← back to rankingCART · Instacart (Maplebear Inc.)
Specialty Retail · mkt cap $9.6B · calls: Q1 FY2026 vs Q4 FY2025
55.0 conviction · conf-adj 52
conf 3/10 partial
enthusiasm:27.0 · trend:8 · quantifies:5 · impact:0 · under_radar:5 · credibility:5 · business_impact:8 · disruption:-6 · commitment:0 · confirmation:3
Enthusiasm latest 9 / prev 8 (rising)
Instacart frames AI as a compounding moat: proprietary grocery data plus in-store operations (Caper, Store View/computer vision) and fulfillment power consumer agentic shopping (Cart Assistant), retailer white-label AI solutions, and ads ranking/generative recommendations—with distribution via ChatGPT, Claude, and other agents. Enthusiasm intensified from Q4's internal productivity claims and net-gainer thesis to Q1's broader product rollout (25% Cart Assistant test, enterprise AI signings, generative rec system). Credibility is mixed: Q4 quantified engineering productivity and a ~1pt Caper basket lift, but most consumer/agentic and ads-AI impacts are qualitative or early data without revenue or margin attribution; management says off-platform agent traffic is immaterial and AI consumer costs are monitored quarter-by-quarter.
GROUNDED NEXT-FY IMPACT vs CONSENSUS
Grounded on actual base — revenue $3.7B · net income $0.4B · net margin 11.9% · diluted EPS 1.6
These are next-fiscal-year annual uplift estimates, not next-quarter numbers.
Aggregate next-FY est. rev uplift: 0.0% · next-FY EPS uplift: 0.0% · vs analysts: inline · priced in: medium (model's call-read: high; verdict above is the hard-data one used for ranking) · confidence: 3/10
| Claim | Figure | Arithmetic | Next-FY Rev % | Next-FY EPS % |
|---|
Engineer output +~40% (10% of team +80%)
productivity · soft | nearly 40% increase in average output per engineer; 10% of team up 80% | Productivity ratio only; no engineering headcount, payroll, or R&D/opex $ disclosed in any claim. Realized over the past year (through Q4 FY25), so largely embedded in FY25 base — not incremental FY26 $ without an opex anchor. Cannot convert % output to $ saving or capacity revenue → rev_uplift_pct and eps_uplift_pct null. | | |
Build production software >4x faster (new projects)
productivity · soft | >4x faster | Velocity multiplier on new projects only; no disclosed eng-spend base (new vs maintenance), project count, or capex/opex. 4x speed ≠ 75% cost cut (capacity reallocation). Overlaps with the +40% output claim. No standalone $ mapping → null. | | |
Caper in-cart 'got everything you need' prompt: ~1pp basket lift
engagement · soft | nearly 1pp avg basket size on Caper-cart baskets with prompt | Segment lift applies ONLY to baskets through Caper smart carts (in-store, 100+ cities) — an undisclosed sliver of GMV. Company rev_uplift_pct = 1.0 × Caper_rev_share, and Caper share not disclosed. Anchored % but no obtainable base → unsizeable; null. | | |
Cart Assistant available to ~25% of US customers
engagement · soft | ~25% reach | Rollout reach only; no conversion, basket, frequency, or retention lift quantified. Reach ≠ revenue. Cannot map 25% to $ without adoption/lift rates → null. | | |
Search users ~5x more likely to place first order
engagement · soft | ~5x first-order likelihood | Observed odds ratio (search vs non-search), not a causal AI increment, with no denominator (search penetration, baseline first-order rate, incremental orders) disclosed. Cannot derive incremental orders or revenue → null. | | |
Caper live in 100+ cities
engagement · soft | >100 cities | Geographic scale/footprint metric only; no carts deployed, GTV, or revenue per city disclosed. Cannot convert city count to $ uplift → null. | | |
Off-platform agentic referral traffic 'not material'
revenue | not material | Mgmt explicit: off-platform/agentic referral revenue immaterial today → incremental rev ≈ $0. $0 / $3,742M rev = 0.0% rev_uplift_pct. Flow-through: $0 incremental NI / $476M consensus FY25 adjusted NI = 0.0% eps_uplift_pct. Only financially-anchored impact statement, and it anchors at zero. | 0.0 | 0.0 |
Assumptions: Next-FY scope = FY2026. Tax rate 21%; default incremental net margin = current ~11.9% (never used — no claim was $-sizeable, no AI software/services claim to justify higher). EPS/NI denominator = consensus adjusted FY25 NI $476M (EPS ~$1.81), not depressed GAAP $447M / $1.60, per earnings-basis rule (FY26 consensus NI $612M). Engineer productivity (+40%) and 4x build speed are backward-looking/capacity metrics with no $ base — treated as largely captured in FY25, not re-counted as FY26 incremental. No bookings figures claimed. Segment mapping: Caper basket lift → undisclosed Caper in-store GMV; engineer productivity → undisclosed R&D/opex. Adopter-side only; no quantified supplier AI revenue.
