Any Nokia stock prediction that starts from the 2026 rally has the story backwards. Nokia (NYSE: NOK) closed at $10.39 on 25 September 2026, up 122% over twelve months on the NYSE ADR line, and almost every write-up frames that as an AI-RAN re-rating still in progress. The tape says the opposite. The ADR peaked at a $16.85 close on 2 June and has given back 38% since. The cleanest scoreboard is not the share price but Nvidia’s stake: it bought 166,389,351 new Nokia shares at $6.01 in October 2025, a position worth about $1.73bn at Friday’s close, down from roughly $2.80bn at the June high. Nvidia is still up around 73% on the trade and has still watched about $1.07bn of paper profit evaporate in under four months.
That gap is the whole argument, and it resolves into one number nobody is putting on the page. Nokia’s full-year guidance is EUR 2.1bn to EUR 2.6bn of comparable operating profit. Through the first half it booked EUR 735m. Hitting the midpoint requires EUR 1.615bn in the second half, roughly 2.2 times what the first half produced, and management has said it expects to land above that midpoint. Every bull and bear case for Nokia between now and the 22 October Q3 print is a bet on whether that back-end load is a phasing artefact or a stretch. The second data point that matters: when Nokia launched AI-RAN with Nvidia it promised more than 100% spectral efficiency gains by 2028, and as of 16 September 2026 the platform has delivered more than 20%. That is real progress against an unusually specific promise, and the distance between 20 and 100 is where the multiple lives.
Key facts
- NOK closed at $10.39 on 25 September 2026 (4:00pm EDT), with an after-hours print of $10.29. 52-week range $4.63 to $17.45 — stockanalysis.com, 25 September 2026
- Q2 2026 comparable operating profit of EUR 434m, up 18%, on net sales of EUR 4,815m; reported operating result was a loss of EUR 50m after EUR 390m of restructuring charges — Nokia Q2 report, 23 July 2026
- Q2 free cash flow was negative EUR 732m, and net cash fell EUR 1,012m in the quarter to EUR 2,776m — Nokia Q2 report, 23 July 2026
- AI & Cloud order intake reached EUR 2.8bn in Q2, with about half expected to convert to revenue within twelve months — Nokia Q2 report, 23 July 2026
- Nvidia subscribed for 166,389,351 shares at USD 6.01, about USD 1.0bn for roughly 2.90% of Nokia — Nokia inside information release, 28 October 2025
- The AI-RAN platform has delivered more than 20% spectral efficiency improvement, against a launch claim of more than 100% by 2028 — Nokia AI-RAN release, 16 September 2026
- Sell-side consensus is a $14.97 average target across 14 analysts, with a $8.50 low and a $21 high — S&P Global data via stockanalysis.com, 25 September 2026
What actually repriced between June and September
Having tracked this ADR since Nvidia’s investment was announced, the June-to-September drawdown looks less like a verdict on AI-RAN and more like the market marking a timing assumption. Nothing in Nokia’s product story broke. The Q2 print on 23 July was, on the comparable line, good: net sales of EUR 4,815m, up 9% at constant currency, comparable gross margin up 70 basis points to 46.0%, comparable operating margin up 70 basis points to 9.0%. Network Infrastructure grew 12% at constant currency, with Optical Networks up 20% and IP Networks up 16%. Sales to AI and cloud customers grew 105%.
What the same report also contained was a reported operating loss of EUR 50m, a 430 basis point deterioration, driven by EUR 390m of restructuring charges taken in a single quarter. And free cash flow of negative EUR 732m. A company guiding to EUR 2.1bn to EUR 2.6bn of comparable operating profit consumed nearly three quarters of a billion euros of cash in three months. Net cash dropped by EUR 1,012m to EUR 2,776m. The comparable numbers and the cash numbers are telling different stories, and the market spent the summer deciding which one to price.
