Goldman Sachs says these telecom stocks are defensive buys despite AI risk

Goldman Sachs estimates agentic AI could reduce European telecom operators' medium-term free cash flow by 5%-20%, based on 25%-100% adoption among out-of-contract customers, although mitigating actions could halve the impact. The broker expects consumer adoption to unfold over 3-5 years, retains unchanged estimates and sees upside to consensus over the next two years. Goldman identifies Orange and KPN as defensive Buys after recent weakness, while BT, Deutsche Telekom, Vodafone and Telefónica offer higher risk-reward.

Analysis

The relevant mechanism is not AI-driven network cost savings; it is a reduction in consumer search and switching friction that could force legacy tariff cohorts toward economically rational pricing. That pressure is most acute where operators carry large high-ARPU back books, distribution commissions are material, and convergence does not meaningfully raise switching costs. The equity risk is therefore asymmetric: a modest revenue rebase can produce disproportionate FCF and valuation damage for highly levered incumbents, because dividend capacity and deleveraging plans are generally underwritten on stable service-revenue growth.

KPN and Orange offer the cleaner defensive expression because market structure, bundled customer bases, and relatively disciplined pricing should preserve more of the value currently embedded in legacy plans. BT and Vodafone have greater operating leverage to a pricing reset, but also the highest upside if management converts lower acquisition/distribution costs into opex reduction before revenue leakage becomes visible. Deutsche Telekom is a more complicated case: its European exposure may be resilient, but the stock’s valuation is driven materially by T-Mobile US, making a Europe-specific AI-pricing thesis an imperfect short catalyst.

Consensus may be over-discounting a near-term disruption. Consumer adoption must be sufficiently scaled, agents need reliable access to plan-level pricing and cancellation workflows, and operators can counter with simplified tiers, retention offers, and AI-enabled targeted pricing. Over the next 1-3 months this is primarily a sentiment and relative-value trade; the investable fundamental catalyst is 2027 service-revenue guidance and evidence of rising churn or lower retention ARPU. A sustained increase in porting rates, a cut to service-revenue outlook, or adverse regulatory mandates on automated switching would validate the bear case over 6-18 months.

Key Decisions for Investors

• Initiate a 3-6 month long KPN / short BT.A relative-value position after confirming no broad European rates-driven telecom rally; target 8-12% relative return. The thesis is resilience of pricing and convergence versus greater UK legacy-plan and balance-sheet sensitivity. Stop if BT delivers service-revenue growth above guidance with stable churn for two consecutive reporting periods.
• Maintain ORA as the preferred defensive long among continental incumbents, sized modestly ahead of the next earnings cycle. Underwrite mid-single-digit upside from multiple stabilization rather than an immediate earnings upgrade; exit if French mobile pricing deteriorates materially or management lowers cash-flow/distribution targets.
• Do not short DTE solely on this theme. Use any Europe-telecom weakness to evaluate a DTE / VOD pair only if T-Mobile US guidance remains intact, as U.S. exposure can overwhelm the European pricing effect; a widening European service-revenue-growth gap versus Vodafone is the required confirmation.
• For higher-beta exposure, place an alert rather than enter VOD outright: initiate only after management demonstrates that customer-acquisition and channel-cost savings are offsetting retention pressure. A failure to protect service-revenue growth at the next results would create a clearer 6-18 month short or put-spread opportunity.
• Monitor quarterly churn, porting activity, retention ARPU, and third-party distribution expense across KPN, ORA, BT.A, VOD, and TEF. These are the earliest falsifiable indicators; headline discussion of consumer AI without measurable switching behavior is insufficient to justify a sector-wide de-rating.