Top line: Hard quantified topline AI impact sums to ~0% of FY25 revenue ($3.742B). The only financially-anchored topline statement — off-platform agentic referral traffic — is explicitly 'not material' (~$0). Caper +1pp basket lift is real but segment-local and cannot scale without undisclosed Caper mix; Cart Assistant (25% US reach) and search (5x first-order likelihood) are reach/correlation stats, not anchored incremental $. Consensus FY26 revenue $4.194B (+12.1% vs $3.742B) is not exceeded by any hard claim math. Value is forward optionality, not a sizeable FY26 number.
Bottom line: No bottom-line $ computable from disclosed claims. The headline EPS lever — engineer output +~40% (10% of team +80%) and >4x faster builds — is a productivity ratio with no disclosed R&D/engineering cost base, so it cannot be converted to an after-tax saving; gains also appear realized over the past year and likely already in FY25 op margin ($498M OI on $3.742B rev). Net margin is healthy (~11.9%), so no thin-denominator artifact — the issue is absence of a base, not distortion. Hard aggregate EPS uplift = 0.0% vs consensus FY25 adjusted NI $476M.
Aggregate hard adopter uplift: 0.0% rev / 0.0% EPS vs FY25 bases ($3.742B rev, $476M consensus NI). Consensus already models FY26 rev $4.194B (+12.1%) and EPS ~$2.43 (+34% vs FY25 $1.81; NI $612M, ~3x revenue-to-EPS leverage) — the same efficiency lever the +40% engineer-output claim represents, so the AI productivity story appears absorbed into the existing margin ramp, not additive. No quantified claim bridges to incremental $ above that trajectory, and the one anchored item is 'not material.' No measurable consensus gap from hard math; AI optionality (ads, enterprise AI, agentic checkout) is forward/unquantified.
MODEL CONSENSUS (impact)
partial
Near-total agreement: 0% rev/EPS hard uplift, adopter-side, all engagement/productivity claims unsizeable, only anchored item ('not material') is zero. Differences were verdict labels.