The 14 September session is the clearest single marker. NOK fell 13.3% on 124m shares, its worst day of the year, with no earnings release, no guidance change and no downgrade attached to it. That is what a crowded positioning unwind looks like when a sector-wide reassessment of AI infrastructure spending arrives. The stock has since stabilised between its 50-day moving average of $10.06 and its 200-day of $9.91, which is a fair description of where conviction currently sits: nowhere in particular.
The index flow was a genuine, separate tailwind that has now passed. Nokia re-entered the Euro Stoxx 50 on 21 September, replacing Volkswagen, a mechanical bid we covered in detail when the reshuffle was confirmed and again when the ETF flow was sized. Index inclusion is a one-off transfer of shares from active to passive hands. It is not a fundamental input, and anyone treating the post-inclusion price as a validated level is reading a technical event as a business event.
What the players are actually doing
The operator list is the part of this story that has moved most and been written about least. On 16 September Nokia named A1 Group, Chunghwa Telecom, du, e&, Mobily, stc, TPG Telecom and Zain Saudi as carriers advancing AI-RAN proofs of concept and live trials on Nvidia’s Aerial RAN Computer, alongside previously disclosed work with NTT DOCOMO, T-Mobile, SoftBank and Indosat Ooredoo Hutchison. Twelve named operators across four regions is a real commercial pipeline rather than a demonstration.
It is also, importantly, a pipeline of trials. None of those names has been disclosed as a volume commercial deployment, and the revenue timing question stays open. Nokia’s own framing on order conversion is the number to hold onto: of the EUR 2.8bn of AI and Cloud orders taken in Q2, management expects around half to convert to revenue over the following twelve months, which puts the bulk of the conversion into 2027 rather than the current guidance year.
Nvidia’s posture has been consistent and worth reading as a supplier signal rather than a shareholder one. The company holds roughly 2.90% of Nokia, a position we sized against its wider equity book in our look at Nvidia’s portfolio stakes. Ronnie Vasishta, Senior Vice President of Telecom at Nvidia, framed the goal in the 16 September announcement: “AI-RAN will transform the mobile network into one of the world’s largest distributed AI infrastructures.” That is a statement about a decade, not about a quarter.
The independent read matters more than either vendor’s. Remy Pascal, Practice Leader for Mobile Infrastructure at Omdia, was quoted in the same release with a deliberately cooler assessment: “The breadth of operator engagement announced by Nokia shows that AI-RAN is evolving from a research topic into a strategic priority for a growing number of service providers… As deployments mature, collaboration between operators, vendors and the broader ecosystem will be critical to realizing AI-RAN’s full potential.” “Evolving from a research topic into a strategic priority” is analyst language for early. Omdia is also the source of the figure Nokia and Nvidia both cite for the prize, a cumulative AI-RAN opportunity exceeding $200bn by 2030.
The China question the market has already decided
Consensus now treats “Nokia is exiting China” as settled fact. The reporting deserves more care than it has been given. The South China Morning Post reported on 18 August 2026 that Nokia would close almost all of its mainland China sites by year end, citing people familiar with the situation and a Shanghai-based source. Nokia did not confirm it on the record in that article. Nokia’s annual report put headcount across mainland China, Hong Kong and Taiwan at about 7,200 at the end of 2025.
Nokia’s own disclosure describes something narrower. The Q2 report lists, under operating model changes, “simplifying Nokia’s operating structure in China”, explaining that having taken full ownership of its Chinese joint venture the company would integrate those operations into its global operating model, with integration charges of EUR 350m to EUR 400m and EUR 200m of expected cost savings, now targeted for completion within two years. That is a restructuring of a consolidated subsidiary, disclosed and budgeted. It is not the same claim as a wholesale market exit, and the two have been merged in coverage.
