IBM-NRF Study: Brands and Retailers Navigate a New Reality as AI Shapes Consumer Decisions Before Shopping Begins

An IBM Institute for Business Value/National Retail Federation study (consumer survey: >18,000 respondents across 23 countries; executive survey: 200 senior leaders) finds AI is reshaping pre-purchase discovery: 72% of consumers still shop in stores while 45% use AI in their buying journeys, including 41% to research products, 33% to interpret reviews and 31% to hunt for deals. Executives report operational frictions—54% cite persistent cross-channel/system challenges and 51% limited AI expertise—leading to recommendations to redesign decision moments, prioritize data readiness and end-to-end testing, deploy AI agents where they reduce uncertainty, and invest in AI skills and partnerships; firms that can operationalize reliable data and agentic commerce stand to gain competitive advantage in retail and commerce technology.

Analysis

Market structure: Winners are AI compute and cloud providers (NVDA, AMD, MSFT, GOOGL, AMZN) and SaaS personalization platforms (SHOP, ADBE, CRM) that convert first‑party data into agentic commerce; advantaged retailers (AMZN, WMT, COST) with clean data stacks will gain share while small/mid retailers and legacy ad‑tech (some META ad products) face margin pressure. GPU/memory supply constraints will keep infrastructure pricing power elevated for 12–24 months, tightening capex cycles for others and lifting energy demand in data‑center regions.

Risk assessment: Key tail risks include 1) regulatory action (EU AI Act/US FTC) within 6–18 months that could force model disclosures or limit personalization; 2) chip export bans (weeks–months) that disrupt supply; and 3) major data breaches or systemic model failures causing reputational losses >$500M for big retailers. Near term (0–3 months) expect event‑driven volatility around earnings; medium/long term (6–36 months) divergence between infrastructure winners and execution‑poor retailers.

Trade implications: Tactical allocation should favor capped‑risk exposure to NVDA/AMD and cloud leaders, paired with selective long SaaS personalization (SHOP/ADBE) and shorts in mall/low‑data retailers or XRT ETF. Use defined‑risk option spreads to express directional views given elevated IV; rotate into logistics/last‑mile (FDX, UPS) and utilities powering hyperscalers for defensive exposure.

Contrarian angles: Consensus underestimates the data‑readiness gap — many “AI” projects will not monetize, so pure AI hype names are vulnerable to de‑rating once ROI tests fail (3–9 months). NVDA and cloud names are richly valued; prefer option structures or relative trades into under‑owned enablers (SHOP) and infrastructure beneficiaries (energy/utilities) that the market is underweight.

Key Decisions for Investors

• Establish a 2–3% portfolio long in NVIDIA (NVDA) over next 2 weeks via a 3‑month call spread: buy ATM calls, sell calls ~+15% strike to target ~+15% upside and cap downside; set tactical stop‑loss at -10% from entry.
• Add 1.5–2% long positions in MSFT and GOOGL (1%–1.25% each) for cloud AI exposure, hold 6–18 months; trim 30% if combined cloud guidance misses consensus by >3% on next earnings.
• Pair trade: Long 1.5% Shopify (SHOP) vs short 1.5% XRT (retail ETF) to capture SaaS personalization gains vs weak mid‑market retailers; size to be delta‑neutral and re‑price after Q1 retail reports (~90 days).
• Use options to express downside in mall/legacy retail: buy 3‑month put spread on Macerich (MAC) or short 1% position in MAC with a 6% stop; target asymmetric payoff if agentic commerce accelerates SKU concentration within 6–12 months.
• Establish a 1–1.5% overweight in utilities/energy beneficiaries of data‑center growth (e.g., NEE or DUK) with 6–12 month horizon; increase allocation by 50% if GPU orders and hyperscaler capex guidance outpace consensus by >10% in next two earnings cycles.