Conflicts reconciled
- vs_analyst_expectations: X=unclear vs Y=inline -> used inline because both compute a 0% hard gap, which 'inline' describes more precisely than 'unclear'
- priced_in: X=medium vs Y=high -> used high (more conservative + Y's reasoning that consensus already embeds the productivity leverage is sound)
| Field | Opus 4.8 | GPT-5.5 |
|---|
| Rev uplift % | 0.0 | – |
| EPS uplift % | 0.0 | – |
| Priced in | medium | – |
| vs analysts | unclear | – |
| Confidence | 3 | – |
| Top line | Effectively unquantifiable / immaterial in the next fiscal year from disclosed claims. The only financially-anchored topline statement — off-platform agentic referral traffic — is explicitly 'not material' (~0% of $3,742M). The genuinely positive signals (Caper +1pp basket lift, search 5x first-order, Cart Assistant at 25% reach) are real engagement wins but apply to undisclosed, small bases (Caper in-store GMV; new-user search cohort), so none can be sized into incremental revenue. Aggregate adopter-side topline uplift ≈ 0% on the demonstrable math; the value is optionality (forward-looking 'meaningful growth driver over time'), not a FY26 number. | – |
| Bottom line | No sizeable bottom-line uplift can be computed. The headline EPS lever — engineer output +~40% (with 10% of the team +80%) and >4x faster builds — is a productivity ratio with no disclosed R&D/engineering cost base, so it cannot be converted to an after-tax saving (saving_$ * 0.79). Critically, consensus already models EPS +34% (2.43 vs 1.81) on revenue +12.1%, i.e. ~3x operating leverage — exactly where these productivity gains would land — so any benefit is plausibly already inside consensus margins rather than incremental to them. Net margin is healthy (~11.9%) so there is no thin-denominator artifact; the issue is purely absence of a base, not a distorted one. | – |
| Reasoning | Consensus FY26 implies rev +12.1% (4,194M vs 3,742M) and net income +37% (612M vs 447M) / EPS +34% (2.43 vs 1.81). That ~3x revenue-to-EPS leverage already assumes substantial efficiency gains — the same lever the +40% engineer-output claim represents — so the AI productivity story appears absorbed into the existing margin ramp, not additive. No quantified claim points clearly ABOVE this trajectory (none is even dollar-sizeable), and the one financially-anchored item is 'not material.' Hence no demonstrable gap vs consensus either way. | – |
Rows highlighted where the two models disagreed.
QUANTIFICATIONS
Engineer output (AI-accelerated internal execution): nearly 40% increase in average output per engineer; 10% of team up 80% (over the past year (through Q4 FY2025), bottomline)
“Over the past year, average output per engineer is up nearly 40%, which includes 10% of our team increasing output by 80%.”
Software build speed for new projects (AI-enabled): more than 4x faster (for new projects (as of Q4 FY2025), bottomline)
“for new projects, we believe AI is now enabling us to build production-grade software more than 4x faster than before.”
Caper in-cart prompt basket lift (AI-powered smart cart): nearly 1 percentage point lift in basket size on average (early engagement (Q4 FY2025), topline)
“a simple got everything you need prompt is driving a nearly 1 percentage point lift in basket size on average.”
Cart Assistant rollout reach: about 25% of U.S. customers (Q1 FY2026 (testing), topline)
“now available to about 25% of U.S. customers.”
Search → first-order conversion: about 5x more likely to place their first order (Q1 FY2026, topline)
“customers who use search are about 5x more likely to place their first order.”
Caper geographic scale: more than 100 cities (Q1 FY2026, topline)
“Caper, our AI-powered smart cart, is now live in more than 100 cities”
Off-platform agentic referral traffic: not material (Q4 FY2025, topline)
“If we look at some of the referral traffic that we get from outside platforms, it's all very -- it's all kind of like not material at this point in time.”
PAST (realized)
- Q4 FY2025 — Chris Rogers: Over the past year, average output per engineer is up nearly 40%, which includes 10% of our team increasing output by 80%.
- Q4 FY2025 — Chris Rogers: for new projects, we believe AI is now enabling us to build production-grade software more than 4x faster than before.
- Q4 FY2025 — Chris Rogers: System reliability is up even as engineering throughput has increased significantly.
- Q4 FY2025 — Chris Rogers: on our enterprise platform, we're onboarding more retailers faster while delivering more customized white glove solutions at scale, which was not possible before.
- Q4 FY2025 — Chris Rogers: a simple got everything you need prompt is driving a nearly 1 percentage point lift in basket size on average.
- Q4 FY2025 — Chris Rogers: we were the first grocery partner to launch native checkout directly on ChatGPT
- Q1 FY2026 — Chris Rogers: we continue to enhance our search functionality, making it faster and more relevant
- Q1 FY2026 — Chris Rogers: we've upgraded our AI-powered replacement flow to better reflect consumer preferences in real time
- Q1 FY2026 — Chris Rogers: we recently launched a new generative recommendation system
- Q1 FY2026 — Emily Maher: I think about 2025 as a point in time where we were more experimenting and seeing where we were seeing adoption
CURRENT (now)
- Q1 FY2026 — Chris Rogers: We recently began testing Cart Assistant, our AI-powered conversational shopping experience, now available to about 25% of U.S. customers.
