The Era of Agentic Commerce: AI Shopping Agents Expected to Re-accelerate Global eCommerce
AI summary card
The Era of Agentic Commerce: AI Shopping Agents Expected to Re-accelerate Global eCommerce
Morgan Stanley believes AI agents will evolve from assisted search to automated purchasing, materially influencing more than 20% of eCommerce GMV over the next five years and driving global eCommerce to approximately US$7 trillion by 2030.
- The report forecasts global eCommerce GMV to reach approximately US$7 trillion by 2030, implying a 2025-2030E CAGR of about +9%, above the +7% seen in 2021-2025.
- Under the five-year base case, about two-thirds of consumers will use agentic commerce tools, and more than 20% of industry GMV will be materially affected by AI search, discovery, or execution tasks.
- The base-case five-year TAM uplift from agentic commerce is approximately 6.2%, with a bear-to-bull range of about 1.4% to 14.2%.
- Large eCommerce platforms are viewed not as passive victims of disruption, but as more likely to control the shopping journey through inventory, fulfillment, customer relationships, ecosystem services, and proprietary AI tools.
- The report's selected core OW beneficiaries include Amazon, Walmart, Alibaba, MercadoLibre, Sea, and Allegro, with additional beneficiaries including Wayfair, Shopify, Chewy, Target, eBay, Meituan, Coupang, LY Corp, Mercari, FSN E-Commerce, and Temple & Webster.
Report interpretation
Overview
This report discusses the impact of agentic commerce on global online retail. Morgan Stanley believes AI-driven shopping agents are moving from search suggestions and product comparison toward cart building, checkout initiation, and even agent-to-agent negotiation. This trend could reduce friction in online purchasing, improve conversion, repeat purchases, and eCommerce penetration, thereby becoming a new catalyst for long-term global eCommerce growth.
Core views
The core view is that market concerns about horizontal LLMs disrupting eCommerce platforms may be excessive. The report argues that large eCommerce platforms still control key links such as inventory, fulfillment, after-sales service, payments, memberships, and retail media, and are more inclined to develop on-platform agents and selective discovery partnerships rather than fully surrender catalogs, checkout, and customer relationships. Early disclosures show Amazon Rufus users are 60% more likely to complete a purchase, Walmart Sparky users have approximately 35% higher average order values, and Lowe's, Sea, Wayfair, and Zalando have also disclosed improvements in conversion, order behavior, or add-to-cart activity.
Analysis framework
The report uses a global eCommerce industry model, AlphaWise consumer surveys, company disclosures, incremental scenario analysis, LLM cost-benefit analysis, and Morgan Stanley's agentic commerce scorecard to assess industry TAM, GMV impact, company relative positioning, and stock beneficiary potential.
Methodology notes
Global eCommerce market size forecast
Using regional and industry models to forecast online GMV and retail penetration, the report expects online retail penetration to rise from 22% in 2025 to 27% in 2030, corresponding to an eCommerce market of approximately US$7 trillion in 2030.
Agentic commerce adoption and incrementality
AI shopping penetration, AI-affected GMV, agent-executed GMV, autonomous GMV, and different layers of incrementality are multiplied to form bear, base, and bull TAM uplift ranges.
Inventory, Infrastructure, Innovation, Incrementality, Income Statement
Evaluates company positioning in the agentic commerce era across five dimensions: inventory, infrastructure, innovation, incremental sales, and income statement impact.
Agentic commerce company scorecard
Builds on the 5 I's by adding qualitative and quantitative indicators such as scale, direct user engagement, ecosystem assets, agent development, and LLM catalog integration to identify the platforms most likely to benefit.
LLM cost versus contribution profit coverage
Estimates the token cost per AI shopping session and compares it with contribution profit from improved conversion, order value, and incremental sales; except in stress scenarios, the report believes contribution profit uplift can cover LLM costs by roughly 2-3x.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- Global scaled eCommerce platformsCore beneficiary asset class
- Strengths
- Possess direct customer relationships, broad product catalogs, fulfillment and after-sales capabilities, membership ecosystems, retail media, and the scale to amortize AI development costs.