Both can be directionally true, and for the equity the distinction is mostly about cost timing rather than revenue. China has not been a material growth market for Nokia for years. The structural point that does matter is the same one reshaping the optical component supply chain: Western and Chinese telecom infrastructure are separating into two stacks, and a vendor with no Chinese exposure is structurally advantaged in Western AI data-centre buildouts. Nokia is paying exit costs now for a cleaner position later.
The numbers, and the chart
At $10.39 the ADR carries a $58.24bn market capitalisation and a $56.33bn enterprise value on 5.60bn shares. Trailing earnings put the multiple at 72 times, which is meaningless for a company mid-restructuring. Forward multiples are the usable frame: consensus has FY2026 EPS at $0.34 and FY2027 at $0.40, putting the stock on 30.6 times this year and 26.0 times next. Price to free cash flow sits at 95.6, which is the cash problem expressed as a ratio.
The sell-side spread is unusually wide and worth reading as information rather than noise. Across 14 analysts polled by S&P Global the average target is $14.97, the median $15, the low $8.50 and the high $21. That high-to-low ratio of 2.5 times on a EUR 19bn-revenue incumbent is not normal dispersion. Recent initiations cluster tightly: Dave Kang at B. Riley Securities started at $15 on 17 September, Michael Genovese at Rosenblatt Securities started at $15 on 14 September, and Jim Kelleher at Argus Research reiterated $15 on 24 July. The $21 outlier is Sandeep Deshpande at J.P. Morgan, maintained on 19 August. September ratings break down as six Strong Buy, four Buy, two Hold and one Sell.
| Bull inputs | Bear inputs |
|---|---|
| EUR 2.8bn AI & Cloud order intake in Q2, half converting within twelve months | Q2 free cash flow of negative EUR 732m; P/FCF of 95.6 |
| Comparable operating margin up 70bps to 9.0%; gross margin 46.0% | Reported operating loss of EUR 50m; EUR 390m quarterly restructuring charge |
| Twelve named AI-RAN operators across four regions | All trials; no disclosed volume commercial deployment |
| Cost programme tracking the high end of EUR 1.2bn gross savings | H2 must deliver about 2.2x H1 operating profit to reach guidance midpoint |
| Optical +20%, IP +16%, insulated from Chinese competition | Share count up 2.76% year on year; EUR 350m more China integration charges by year end |
One more structural detail: dilution is real but small. The Nvidia issuance is the main reason the share count rose 2.76% year on year. Shareholders gave up roughly 2.9% of the company to secure the silicon partnership that now underpins the entire AI-RAN thesis, at a price of $6.01 when the stock traded near $6. Judged on the strategic logic that looks like a fair trade. Judged on price, Nvidia bought well.
The call: base $12.40, bull $17.00, bear $7.80
Any honest Nokia stock prediction has to route through 22 October, when the company publishes Q3 and January-September results before the US open. Management has told the market what to expect: net sales up 3% to 7% quarter on quarter, comparable operating profit roughly flat from Q2 into Q3 because of software revenue phasing, and a meaningful increase in Q4. That guidance structure means Q3 is not where the year is won. Q3 is where the credibility of the Q4 hockey stick is either confirmed or broken.
Base case, $12.40, 19.3% above spot, 45% probability. Q3 lands inside the guided range, free cash flow turns positive or close to it as restructuring outflows peak, and management reiterates the EUR 2.1bn to EUR 2.6bn range. On 31 times FY2027 consensus EPS of $0.40 that is $12.40, below the sell-side average and roughly where the stock traded in April. This is the outcome the guidance as written implies.
Bull case, $17.00, 63.6% above spot, 25% probability. Requires two things together: a Q3 beat at the top of the 3% to 7% sequential range, and the first disclosed conversion of AI-RAN trial operators into commercial contracts with stated values. Order intake would need to exceed the Q2 EUR 2.8bn. That combination would take the stock through the 2 June high of $16.85 to a new post-rally peak and validate the J.P. Morgan $21 as a twelve-month path rather than an outlier. The mechanism is multiple expansion on evidence, not on narrative.