- Q1 FY2026 — Chris Rogers: we're working with partners like Kroger and Sprouts to bring these capabilities to life
- Q1 FY2026 — Chris Rogers: we've decided to integrate Instacart with AI platforms like ChatGPT and most recently with Claude.
- Q1 FY2026 — Chris Rogers: Caper, our AI-powered smart cart, is now live in more than 100 cities
- Q1 FY2026 — Chris Rogers: we've begun piloting Store View with partners like McKeevers, Sprouts and more, who are leveraging real-time computer vision
- Q1 FY2026 — Chris Rogers: AI continues to meaningfully accelerate our progress.
- Q1 FY2026 — Emily Maher: we are looking for metrics around ways to, again, offset those costs through things like better customer engagement, better conversion, et cetera. So I think it's still early, something we're monitoring and adapting to really quarter-by-quarter.
- Q4 FY2025 — Chris Rogers: we're now getting ready to launch AI solutions starting with Cart Assistant.
- Q4 FY2025 — Chris Rogers: We're in beta testing now
- Q4 FY2025 — Chris Rogers: If we look at some of the referral traffic that we get from outside platforms, it's all very -- it's all kind of like not material at this point in time.
FORWARD (guidance)
- Q1 FY2026 — Chris Rogers: Over time, we expect this to translate into additional benefits across our ecosystem, including better availability, better recommendations, more efficient fulfillment and more valuable advertising for retailers and brands.
- Q1 FY2026 — Chris Rogers: we do think AI overall is going to be a meaningful growth driver for the category over time.
- Q1 FY2026 — Chris Rogers: that should accelerate online adoption by driving better conversion and higher retention and larger baskets and more frequent ordering.
- Q1 FY2026 — Chris Rogers: we're moving from more of a front-end AI experience to actually -- experience that will actually help you complete your order
- Q1 FY2026 — Chris Rogers: I firmly believe is going to be one of the most interesting advertising opportunities for brands in-store.
- Q1 FY2026 — Chris Rogers: building leading-edge AI solutions
- Q4 FY2025 — Chris Rogers: we think we will excel and be a net gainer in an AI-driven world.
- Q4 FY2025 — Chris Rogers: we have plans to roll out on Instacart marketplace by the end of Q1.
- Q4 FY2025 — Chris Rogers: we expect every generative AI company will connect into our grocery engine to drive demand for our retailers.
- Q4 FY2025 — Chris Rogers: we think if we nail that there will be lots of opportunities to monetize that down the road.
- Q4 FY2025 — Chris Rogers: That intersection, we think, long term, is going to be a benefit to our business.
TRACK RECORD — PROMISE vs DELIVERY
58/100 track record mixed 6 calls reviewed
Instacart rarely sets forward-looking, dated AI targets on earnings calls; management mostly cites retrospective productivity and pilot metrics. Delivery is credible on shipped product milestones (Cart Assistant, engineer throughput), but several quantified AI KPIs were dropped from the narrative without follow-up, and Caper impact claims were scaled back in later disclosures.
Caper pilot deployments driving double-digit basket size increases to justify broader rollout — promised Q4 FY2024
partial Later calls still cite Caper basket benefits but never reaffirm double-digit lifts; Q4 FY2025 highlighted only ~1pp basket lift from a specific in-cart prompt.
87% of Q1 code developed with AI assistance as an operating benchmark — promised Q1 FY2025
quietly-dropped Q2 FY2025 reported 80%+ AI-assisted deployed code, but the metric vanished from Q3 FY2025 onward with no update or explanation.
30% YoY increase in average engineer merges from AI-assisted development — promised Q2 FY2025
delivered Q4 FY2025 reported average engineer output up nearly 40% over the past year, beating the earlier productivity claim.
AI automation tripled sales outreach to priority accounts and doubled meetings booked — promised Q2 FY2025
quietly-dropped No later earnings call restated or updated these AI-driven sales productivity metrics.
