, whether it's product velocity more engagement on the platform, more monetization opportunities or expense efficiency.
A: Obviously, I think now, something like 40% of new code at Snap is AI generated. We made a ton of headway with trust and safety and customer service. in terms of automating those workflows. I think there's a lot of opportunity for the sales workflow as well to empower our sales team but also to automate quite a bit of that. So certainly, we're seeing gains across the board and how we're operating our business today. I also think this can be a real accelerant for our own creativity... we're just running as fast as we can to roll out new agents across the enterprise new tools. And it's especially for a small team, like the one we've got at Snap, this is just a massive force multiplier.
Q (Q4 FY2025, Benjamin Black (Deutsche Bank)): Can you talk about the decision to moderate infrastructure spending at a time when others are ramping spend to drive ad performance?
A: the big driver in the ramp of infrastructure investment over the last couple of years has been a really significant growth in our ML and AI investment, and that was to both support the rebuild of the Ad Platform and the DR advertising business. and also to support the content business and banking and personalization and all of the work that we've done there. And I think I would say, first and foremost, we intend to continue to invest pretty heavily there. And so that's not an area of focus for pulling back.