Bear case, $7.80, 24.9% below spot, 30% probability. A second consecutive quarter of materially negative free cash flow, or any language softening the Q4 step-up, removes the growth multiple. At 23 times FY2026 consensus EPS of $0.34 the stock is $7.80, below the $8.50 street low and back toward the pre-rally base. The trigger is cash, not orders. Nokia can announce operator wins every month and still de-rate if the cash conversion does not arrive, because a 95.6 price-to-free-cash-flow ratio prices a turn that has not happened yet.
What would change my mind. On the bullish side, a disclosed AI-RAN contract with a named operator and a stated euro value before year end would move the base case up, because it converts the single biggest open question from timing to arithmetic. On the bearish side, the invalidation is specific: if Q3 free cash flow is negative by more than EUR 300m, the EUR 2.1bn to EUR 2.6bn range stops being credible regardless of what the comparable operating line prints, and the bear probability should rise above the base. A guidance range narrowed downward on 22 October would do the same. The $7 bear on our previous Nokia page sat below where the stock has actually traded all year; $7.80 is the level the current cash profile supports, and it is still 25% below Friday’s close.
On the income line, the board distributed EUR 0.04 per share on 6 August with a remaining authorisation of up to EUR 0.06. Used in full, that is EUR 0.10 for the year, about $0.11 per ADS at the ECB reference rate of 1.1403 on 25 September, or roughly a 1.1% yield. Nobody owns this for the dividend, and the buy case does not improve if they do.
FAQ
What is the Nokia stock price today?
Nokia’s NYSE ADR closed at $10.39 on 25 September 2026 at 4:00pm EDT, down 0.38% on the session, with an after-hours print of $10.29. The 52-week range is $4.63 to $17.45. All figures in this analysis are on the USD ADR line, not the Helsinki EUR listing, and the two should not be compared directly.
When does Nokia report Q3 2026 results?
Nokia publishes its third quarter and January-September 2026 results on 22 October 2026, before the US market open, as stated in its Q2 report. Management has guided to a 3% to 7% sequential increase in net sales and roughly flat comparable operating profit from Q2 into Q3, with a meaningful increase expected in Q4.
How much of Nokia does Nvidia own?
Nvidia subscribed for 166,389,351 new Nokia shares at $6.01 each on 28 October 2025, an investment of about $1.0bn representing roughly 2.90% of Nokia after the issuance. The shares were delivered as American Depositary Shares. At $10.39 that stake is worth about $1.73bn, having peaked near $2.80bn at the 2 June high.
Is Nokia really leaving China?
Reporting from the South China Morning Post on 18 August 2026, sourced to unnamed people rather than confirmed on the record, said Nokia would close almost all mainland China sites by year end. Nokia’s own Q2 disclosure describes something narrower: integrating a wholly owned Chinese joint venture into its global operating model, with EUR 350m to EUR 400m of charges and EUR 200m of expected savings.
What is AI-RAN and why does it matter to Nokia?
AI-RAN runs radio access network functions and AI workloads on the same accelerated computing platform, letting operators add capacity through software rather than new spectrum. Nokia’s platform has delivered more than 20% spectral efficiency improvement so far, against a launch target of more than 100% by 2028. Twelve operators were in trials as of 16 September 2026.
Why is the analyst range on Nokia so wide?
The 14 analysts polled by S&P Global span $8.50 to $21, a 2.5-times spread. Any Nokia stock prediction inherits that disagreement, and it is about timing rather than technology. Bulls price the EUR 2.8bn AI and Cloud order book converting on schedule; bears price the negative EUR 732m quarterly free cash flow persisting. Both are reading the same Q2 report.
This article is analysis, not investment advice. It does not constitute a recommendation to buy or sell any security. Price levels are the author’s own estimates derived from publicly available data and may prove wrong. Capital is at risk and past performance does not indicate future results.