Time to Rethink Nokia's Valuation: AI Revenue Under 10% Drives 60% of Growth

In the second quarter of 2026, Nokia's revenue increased by 372 million euros year-over-year, with 226 million euros originating from AI and cloud customers. In other words, while these customers accounted for only 9.3% of the quarter's revenue, they contributed 61% of the year-over-year growth. The growth engine has shifted, but profit and cash flow have yet to fully keep pace.

First, a word on the profit metrics used throughout this article. Nokia reports both "comparable" and "reported" results: comparable metrics exclude company-defined adjustments such as restructuring costs and acquisition-related amortization, making them better suited for evaluating core operations; reported metrics account for these costs under IFRS, reflecting the complete income statement.

In Q2, comparable operating profit reached 434 million euros, up 18% year-over-year; however, on a reported basis, the company posted a loss of 50 million euros, with a free cash outflow of 732 million euros. While comparable results show improving core performance, the excluded restructuring expenses still represent real cash spending.

To state the conclusion upfront: I am moderately optimistic on Nokia, but not because it is cheap. The current stock price is betting on three things—continued high growth in the AI networking business, Network Infrastructure operating margins rising toward the mid-teens, and AI-RAN potentially delivering recurring software subscription revenue. The first has already been validated by orders and revenue, the second shows only early signs of improvement, and the third remains a call option.

Stop Treating It as Just a 5G Stock: The Largest, Fastest-Growing, and Most Profitable Businesses Are Not the Same

Today's Nokia does not sell mobile phones; it sells networks—connecting homes, mobile devices, and data centers, while enabling high-speed data flow between different locations.

Starting in 2026, the company restructured its operations into two main segments: Network Infrastructure and Mobile Infrastructure. The patent business, Nokia Technologies, which was previously reported separately, was integrated into Mobile Infrastructure and renamed Technology Standards.

The remaining ~2% comes from discontinued operations and other items.

Prior to the restructuring, Nokia Technologies was a standalone segment, allowing direct visibility into its 2025 profitability: revenue of 1.501 billion euros, operating profit of 1.059 billion euros, and an operating margin of 70.6%. In other words, for every 100 euros in revenue, roughly 71 euros converted into operating profit.

However, the 15% growth in Technology Standards during Q2 included catch-up licensing revenue: following new agreement signings, Nokia recognized unbilled license fees for past periods covered under those agreements in a lump sum. Consequently, this revenue will not recur quarterly, and the 15% growth rate cannot be directly extrapolated.

To understand Nokia, remember just three things: the largest business is Radio Networks, the fastest-growing is Optical and IP Networks, and the most profitable is patents. Lumping all three under a single P/E ratio makes it easy to misjudge the company.

What's Really Taking Off Isn't Base Stations, But Networks Connecting AI Data Centers

In the past, Nokia relied primarily on telecom operators to build and upgrade communications networks; today, marginal growth is increasingly driven by AI and cloud customers.

AI data centers buy more than just GPUs. The more GPUs deployed, the more data must be exchanged among servers, storage, and distinct campuses. If the network bottlenecks, expensive GPUs sit idle waiting for data. As Big Tech expands AI infrastructure, networks must upgrade in tandem.

Nokia does not handle compute. Its established position is at data center exits, inter-campus optical transport, and telecom backbone networks: IP Networks route data rapidly to the correct destination, while Optical Networks move massive volumes of data to another campus, city, or region.

Intra-data center networking is a market where Nokia is expanding rather than holding an established advantage. The company has disclosed design wins and order growth in this segment, but has not yet disclosed market share data to confirm its competitive standing.

In Q2, Network Infrastructure generated 2.037 billion euros in revenue, with Optical and IP Networks contributing a combined 1.547 billion euros, or 76%. Growing by 20% and 16% respectively, these two units drove segment operating profit up 42% year-over-year to 166 million euros.

This does not mean the entire 1.547 billion euros comes from AI, but it shows that the two units benefiting most from data center construction are driving the segment's growth trajectory.

Customer and order signals are even more direct. AI and cloud revenue rose from 220 million euros in Q2 2025 to 446 million euros in Q2 2026, while orders surged from 1.0 billion euros in Q1 2026 to 2.8 billion euros in Q2. Management expects roughly half of these orders to convert into revenue over the next twelve months.

