This thread is a continuation of the comment: Nokia sijoituskohteena (Osa 3) - #11430 käyttäjältä 0-Kelvin.
Previous threads:
This thread is a continuation of the comment: Nokia sijoituskohteena (Osa 3) - #11430 käyttäjältä 0-Kelvin.
Previous threads:
Alrighty, a new thread has been opened,
An interesting thing came across on LinkedIn.
There is speculation in the market about a consortium (likely Microsoft and Meta) developing a standard for slower but wider 25-50G connections using traditional NRZ coding (Non-Return-to-Zero).
Technical challenge: optical complexity: To keep the fiber count under control, multiple wavelengths (WDM) are needed. This requires a lot of power or integrated lasers, which can be a “nightmare” for system reliability.
FibeReality mentions that silicon-based ring modulators and integrated lasers on silicon are challenging due to thermal management and reliability. So, InP (Indium Phosphide) technology is likely needed again, as it is inherently a better material for lasers and optical components than silicon due to its significantly higher heat resistance – plus, it is more energy-efficient.
In that sense, this is again a positive rumor for Nokia, as optical expertise is once again at the center. Both Microsoft and Meta have, in my opinion, been rumored to be Nokia’s (Infinera’s) data center customers, so a link to Nokia would be natural regarding this development as well.
One AI observation:
I looked into Nokia’s (and Infinera’s) recent patent applications. Particularly interesting is a patent dealing with a “multi-channel integrated optical transmitter with passive thermal management”.
- Why is this important? fibeReality mentioned that managing multiple wavelengths (mux/demux) can be a “nightmare” for reliability.
- Nokia’s solution: Nokia has patented a way to integrate these laser wavelengths onto an Indium Phosphide (InP) chip so that they do not require constant electrical heating to remain stable. This removes the “nightmare” that Microsoft and Meta are worried about.
On another forum it was claimed that 1) Nokia is not Ciena and 2) Nokia’s and Ericsson’s stock price curves move quite similarly. Although both claims are factually true (e.g., share price development since the beginning of the year: Nokia +16% and Ericsson +20%), there is reason to examine both matters more closely.
To put it bluntly, one could also say that half a year ago, Ciena wasn’t Ciena either. In six months, its share price skyrocketed 270% from about $90 to $335. Ciena is certainly a pioneer with a strong order backlog, but Nokia has advantages in optical networks such as vertical integration that Ciena does not have. More on the subject at the link.
Let’s look at P/E ratios based on figures from Marketscreener.
Ericsson: 10.6 (2025) 17.4 (2026) 15.9 (2027)
Nokia: 46.4 (2025) 26 (2026) 21.5 (2027)
In other words, Nokia’s current earnings are valued much higher than Ericsson’s, reflecting Nokia’s expected improvement in results. Nokia is now seen as an “AI infrastructure stock” thanks to Infinera and the growth of optical networks, whereas Ericsson is much more dependent than Nokia on the slowly recovering RAN market.
Nokia has also begun to invest more in NI’s (Network Infrastructure) product development: €1,536 million last year (+€329 million increase from the previous year). However, the gross margin, including NI’s R&D investments, has remained quite stable. It was 44.6% (Q4/2025) and 45.4% (Q4/2024), and for the full year 41.1% (2025) and 42.0% (2024). This suggests product competitiveness and pricing power.
The integration of Infinera targets €200 million in annual synergies, but it requires front-loaded investments. A significant cost item is the modernization of the San Jose optoelectronics factory, which is scheduled to be production-ready during 2026. If the project succeeds, production costs for InP (Indium Phosphide) chips will decrease significantly with new automation and larger 6-inch (150 mm) wafers (2.25x the surface area compared to the current 4-inch). The impact on results will be fully reflected only in the 2027 figures.
You could also add the P/B value there. Nokia 1.48, Ericsson 2.74, and Ciena 9.81 (source: MarketScreener). Let’s include some of the AI’s insights in the following. If assessed by P/B, the market does not expect significant growth from Nokia. Otherwise, Nokia’s P/E value could also be around 20; it corresponds to the true value. For instance, Inderes uses it. It reflects “return” more. Below is the AI’s analysis.
Important insight: P/B and P/E together = ROE expectation
There is a simplified connection:
Thinking about this in reverse:
That is, high P/B + high P/E means that the market is pricing in a very high future return on equity.
For Ciena, the market expects:
either a very strong ROE improvement
or a long, rapid period of growth
For Nokia, the market is not pricing in a corresponding improvement.
I am also posting my response to your message from the other thread here.
Historical stumbling is certainly a reality, but the Nokia of 2026 is a different company than the Nokia of 2020. The market hasn’t yet woken up to what Nokia’s technology parity and in-house production mean for the competitiveness of Nokia’s optical networks.
Nokia and Ericsson today announce a landmark collaboration aimed at significantly advancing intelligent automation across purpose-built, cloud RAN, and Open RAN networks.
Nvidia’s bid to open-source 6G may rattle Ericsson and Nokia
https://www.lightreading.com/6g/nvidia-bid-to-open-source-6g-may-rattle-ericsson-and-nokia
I don’t think this press release has been posted here yet. MWC2026 has started and news about the RAN side (MI) might start popping up.
Previously announced operator partners:
- T-Mobile U.S. (Seattle): First “over-the-air” test where Nokia’s Massive MIMO radio and NVIDIA Grace Hopper 200 server simultaneously ran network traffic and heavy AI applications (such as generative AI).