- Weaknesses
- Need to bear LLM token costs, product development investment, and user trust-building costs.
- Comparison
- Compared with smaller vertical platforms, scaled platforms are better able to control the shopping journey and retain customer relationships.
- Risks
- If horizontal LLM platforms control traffic, catalog discovery, and checkout entry points, platform economics could be diluted.
- Amazon (AMZN.O)Top-six scorecard leader; OW beneficiary
- Strengths
- Rufus already has large-scale usage disclosure, with users 60% more likely to complete a purchase, and Amazon has strong inventory, fulfillment, and ecosystem advantages.
- Weaknesses
- AI costs and diversion by external agents remain variables to monitor.
- Comparison
- Ranks among the core leading platforms in the agentic commerce scorecard.
- Risks
- If horizontal agents weaken on-site search and retail media, the profit structure could be affected.
- Walmart (WMT.O)Top-six scorecard leader; OW beneficiary
- Strengths
- Sparky users have average order values about 35% higher, and the company benefits from the combination of offline retail, online fulfillment, and membership ecosystems.
- Weaknesses
- eCommerce margins and returns on technology investment still require continued validation.
- Comparison
- Compared with pure online platforms, Walmart has omnichannel and physical fulfillment advantages.
- Risks
- If AI tools fail to sustainably raise conversion or order value, returns on investment may come in below expectations.
- Alibaba (BABA.N)Top-six scorecard leader; OW beneficiary
- Strengths
- The Qwen AI shopping experience has disclosed about 140mn users, and Alibaba has China's eCommerce ecosystem, merchant network, and AI capabilities.
- Weaknesses
- China's consumer environment, competitive intensity, and regulatory variables will affect the pace of monetization.
- Comparison
- Among Asian platforms, it has relatively strong potential for AI agent and ecosystem integration.
- Risks
- Weak macro consumption or intensified platform competition could pressure valuation recovery.
- MercadoLibre (MELI.O)Top-six scorecard leader; OW beneficiary
- Strengths
- Has clear strengths in Latin American eCommerce, payments, and logistics ecosystems, while regional eCommerce penetration still has substantial room to rise.
- Weaknesses
- Valuation and regional macro volatility are relatively high.
- Comparison
- Compared with traditional retailers, MELI benefits more directly from Latin America's online migration and ecosystem synergies.
- Risks
- Fluctuations in exchange rates, interest rates, competition, and consumer purchasing power may affect growth.
- Sea Ltd (SE.N)Top-six scorecard leader; OW beneficiary
- Strengths
- The company disclosed that AI tools supported a 14% improvement in purchase conversion in 1Q26 and reduced customer service contact costs by about 30%.
- Weaknesses
- Regional competition and earnings quality remain key concerns.
- Comparison
- Has real deployment scenarios for agentic tools within Southeast Asia's eCommerce and digital ecosystem.
- Risks
- Competitive subsidies, macro consumption, and AI costs may affect profit realization.
- Allegro (ALEP.WA)Top-six scorecard leader; OW beneficiary
- Strengths
- Has a localized platform and customer relationship base in the European regional eCommerce market.
- Weaknesses
- The table shows negative implied upside to target price, indicating more complex valuation or timing factors between market expectations and rating logic.
- Comparison
- Smaller in scale than global giants, but still has platform control in its local market.
- Risks
- Regional competition, the consumer environment, and changes in external traffic could weigh on performance.
- Additional OW beneficiariesSupplementary beneficiary list
- Strengths
- Includes Wayfair, Shopify, Chewy, Target, eBay, Meituan, Coupang, LY Corp, Mercari, FSN E-Commerce, and Temple & Webster, covering U.S. and international eCommerce, retail, and platform assets.
