Nokia surpasses 1,000 private network contracts in Q4 2025
https://www.fierce-network.com/wireless/nokia-passes-1000-private-network-contracts-q4
https://www.fierce-network.com/wireless/nokia-passes-1000-private-network-contracts-q4
And Coherent reported its earnings â here are a few highlights. Lumentum rose 7% yesterday, but Coherent fell 8%.
- Massive CPO order: Coherent has received an âexceptionally largeâ order confirmation from the market leader (likely referring to Nvidia) for a CPO solution that includes their new high-power CW laser. Deliveries will begin in late 2026.
- 6-inch InP leadership: Coherent claims to be the worldâs first to have 6-inch InP (Indium Phosphide) production lines operational in Sherman (Texas) and JĂ€rfĂ€lla (Sweden). This quadruples the chip count per wafer and halves costs compared to traditional 3-inch wafers.
- 1.6T and 800G side-by-side: Demand for 1.6T modules is accelerating, but 800G will remain the volume product throughout 2026.
- OCS growth: The order backlog for Optical Circuit Switches (OCS) grew, and they already have over 10 customer projects. The market is estimated to be over $2 billion by the end of the decade.
But Iâm not quite sure about that claim by Coherent.
The G10-AsP will enable Nokia to produce 6-inch Indium Phosphide (InP) wafers, marking a significant milestone in the evolution of Photonic Integrated Circuits (PICs).
The press release mentions, by the way, that the PIC market will grow from $12 billion (2023) to $41 billion (2031).
Nokia made this order in 04/2025, likely for the new San Jose factory. So in that sense, it is possible that Nokia is behind Coherent here since the factory is not yet in production.
Competitors are reporting wins; it would be nice to hear Nokia also explaining in more detail what wins they have had, or if they have any?
It seems a bit like Nvidia is grabbing all the worldâs InP capacity for itself, and Nokia, Coherent, and Lumentum are probably all involved in the Rubin architecture.
@lexus apparently 6⏠is pulling like a magnet in the technical chart? In the US as well, Nokia is once again swimming against the tide and the close was excellent while other tech stocks were mostly falling freely.
Yeah, thatâs where the resistance is, so itâs essentially acting like a magnet trying to push the price lower. The USD-denominated one is likely more dominant, and the resistance zone is 6.80-7.00 USD. If it breaks through (with a close above as well), there would likely be a fairly brisk rally.
To clarify, 5G radio network coverage is ~ 95%. That 2% refers to 5G SA Core, which hasnât progressed at the same pace as the radio network due to 5G NSA. Itâs not a very substantial investment for operators now that the LTE Core is already VNF/CNF. AI could prove to be a significant booster for the transition as the number of AI-supporting terminals increases..
https://x.com/nokiaindustries/status/2019602627902804327?s=20
Nokia and #CominAsia have signed an MoU to deliver a full end-to-end AI on premises #datacenter solution in Southeast Asiaâcombining Comin Asiaâs engineering systems integration expertise with Nokiaâs networking technology to enable secure, scalable #AI for enterprises.
I came across this new analysis.
But Nokia isnât just a telecom equipment vendor anymoreâitâs building an end-to-end AI infrastructure platform that could dominate the network equipment market for hyperscalers and service providers over the next decade.
With Infinera integration progressing ahead of schedule, the NVIDIA partnership announced in Q4, and a backlog approaching EUR 1.1 billion, Nokia is positioning itself as the critical infrastructure provider for the AI supercycle.
The AI & Cloud revenue story is the key unlock. Nokia generated EUR 2.4 billion from AI & Cloud customers in 2025 , representing roughly 30% of Network Infrastructure revenue. This is growing 50%+ annually as hyperscalers build out AI data centers and need optical transport to connect GPU clusters. The gross margins on these deals are superior to traditional telecom sales because hyperscalers value performance and reliability over absolute lowest cost.
Edit: @ruuki I didnât start checking if the figures are correct. They are writing about the right themes, though.
Wait, where is this from?
Didnât Nokia say regarding some NI segment (or the whole NI) that Book-to-Bill > 1?
Taking a bit of a shortcut, I would understand that to be more than the Q4 NI net sales?