- Indosat (IOH): Southeast Asia’s first AI-RAN-based 5G call.
- SoftBank: Demonstrated how the network’s excess computing capacity can be sold for external AI tasks (monetization).
Nokia has also been able to expand the ecosystem. Quanta and SuperMicro are now mentioned as new names:
- Server vendors: Dell Technologies, Hewlett Packard Enterprise, Quanta (QCT) and SuperMicro.
- Software platforms: Red Hat (OpenShift), which enables cloud-native management.
- Chips: NVIDIA Blackwell-based GPUs (such as RTX PRO Blackwell) are now integrated directly into Nokia’s anyRAN architecture.
And finally, ”Supporting Quotes” statements from the following:
BT, Elisa, NTT DOCOMO, Vodafone..
Ronnie Vasishta, Senior Vice President of Telecommunications at NVIDIA: The telecommunications industry is at a turning point. Instead of upgrading networks in big, hardware-heavy cycles, we now have the opportunity to build them as fully software-driven systems. By running AI and radio access networks on the same accelerated computing platform, we make sure the network supports business needs — not the other way around
Nokia is launching the new generation Doksuri Remote Radio Heads (RRH) product family. These are not just ordinary radio devices; they are specifically designed for the AI-RAN architecture.
- Energy efficiency: Up to 30% better efficiency. This is critical as AI applications increase network power consumption.
- Lightness and speed: The radios are 25% lighter, and a new mounting mechanism reduces installation time by as much as 70%. This saves operators millions in installation costs.
- ReefShark SoC: The radios run on Nokia’s latest ReefShark chip, which brings intelligence directly to the antenna (Edge computing).
So, there are still no GPUs crammed into the radios, just SoCs — as was to be expected.
Pallavi Mahajan, chief technology and AI officer at Nokia, stated during a panel the trial with T-Mobile “is a good proof point of showing the world” where the vendor and ecosystem is headed.
“Now, 2026 is where we will do our first commercial trial, and 2027 is where we will have our first commercial release out,” she said. “This is stuff that actually excites us a lot.”
Nokia explained the integration represents a major advance in evolving the RAN into an AI‑enabled platform which can host third‑party AI workloads, unlocking new service and revenue opportunities beyond traditional connectivity while adhering to AI‑RAN Alliance architectural principles.
In addition to those operators, BT Group, Elisa, NTT Docomo and Vodafone are engaged in trials with Nvidia and Nokia to confirm AI‑RAN’s readiness for real‑world deployments.
https://www.fierce-network.com/wireless/nokia-promises-commercial-ai-ran-2027
CTO Mahajan said commercial AI-RAN would arrive in 2027
If anyone can be bothered to wade through it, Nokia’s MWC 2026 Press and Analyst Event from yesterday.
March
02,
2026
even though interesting news is coming out, there might be good buying opportunities today/this week
Frankfurt (not significant, I know) at the moment:
and for example German futures -2.28%

Hi, it’s big news by the way that Nokia and DT are expanding their collaboration. And especially this next sentence:
Nokia will collaborate with Deutsche Telekom as its strategic co-creation partner for AI‑native RAN (AI-RAN) development.
So, are there small signs here of a situation where Nokia is gaining a first-mover advantage? Nokia’s baseband is thus ahead of many competitors regarding the Nvidia collaboration. Although, of course, GPUs (or Nvidia) weren’t mentioned in the press release. However, it was stated: "As we move toward fully programmable and autonomous networks…)
As grim as the situation in Iran is, I suspect that the network spike caused by (even) that situation will demonstrate the need for additional capacity. The world is moving towards AI - is RAN fully involved in that or not?
That will likely bring additional demand for military networks as well.
Pressi:
https://www.lightreading.com/5g/t-mobile-eager-to-test-nokia-s-ai-ran-this-year
T-Mobile eager to test Nokia’s AI RAN this year
T-Mobile chief technology officer John Saw said he would be “really upset” if Nokia did not deliver an AI RAN radio that he can test in the field in the next 12 months, speaking at Nokia’s press conference on the eve of Mobile World Congress.
Hotard assured Saw that he would “definitely” get his wish for equipment to try out.
According to Nokia’s product timeline, field trials are expected in the fourth quarter this year for 5G software featuring time division duplex massive MIMO, multi-user MIMO and higher order quadrature amplitude modulation.
Commercial 5G offerings are slated to be ready in the fourth quarter of 2027. For 6G products, that would be "just a software update," but Nokia did not provide a date for availability.
This LightReading article had a different writer than usual, which likely explains the pro-Nokia headline. If it had been the usual fare, it could have been something like: “Nokia’s $1B Nvidia rescue bid: Finnish soul has been sold for American silicon and NO real progress”.
Same underlying story, but depending on whether the attitude is positive or negative, you can get many different versions of it. ![]()
- NVIDIA ANNOUNCES STRATEGIC PARTNERSHIP WITH LUMENTUM TO DEVELOP STATE-OF-THE-ART OPTICS TECHNOLOGY
- NVIDIA CORP - IN ADDITION, NVIDIA IS INVESTING $2 BILLION IN LUMENTUM TO SUPPORT RESEARCH AND DEVELOPMENT, FUTURE CAPACITY AND OPERATIONS
- NVIDIA CORP: MULTIYEAR STRATEGIC AGREEMENTS WITH LUMENTUM HOLDINGS
Interesting, Lumentum is strongly involved in the optics game.
And Coherent as well.
“Ouch”
Apparently the same 2 yards to both