Caper Carts in nearly 20% of Wakefern stores soon — promised Q3 FY2025
partial Q1 FY2026 cited continued Wakefern Caper expansion but did not confirm reaching ~20% store coverage.
Launch Cart Assistant generative AI with enterprise partners (e.g., Sprouts) — promised Q4 FY2025
delivered Q1 FY2026 confirmed Cart Assistant testing with ~25% of U.S. customers and live rollouts with Kroger, Sprouts and additional signed partners.
PRICED-IN (REFINED)
MEDIUMEst. revisions rising · Fwd P/E 22.1 · EV/Sales 2.3x
AI claim maps to Advertising And Other, Transaction
Over the past six months, bullish ratings rose (buy+strongBuy 16→18) while holds fell (15→12), and price targets stepped up (lastYear $50.74 → lastQuarter $53.20 → lastMonth $54.67), with consensus EPS climbing from $1.81 (FY25) to $2.43 (FY26) and $2.98 (FY27)—so estimate-revision momentum is already positive. At ~22x forward P/E and ~2.3x EV/sales the stock is not extreme for a profitable grower (PEG ~0.87, price ~27% below avg PT), so multiples alone do not scream fully priced-in. AI-driven upside (ads/retail media, ops efficiency) maps mainly to Advertising And Other and Transaction; ads growth is already visible in the mix (~$1.07B, +11% YoY), so the thesis is partly in the numbers but not fully stretched—medium.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
5Q4 FY20247Q1 FY20258Q2 FY20258Q3 FY20258Q4 FY20259Q1 FY2026
AI enthusiasm across 6 calls — trend ↗ rising
From ad ML only to AI-first engineering, fulfillment models, retailer AI suite, and live agentic consumer products.
RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY
7/10 qualitative impact material medium-term · mixed evidence
Where AI matters: consumer conversion, retail media/ads, fulfillment ops, engineering productivity
Real deployments—Cart Assistant (~25% US), Caper/Store View, semantic ads/recs, and quantified eng throughput (+~40% output, >4x new-build speed)—target core order completion and the ~$1B ads line, but mgmt attributes almost no incremental revenue (off-platform agent traffic explicitly not material) and no P&L bridge beyond a ~1pp Caper basket lift on an undisclosed GMV slice.
Caveats: No disclosed conversion, basket, frequency, or ads $ lift from Cart Assistant despite 25% reach; Hard quantified topline AI uplift ~0% of revenue; agent referral traffic explicitly not material; Engineering productivity gains likely already embedded in FY25 margins, not incremental FY26 EPS; Caper ~1pp basket lift is segment-local with undisclosed revenue mix
AI DISRUPTION / CANNIBALIZATION RISK two-sided · 4/10
Agentic interfaces (ChatGPT/Claude) can commoditize discovery and route orders among Instacart, DoorDash, and Uber Eats, but grocery’s physical fulfillment, retailer integrations, and proprietary catalog/substitution data make the completion layer harder to disintermediate than a pure software marketplace.
OPTIONS / MARKET STRUCTURE
option liquidity: fair
proxy inputs — dollar-ADV $161M · beta 0.965 · px $40.13
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 6/10 measured.
INSIDERS selling 11 open-market sell(s) vs 0 buy(s) — net distribution
INSTITUTIONS (13F) adding as of 2026-03-31: 94 new / 82 closed positions; 263 increased / 137 reduced; institutional ownership +1.62pp; +11 net 13F holders
MGMT LANGUAGE 6/10 measured Real shipped AI with 25%/100-city metrics; balanced by testing, pilots, and over-time expectation language.
commit “We recently began testing Cart Assistant, our AI-powered conversational shopping experience, now available to about 25% of U.S. customers.”
commit “early data shows that this is driving higher engagement and better results for advertisers.”
commit “Caper, our AI-powered smart cart, is now live in more than 100 cities”
VERBATIM AI QUOTES
“we're increasingly using AI to make our experience more personalized and intuitive.”
— Chris Rogers, Q1 FY2026
“we've upgraded our AI-powered replacement flow to better reflect consumer preferences in real time, a strong example of how data and technology have come together to improve outcomes for both shoppers and customers.”