Or is that âŹ1.1bn that 30% of Optical net sales? Maybe some shortcuts were taken there too ![]()
Regarding that Nvidia $1.1B backlog, itâs highly likely they are related to Rubinâs upcoming products.
There has been talk about the ICE-D optical interface, but what about this PSE-6e optical processor, which is apparently also required for the GPU to route data? Is the reason Nvidia is already using Nokiaâs Linux networking system to ensure it directly supports the new processor in Rubin later on?
Nvidia is moving from the Blackwell architecture towards the Rubin architecture (estimated 2026â2027), and Nokiaâs optical technology plays a crucial role in this transition.
Here is a comparison of how data transfer challenges change from Blackwell to Rubin and how Nokiaâs solutions (such as ICE-D) help:
Comparison: Nvidia Blackwell vs. Nvidia Rubin (with Nokia technology)
| Feature | Blackwell (2024â2025) | Rubin + Nokia Optical (2026â2027) |
|---|---|---|
| Connectivity Technology | Primarily copper (NVLink Switch) and traditional optics. | CPO (Co-Packaged Optics) and advanced ICE-D optics. |
| Bottleneck | Copper cable length limitations and power consumption. | Conversion of electrical signal to light (I/O energy). |
| Nokiaâs Role | SR Linux software on the Spectrum-X platform. | ICE-D chips directly near the GPU packages. |
| Bandwidth | Approximately 800G â 1.6T per link. | 3.2T â 6.4T+ per link with Nokia technology. |
| Cooling | Liquid cooling is essential. | Optics generate less heat than copper at high speeds. |
Why is Nokiaâs ICE-D the key to the âRubin eraâ?
Nvidiaâs Rubin architecture requires such massive amounts of data that traditional âpluggableâ optical modules (inserted into the switchâs front panel) consume too much electricity and take up too much space.
ICE-D (Intra-Datacenter): This technology allows optics to be brought directly next to the processor. This minimizes electrical loss and enables Nvidiaâs GPU clusters to scale to tens of thousands of units without the network slowing down.
PSE-6s Boost: While Blackwell still relies heavily on Ethernet-based Spectrum-X, in the Rubin era, Nokiaâs PSE-6s and its successors will ensure that connections between data centers keep up with GPU speeds at 1.2 Terabit rates.
Energy Efficiency: On the Rubin platform, the biggest challenge is power availability. Nokiaâs technology promises to reduce the power consumption of optical transmission by up to 40â60%, which is a critical factor for Nvidiaâs customers (such as Microsoft or Meta).
Summary: While Blackwell is still a âhybrid eraâ (copper and traditional optics), for Rubin, Nvidia needs a partner like Nokia to integrate light directly into the computing units. This is why Nvidia is investing in Nokia right now: to ensure that the optical network does not become a bottleneck for Rubin.
The Infinera acquisition was a strategic âhome runâ for Nokia, specifically regarding the Nvidia collaboration and the upcoming Rubin era. It gave Nokia two critical weapons it didnât previously have to the same extent:
1. Access Inside the Data Center (Intra-Data Center)
Before Infinera, Nokia was strong in long-distance backbone networks. Infineraâs ICE-D technology is specifically designed for short distances (under 500mâ2km), where Nvidiaâs GPU clusters communicate with each other.
2. Vertical Integration (In-house chip production)
Infinera brought with it its own Optical Integrated Circuit (PIC) manufacturing capacity in the United States.
3. Strengthening Market Position (United States)
The deal gave Nokia a massive foothold among North American cloud giants (Hyperscalers, such as Meta, Google, and Microsoft).
Iâll try to dig into this more when I have a moment. Here are a few points below.
Iâve started to wonder if Nokia might well be a buyer of Lumentumâs InP (Indium Phosphide) production in 2026, as their own production hasnât ramped up yet. And because of this, Lumentumâs order book wouldnât be full or allocated from the start of 2027. Regarding ICE-D, the InP laser is one area, but the PIC (Photonic Integrated Circuit) component isnât completed by that alone. Lumentum or Coherent probably donât have the full expertise for this, unlike Nokia. And to get the CPO (Co-Packaged Optics) stamp on it, will something more be added on top of Nokiaâs ICE-D..?
