Why Nokia Could Be A Safer Bet On AI Infrastructure After The NVIDIA Deal

Nokia’s Q1 2026 results showed revenue of 4.5 billion euros (+4% YoY), operating profit of 281 million euros (+54%), and EPS of 0.05 euros versus 0.03 euros a year ago, beating forecasts by about 31%. AI and cloud sales surged 49%, the company booked 1 billion euros in new cloud orders, and it raised full-year guidance for network infrastructure growth to 12%-14%, optical/IP networks to 18%-20%, and operating profit to 2.0-2.5 billion euros. The Nvidia-backed AI-RAN strategy, with T-Mobile as the initial deployment partner, continues to reshape sentiment and supports the stock’s 140% year-to-date rally.

Analysis

Nokia is no longer trading as a telecom utility; it is being rerated as a scarce picks-and-shovels proxy for AI capex with a network-edge twist. The key second-order effect is that Nvidia’s equity investment does more than validate the story — it likely lowers customer adoption friction because hyperscalers and operators now view Nokia as part of a broader AI supply chain rather than a standalone carrier vendor. That should help funnel more pilot demand into commercial deployments, but it also means the stock is now partly hostage to the durability of AI infrastructure spend and the market’s willingness to pay for a multi-year option on that theme.

The immediate competitive implication is that Nokia is gaining share-of-wallet in an area where incumbents have been slow to monetize AI-RAN economics. If AI-RAN proves real, the winners are not just NOK and NVDA; AMD, Keysight, and optical component suppliers can benefit from lab-to-field-trial conversion, while traditional RAN peers face a valuation overhang if investors conclude they are missing the AI layer. The downside is that telecom operators are notoriously conservative on ROI, so the first failed field economics readout would hit the multiple harder than a modest revenue miss would hit fundamentals.

Catalyst timing matters: the next 6-10 weeks likely trade on expectations into Q2 and early field-trial headlines, but the real inflection is later in the year when trials either expand or stall. That makes this a classic event-driven momentum setup with asymmetric downside if management commentary shifts from deployment to experimentation. The move is also likely over-owned by fast money after a 140% YTD run, which raises the probability of a sharp air-pocket on any sign that AI order growth is lumpy or supply-capex is front-loaded rather than demand-led.

The consensus seems to be underestimating how much of the valuation now rests on operating leverage from optical and cloud orders, not just the AI-RAN narrative. If the company can convert elevated demand into sustained margin expansion while keeping capex disciplined, the rerating can extend; if not, the stock can quickly revert to a more ordinary infrastructure multiple. In short, the setup favors staying long only through clearly defined catalysts, not as an unconstrained core hold.

Key Decisions for Investors

• Stay long NOK into Q2 earnings as a catalyst trade, but size it as a 6-8 week event position rather than a strategic hold; target further upside if management re-accelerates AI/cloud order commentary, cut if order conversion stalls.
• Use NOK call spreads instead of outright stock for the next catalyst window; upside remains meaningful, but after the sharp rerating implied volatility is likely cheaper than the risk of a post-earnings compression.
• Pair long NOK vs short ERIC on the thesis that AI-RAN and optical monetization are creating a cleaner growth narrative for NOK while legacy RAN peers face slower rerating; monitor for any evidence of peer catch-up.
• Add small tactical longs in NVDA and KEYS on NOK weakness if field-trial momentum improves, as both benefit from broader AI networking validation with lower company-specific execution risk.
• Take profits on NOK into any post-earnings spike above prior highs; the risk/reward shifts sharply if the market starts pricing perfection before real-world operator economics are proven.