- Weaknesses
- These companies differ significantly in their control over agentic AI, traffic sources, category exposure, and earnings elasticity.
- Comparison
- The report believes these companies may also benefit from AI-driven discovery, improved conversion, and expanding eCommerce penetration.
- Risks
- Lower-scoring or more externally traffic-dependent companies face a wider bull-bear distribution and greater downside skew.
Key data
- 2030 global eCommerce market size forecastapproximately US$7.0tnEquivalent to a 2025-2030E CAGR of about +9%.
- Online retail penetration22% in 2025; 27% in 2030EPenetration still has upside after agentic tools improve the shopping journey.
- Five-year base-case AI shopping penetration65%The report expects nearly two-thirds of consumers to engage with agentic commerce over the next five years.
- Industry GMV materially affected by agents>20%Includes touchpoints such as AI search, discovery, recommendations, and deeper task execution.
- Agent-executed GMVapproximately 8% in the base caseIncludes higher task-execution layers such as cart building and checkout initiation.
- Autonomous GMVapproximately 2% in the base caseIncludes higher-autonomy purchasing and agent-to-agent negotiation.
- Five-year TAM uplift6.2% base case; 1.4% bear case; 14.2% bull caseThe report says the roughly 6% base-case uplift supports a re-acceleration in industry growth.
- Current share of eCommerce web traffic from GenAI platformsaverage <0.5%The direct traffic impact on profits remains limited in the near term, but valuation pressure may persist.
- Average upside for top-six scorecard leaders+26%The six companies are Amazon, Walmart, Alibaba, MercadoLibre, Sea, and Allegro.
- 2025-2028E growth for top-six scorecard leadersGMV CAGR +15%; revenue CAGR +16%; EBITDA CAGR +18%Above the report's stated overall bottom-up GMV growth of about +12%.
Impact & implications
The investment implication is that agentic AI may not weaken all eCommerce platforms; instead, it may expand online retail TAM and benefit platforms with scale, user access, fulfillment capabilities, data, and ecosystem assets. Since eCommerce sector valuations have already de-rated due to concerns about AI disruption, further company disclosures showing AI tools driving conversion, order value, or sales uplift could become catalysts for re-rating.
Risks
- Horizontal LLMs or external agents may shift traffic away from eCommerce platforms' owned channels, weakening direct visits, retail media revenue, and control over customer relationships.
- LLM token costs, model invocation costs, and agent product development costs may erode incremental sales profits, especially in stress scenarios.
- Consumer concerns around trust, privacy, automated purchasing, and data usage may slow adoption.
- Traditional shopping preferences and the need for human confirmation may keep high-autonomy agent penetration below forecasts.
- Valuations in the eCommerce sector may remain pressured by the AI disruption narrative, even if near-term traffic impact stays small.
- Smaller, vertical, or more externally traffic-dependent platforms face higher risks of economic dilution and competition.
- If company-disclosed AI uplift does not continue to translate into auditable incremental GMV, revenue, and profit, the re-rating thesis will weaken.
What to watch
- Disclosures on active users, conversion, order value, and sales uplift from agentic tools such as Amazon Rufus, Alibaba Qwen, and Walmart Sparky.
- Whether the share of eCommerce web traffic contributed by GenAI platforms rises quickly from the current average of <0.5%.
- Whether horizontal LLM platforms direct traffic to marketplace platforms or instead attempt to control product discovery, catalogs, checkout, and after-sales within their own environments.
- Token costs of AI shopping tools, model efficiency, and contribution profit coverage multiples.
- Changes in usage rates for AI shopping, product research, price comparison, and automated purchasing in consumer surveys, especially in markets such as the U.S., China, Mexico, and Brazil.
- Catalog integration and partnership models between large eCommerce platforms and LLM, search, social, or browser entry points.
- Whether the EV/revenue valuation gap between the eCommerce sector and Nasdaq narrows as more evidence of AI uplift emerges.