I agree that the NVLink switch-level optical solution is likely the key area of the Rubin architecture where Nvidia is dependent on Nokiaâs expertise. Nokiaâs (Infineraâs) ICE-D is likely the technology being applied here. Upscale-level GPU-to-GPU/GPU cluster connections are the most critical in terms of latency and bandwidth, and these apparently cannot be implemented with competitorsâ optical components. However, Coherent will likely supply the optical mirror switch (MEMS) for NVLink. Certainty on this should be gained at Nvidiaâs GTC event in March.
In contrast, outscale-level optical connections, starting from the Spectrum-CPOâs ELS laser source, could likely also be handled with competitorsâ components. Of course, Nokia is likely heavily involved in outscale-level intra-DC and DCI connections as well. The mentioned PSE-6s optical engine is very competitive specifically in coherent DCI connections. A 1.6T coherent pluggable will, however, require further development (product launches at the beginning of the year).
So, just under 30m shares short
On the NYSE on Friday and Monday, it closed above 7.00 USD; what does the next resistance level look like technically?
I started projecting the situation using Fibonacci levels. We bottomed in November, and since then, a new upward wave has been forming. The first phase of this wave was 5.86 USD â 7.09 USD (large chart at the top). And this gives the following ranges for the ongoing wave:
1x first wave 7.29 USD
1.618x first wave 8.05 USD
2.618x first wave 9.28 USD
4.236x first wave 11.27 USD
So potential resistance is already in sight, but I do believe that at a minimum, we are on our way to testing those previous peaks slightly above 8 USD. Then, to my eye, a bullish symmetrical triangle pattern formed there (an upward flagpole with a narrowing triangle at the end), which would give a target at around the 9.28 USD level.
And then in the monthly chart (bottom right) based on the first wave, i.e., 2.94 USD â 5.48 USD⊠the following targets come up for this current wave.
1.618x = 8.11 USD (already reached)
2.618x = 10.65 USD
4.236x = 14.76 USD
This isnât an easy task, however. It looks good, but how far the potential takes us, how the future will be priced in. It also depends a lot on the general market how much enthusiasm will still followâor if it will at all. Iâve been projecting for a while that around the 10 USD level would be a good target for this year, and for the current upward wave. But.
I would like to see something happen on the fundamental side that would generate strong buying interest. Perhaps Nvidia and Rubin? Without something like this, the stock probably wonât get an especially strong upward wave behind it.
Oh right, I donât think volumes are really helpful for the analysis⊠since it has been languishing at a clearly lower level for years.
Found this in the Nordnet comments? :
On Monday, February 9, 2026, Morgan Stanley initiated coverage of Nokia with an âOverweightâ rating and set a price target of USD 8.00 for the stock. This follows an update on January 15, 2026, when the bank added Nokia to its âTop Picksâ list for 2026.
found a more official source: https://www.nasdaq.com/articles/morgan-stanley-initiates-coverage-nokia-oyj-depositary-receipt-nok-overweight
TelefĂłnica today announced a collaboration with Nokia to explore the use of advanced Agentic AI methods in accelerating the adoption of Network APIs.
âTelefĂłnica is committed to driving open innovation in network exposure and automation. By working with Nokia to pilot these advanced AI protocols, we are laying the groundwork for faster and more scalable adoption of Network APIs. Being prepared to expose AI Agents is key to continuing to monetize our capabilities,â said Cayetano Carbajo, Core, Transport and Ecosystem Director at CTIO , speaking on behalf of TelefĂłnica .
https://www.nokia.com/blog/hollow-core-fiber-not-just-for-low-latency/
Optical fiberâs ability to carry petabit-scale data rates over thousands of kilometers at low cost is the foundation of todayâs networks and the global digital economy. However, the requirements of emerging applications are beginning to stress the limits of conventional silica-core fiber (SCF). Hollow-core fiber (HCF) is evolving rapidly and could offer solutions to these problems.
I have a hunch that these have been earned.
Name: Heard, David
Position: Other senior management
Issuer: Nokia CorporationNature of the transaction: RECEIPT OF SHARE REWARD
Volume: 352,752
Edit:
A few more observations about this.
So whatever the case, itâs positive regardless.