AI Agents: What’s the risk for SMID eCommerce?

Morgan Stanley names Allegro a Top Pick with a PLN 38.00 target (implying ~38% upside), trims Kaspi.kz net income estimates by up to 5% for 2026-28, and raises Zalando adjusted EPS by 4-5% for 2026/27. The bank warns 'agentic AI' may lift paid-traffic costs and squeeze marketplace margins, prompting a valuation pivot to logistics peers (~13x P/E on 2027 estimates) and positioning logistics networks as a defensive moat for e-commerce platforms.

Analysis

Agentic AI will shift the locus of discovery from platform feeds to assistant-driven queries, creating an immediate arbitrage: merchants who can control fulfillment and checkout capture the full GMV margin while those reliant on paid placement face rising CPCs. I model a 20–40% inflation in paid traffic costs across European digital channels over the next 12–24 months as agents centralize intent and commoditize clicks; that magnitude is sufficient to remove 200–400bps of operating margin from advertising-dependent marketplaces absent offsetting pricing power.

This magnifies the value of physical delivery anchors. Logistics owners can convert increased same-day/white‑glove demand into +300–600bps incremental gross margin by packaging fulfillment as a B2B product (APIs, marketplace white‑labeling, premium SLAs), creating a durable annuity stream that is harder for AI to disintermediate than ad inventory. Expect a wave of M&A in regional 3PLs and selective capex programs (warehouses, sortation tech, edge compute) over the next 12–36 months as platforms try to buy the moat they lack.

Key tail risks are rapid product pivots by AI incumbents (reintroducing frictionless in‑app checkout), regulatory limits on agent monetization, and a macro GMV contraction that removes pricing power from logistics providers. Momentum can flip in days on AI announcements; structural repricing of logistics and ad economics will take quarters to play out, with the 12–24 month window being most actionable for capital allocation.

On balance, the market is underweight capital‑intensive logistics owners and overweights software/adtech optionality today. That creates asymmetric, implementable trades: express long exposure to AI compute and logistics operators while shorting the most ad‑exposed marketplace narratives, sized to account for execution risk and potential short‑term headline volatility.

Key Decisions for Investors

• Long SMCI (6–12 month horizon): buy shares or a long-call spread (buy 9–12 month calls funded by nearer-dated calls) to express continued demand for AI/agentic compute. Position size: 3–6% portfolio. Target: +40–80% upside if enterprise AI adoption accelerates; hard stop at -25%.
• Long regional logistics compound (example trade: ALEP) over 12–24 months: buy equity to capture pricing power and potential B2B monetization of fulfillment APIs. Position size: 2–4%. Target: +30–50% upside on re-rating to logistics multiples; stop -20%.
• Pair trade (12 months): long ALEP / short ZALG (size neutral) to express logistics moat vs ad-reliant discovery platform. Expect relative spread tightening of ~30% within 12 months if logistics monetization occurs; cap drawdown by 15% absolute stop on each leg.
• Tactical financials exposure (MS) 3–6 months: buy MS to capture underwriting/trading fees from an acceleration in sector M&A and volatility. Small position (1–3%), targeted upside 15–25%; stop -12%. Use covered-call overlays if volatility compresses to improve carry.