— Chris Rogers, Q1 FY2026
“we're using that to build the gold standard of agentic grocery AI.”
— Chris Rogers, Q1 FY2026
“We recently began testing Cart Assistant, our AI-powered conversational shopping experience, now available to about 25% of U.S. customers.”
— Chris Rogers, Q1 FY2026
“Early feedback is encouraging with customers using it to discover recipes, build meal plans, quickly assemble baskets and research products.”
— Chris Rogers, Q1 FY2026
“we've decided to integrate Instacart with AI platforms like ChatGPT and most recently with Claude.”
— Chris Rogers, Q1 FY2026
“Retailers choose Instacart because of our purpose-built grocery technology across e-commerce, retail media, in-store and AI solutions.”
— Chris Rogers, Q1 FY2026
“we're extending that approach with our newest enterprise offering, AI solutions. As we build leading AI capabilities on our marketplace, we're bringing those same tools to retailers' owned and operated channels.”
— Chris Rogers, Q1 FY2026
“we recently launched a new generative recommendation system that can use real-time context to better understand the consumer's intent.”
— Chris Rogers, Q1 FY2026
“early data shows that this is driving higher engagement and better results for advertisers.”
— Chris Rogers, Q1 FY2026
“our AI-powered recommendation engine in our self-service platform, Ads Manager, continues to gain traction.”
— Chris Rogers, Q1 FY2026
“Caper, our AI-powered smart cart, is now live in more than 100 cities”
— Chris Rogers, Q1 FY2026
“we've begun piloting Store View with partners like McKeevers, Sprouts and more, who are leveraging real-time computer vision to improve shelf availability and accuracy.”
— Chris Rogers, Q1 FY2026
“AI continues to meaningfully accelerate our progress. It's allowing us to onboard retailers faster. It's allowing us to customize for retailers faster.”
— Chris Rogers, Q1 FY2026
“the innovation around ads is going to come from AI and everything related to how AI can advance advertising mediums and platforms.”
— Chris Rogers, Q1 FY2026
“AI is completely core to our advertising efforts at this point”
— Chris Rogers, Q1 FY2026
“we're building the gold standard of agentic experiences using all of our proprietary data directly on Instacart and on our retailer partner side.”
— Chris Rogers, Q1 FY2026
“we do think AI overall is going to be a meaningful growth driver for the category over time.”
— Chris Rogers, Q1 FY2026
“I think about 2025 as a point in time where we were more experimenting and seeing where we were seeing adoption and allowing people to try to adopt AI across their workflows and then, of course, across the consumer experience as well.”
— Emily Maher, Q1 FY2026
“And with generative AI accelerating execution across our platform, we are increasing our velocity, compounding our advantages and driving greater efficiency, all while strengthening the value of our first-party data.”
— Chris Rogers, Q4 FY2025
“we believe these shifts favor platforms like Instacart, that combine technology with real-world operations and unique data at scale.”
— Chris Rogers, Q4 FY2025
“we think we will excel and be a net gainer in an AI-driven world.”
— Chris Rogers, Q4 FY2025
“our physical operations make our data better, and that data makes our technology smarter, more unique and more effective”
— Chris Rogers, Q4 FY2025
“Over the past year, average output per engineer is up nearly 40%, which includes 10% of our team increasing output by 80%.”
— Chris Rogers, Q4 FY2025
“for new projects, we believe AI is now enabling us to build production-grade software more than 4x faster than before.”
— Chris Rogers, Q4 FY2025
“on our enterprise platform, we're onboarding more retailers faster while delivering more customized white glove solutions at scale, which was not possible before.”
— Chris Rogers, Q4 FY2025
“our white label AI assistant, known as Cart Assistant”
— Chris Rogers, Q4 FY2025
“building physical AI capabilities in-store with Caper Carts and Store View”
— Chris Rogers, Q4 FY2025
“we're now getting ready to launch AI solutions starting with Cart Assistant.”