The 2.8 billion euros in orders represents a queue for future revenue rather than a new quarterly run-rate. The company noted that supply constraints are prompting customers to secure longer-term orders in advance. While this enhances revenue visibility, it also means a subsequent moderation in order intake would not necessarily signal a reversal in demand.

Elongated lead times also heighten the impact of cost fluctuations, spec changes, and customer project delays on gross margin. Thus, while the 2.8 billion euros proves robust demand, it does not mean an equivalent level of profit is locked in.

Nokia is easing delivery bottlenecks through the integration of Infinera (acquired in 2025) and capacity expansion in U.S. optical chips, testing, and packaging. Management targets net comparable operating profit synergies of over 200 million euros by 2027; while this could enhance delivery capacity and cost control, it remains a target rather than realized profit.

Currently, the most credible AI narrative centers on Optical and IP Networks. It is already supported by orders, revenue, and profits, but fundamentally remains an infrastructure business subject to capex cycles, market share shifts, and pricing pressure—not a software subscription business.

€2.8 Billion in Orders Is Just the Start; Margins Are the Valuation Switch

Orders answer whether customers want to buy; margins answer how much money Nokia gets to keep.

In early 2026, the company's full-year growth assumption for Optical and IP Networks was just 10%–12%; by Q2, this was raised to 18%–20%, yet the three-year CAGR target remains at 10%–12%.

Backing into the figures on the same basis, growth in 2027 and 2028 would only need to average ~5%–9% annually to reach the target. Management has not extrapolated current high growth linearly, keeping the mid-term bar unaggressive.

Margins provide even greater leverage. On the latest restated basis, Network Infrastructure generated revenue of 7.646 billion euros and operating profit of 760 million euros in 2025. If revenue grows at a three-year CAGR of 6%–8%, it will reach 9.107 billion to 9.632 billion euros by 2028. Applying the target margin of 13%–17%, operating profit would rise to 1.184 billion to 1.637 billion euros, representing a growth of ~56%–116% compared with 2025.

Revenue growth is only the first layer; margins are what truly amplify profits. If operating profit retained per 100 euros of revenue increases from 9.9 euros to 13–17 euros, segment profits will outpace revenue growth significantly. Upside drivers include Infinera synergies, economies of scale, and an improved product mix; conversely, incremental R&D, capacity expansion depreciation, and price competition could delay this progress.

Ultimately, however, profit must convert into cash. In Q2, free cash flow was a negative 732 million euros, with trailing 12-month free cash flow at ~550 million euros; the company projects restructuring cash outflows of 700 million to 800 million euros in 2026. Nokia generated 1.465 billion euros in free cash flow in 2025, representing a conversion rate of ~70%. Thus, the question is not whether it can generate cash, but whether cash generation can sustainably recover once restructuring concludes.

Also needing clarification is the headline positive of the "guidance hike." The upward adjustment in full-year comparable operating profit from 2.0–2.5 billion euros to 2.1–2.6 billion euros was merely a technical adjustment resulting from moving two loss-making units out of continuing operations; management explicitly stated that the operational outlook remains unchanged.

Legacy businesses have yet to yield incremental profits. Mobile Infrastructure Q2 revenue grew 7% at constant currency, yet operating profit remained flat at 310 million euros, with margin declining from 12.2% to 11.6%. While Nokia does not disclose standalone profits for Radio Networks, Core Software, and patents—precluding a definitive claim that the wireless unit is unprofitable—it is clear that incremental revenue generated no incremental segment profit.

Nvidia Is Not Betting on a Batch of Base Stations, But a New Monetization Model

Traditional wireless networks rely on hardware refreshes every few years. AI-RAN utilizes AI-accelerated compute to run and optimize the radio access network—the system connecting mobile devices to base stations. Its commercial significance lies in enabling operators to continuously upgrade existing networks via software, rather than waiting for the next hardware cycle.

On October 28, 2025, Nvidia invested $1 billion in Nokia, taking an approximate 2.9% stake. Beyond collaborating on AI-RAN, the two companies are exploring the integration of Nokia's data center switching and optical networking technologies into Nvidia's future AI infrastructure architecture.

In July 2026, Nokia stated that AI-RAN demonstrated a more than 20% improvement in spectral efficiency. Put simply, according to company tests, the same scarce spectrum can carry over 20% more traffic, relieving operators of the pressure to acquire new spectrum or build additional sites, while giving Nokia's software upgrades a clear ROI proposition.