— Chris Rogers, Q4 FY2025
“a simple got everything you need prompt is driving a nearly 1 percentage point lift in basket size on average.”
— Chris Rogers, Q4 FY2025
“It is fair to say that we are using AI across the board to accelerate and improve all aspects of our business.”
— Chris Rogers, Q4 FY2025
“we're well positioned to be the clear winner with AI.”
— Chris Rogers, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Shweta Khajuria (Wolfe Research)): Opportunity and/or risk with agentic AI platforms where a consumer orders delivery and the AI chooses Instacart vs DoorDash vs Uber Eats, potentially risking organic traffic — how do you view this and what is your pushback?
A: Strategy: be wherever customers are while maintaining control of experience and data; build gold-standard agentic experiences on Instacart and retailer channels with proprietary data. Third-party integrations (OpenAI, Anthropic) are early but viewed as incremental demand in an underpenetrated category; co-created experiences should accelerate online grocery adoption where Instacart wins as category leader. Data shared in a controlled way to limit disintermediation risk.
Q (Q1 FY2026, Josh Beck (Raymond James)): Cart Assistant at ~25% of customers — early learnings on conversion, basket size, ad monetization? Is search moving from keyword index to something more LLM-driven?
A: Cart Assistant still early; users save time via recipes, meal planning, product research; studying engagement across the journey (e.g., gluten checks). Believes AI will drive category growth via simpler shopping → better conversion, retention, baskets, frequency; unique data + fulfillment + retailer integrations enable completing orders, not just front-end chat. On search: rich catalog data drives relevance; customers who search are ~5x more likely to place first order; will keep investing in search relevance — did not confirm LLM-based search architecture.
Q (Q1 FY2026, Justin Patterson (KeyBanc)): Guardrails around AI costs / token consumption; how advances in recommendation and ranking benefit conversion and ads.
A: Emily: 2025 was experimentation; monitoring adoption real-time, offsetting workflow costs where possible; consumer AI still early — looking for engagement/conversion offsets, adapting quarter-by-quarter. Chris: AI enables semantic-intelligence recommendations that help consumers complete perfect orders and benefit advertising.
Q (Q1 FY2026, Eric Sheridan (Goldman Sachs)): Most interesting advertising growth opportunities over 12–18 months (AI-adjacent in answer).
A: Chris: ad innovation will come from AI — ranking/relevance/personalization for sponsored products; AI-powered advertiser tooling and recommendations; third pillar is conversational/agentic commerce with ads informed by agentic shopping engagement, incorporating online, in-store, and Caper learnings.
Q (Q4 FY2025, Josh Beck (Raymond James)): Tailwinds from agentic experiences on-platform (easing large baskets) vs economics off-platform (larger customer pool).
A: Building best agentic experiences on Instacart with first-party data from 1.6B lifetime orders. Also partnering off-platform (OpenAI, Google, Microsoft) to be wherever customers shop; OpenAI example was first grocery native checkout on ChatGPT — channel to attract incremental demand and transact on Instacart.
Q (Q4 FY2025, Andrew Boone (Citizens)): ChatGPT Instant Checkout live — advertising intensity on those orders; evolution of retail media as agent channels mature.
A: Priority is discoverability and incredible experience wherever customers shop; monetization later if execution is nailed. Ads team innovates alongside agentic consumer experiences; ads informed by agentic engagement. Using AI for ranking, relevance, personalization, and AI-powered advertiser recommendations (e.g., ROAS headroom, new-to-brand coverage).
Q (Q4 FY2025, Deepak Mathivanan (Cantor Fitzgerald)): Meaningful change in grocery shopping behavior from Cart Assistant / ChatGPT integrations? When will cart assistance be more broadly available? Competition where Uber/DoorDash launched.
A: Agentic behavior will shift over time but still very early; outside referral traffic not material. Investing for future trends; Cart Assistant built on proprietary data — goal is best/most relevant experience, extending to Sprouts/Kroger; beta testing, rollout on marketplace by end of Q1. On DoorDash/Kroger: grew despite launch; non-exclusive retailers steady state; other platforms plateau with small baskets.