The company plans to launch trials by late 2026 and achieve commercial availability in 2027 using a software subscription model. Looking further ahead, edge compute resources at base stations could also handle low-latency edge AI inference. However, until standalone AI-RAN revenue, renewal rates, and gross margins are demonstrated, this partnership should be treated as upside potential rather than baseline earnings.

The Stock Isn't Cheap, But Management May Underestimate Its Own Upside

The following valuation reflects fair value as of late 2027, using 2028 financial performance as a baseline. Rather than representing an immediate fair value today, it serves as a scenario midpoint as the market prices in the Infinera integration, optical expansion, and margin trajectory over the next 16 months.

As of the close on August 12, 2026, NOK stood at $10.32. Based on ~5.655 billion shares outstanding and current EUR exchange rates, market capitalization is ~50.5 billion euros; subtracting Q2-end net cash minus August dividend payouts yields an Enterprise Value (EV) of ~48.0 billion euros. This price corresponds to ~15–18x management's 2028 target of 2.7–3.2 billion euros in comparable operating profit. For context, mature, low-growth telecom equipment vendors typically trade at 8–12x forward EV/EBIT, indicating that the market is already paying a premium for Optical and IP growth, margin expansion, and the AI-RAN option—rather than valuing Nokia as a slow-growth 5G hardware vendor.

The valuation model breaks down into two components:

• Network Infrastructure determines growth and the valuation ceiling;
• Mobile Infrastructure—comprising Radio Networks, Core Software, and high-margin patents—serves as the profit foundation.

Higher multiples are applied to Network Infrastructure across scenarios due to its accelerating growth and margin expansion; lower multiples for Mobile Infrastructure reflect legacy telecom cycles, hardware competition, and patent revenue lumpiness. AI-RAN is not assigned separate standalone value; its potential contribution is subsumed within Mobile Infrastructure's profits and multiples under the Bull Case.

Below are my estimated scenario midpoints, not exact price targets:

One NOK American Depositary Receipt (ADR) represents one Nokia ordinary share. To keep the model concise, discontinued operations and minor items are not valued separately; their impact is absorbed into conservative segment multiple assumptions.

The Bear Case does not require AI demand to evaporate. It simply assumes order growth cools noticeably after 2026, Network Infrastructure margin stalls at 12%, and Mobile Infrastructure earnings remain under pressure. Under this scenario, the current share price lacks downside support.

The Base Case assumes Nokia largely delivers on management targets: Network Infrastructure maintains high-single-digit growth in 2027–2028 with margins reaching 16%, while Mobile Infrastructure remains stable. This yields an estimated value of ~$10, close to the current stock price. In short, the market has already priced in execution of the target transformation, meaning merely meeting guidance is unlikely to yield exceptional capital returns.

The Bull Case requires three catalysts aligned: Network Infrastructure sustains low-double-digit growth through 2027–2028, operating margins exceed the upper target bound of 17% toward 20%, and AI-RAN begins generating verifiable software revenue. In this scenario, late-2027 value reaches ~$16.

Thus, current valuation is not bargain-basement, yet it does not fully discount the Bull Case either. The market has paid for targets met, but not yet for sustained upside surprises.

My Optimistic Stance: The Market Has Priced in Meeting Guidance, But Not Outperformance

I view Nokia constructively, not because it will become the next Nvidia, nor because every segment is recovering. My optimism stems from three verifiable dynamics: AI and cloud customers are driving the lion's share of growth, Optical and IP order intake and revenue have accelerated, and mid-term management targets do not rely on current high growth rates persisting indefinitely.

The legacy Nokia business does not need to vanish. As long as Radio Networks and Fixed Networks remain steady, patent revenue provides a solid earnings baseline for transformation, Optical and IP spearhead growth, and AI-RAN supplies additional upside.

Over the next four quarters, I am monitoring three key metrics: whether the 2.8 billion euro order backlog converts into revenue on schedule, whether Network Infrastructure operating margins sustainably return to double digits toward 13%–17%, and whether free cash flow normalizes post-restructuring.

Nokia is not a cheap speculative AI play, but an established enterprise reshaping its earnings quality through tangible order backlog. If Network Infrastructure growth decelerates less sharply post-conversion than implied by mid-term guidance while margins track toward the upper end or beyond, I believe Nokia holds room to rerate higher.

Disclaimer: The analysis in this article represents a research framework based solely on publicly available information and does not constitute investment advice.