AI confirmed the view: Yes — a very credible interpretation is that Nokia’s current AI scale-across architecture thinking is built around the 7220 IXR + SR Linux + coherent-lite layer, and at a later stage, very low-power LPO solutions for shorter optical fabric connections could also be introduced into the same environment.
Furthermore, regarding Nokia’s goal to bring coherent optics to the scale-across level (coherent lite launch H2/27) and the role of the current ICE-D, AI states: Yes — a very credible interpretation is that ICE-D currently serves as Nokia’s ultra-low-power intra-DC optical engine platform (LPO/CPO/retimed optics), while the actual coherent-lite architecture is still being standardized and commercialized over the next 1–2 years. Thus, ICE-D is likely not a replacement for coherent-lite, but rather its potential photonics base or “foundation layer.”
Nokia appears to be building a unified optical stack:
- intra-DC → scale-across → DCI
13 Likes
There hasn’t been enough capacity, and no manufacturer can prepare for it when demand grows by “these kinds of numbers.” It requires major investments in supplier and ODM (Original Design Manufacturer) capacity. It will pay for itself. The biggest problem/risk is getting production up and running in time, and that kind of rapid capacity expansion is only possible in China, or perhaps Vietnam. Well, even US firms have partially abandoned their “out of China” strategies, so I expect this will turn into cash flow soon enough.
10 Likes
Everything positive is now impacting the share price with a 10x multiplier.
Still, it’s worth keeping an eye on the share prices of companies involved in optical transceiver/laser technologies. There are a few others that have seen their prices rise at the same gradient as Nokia. I believe this signals a boost from the AI-optics hype. Similarly, notice the panic among interconnection companies as they snap up small startups developing CPO technology (Co-Packaged Optics), especially lasers. Everyone is terrified of being left behind as copper switches to fiber, and business flows toward players previously known as transceiver and infra firms.
I spoke with an acquaintance in the industry from the copper side, a big player in the field. They have no CPO products, not even the technology, and they’ve been caught with their pants down—panicking over how to build a product. Management had issued orders not to comment on related questions at trade fairs, for example, because investors are currently trying to pry for information solely about CPO to figure out which horse to bet on.
These are the ripple effects of the bubble.
Nokia is in a good “striking position” – to borrow an old sporting term 
14 Likes
The market is currently still not pricing Nokia as a telecommunications network company. I believe you did a sum-of-the-parts calculation yourself a while ago. We are gradually moving towards those part *(fair valuation multiple) prices, but we aren’t there yet, especially within the context of Q1 growth figures.
In Nokia, defense + edge AI + network security, partly data center capacity/vertical integration, and largely the growing volume of data and the transformation of mobile data types (download/upload) as we move towards the AI era, remain perhaps unpriced. There is demand for optical, and Nokia’s full stack is highly competitive, and in the Western world, even a monopoly.
Ciena: “Best optical transport system wins.”
Nokia: “The best integrated transport architecture wins.”
Neither of us would probably be posting here without a significant position in Nokia. You likely know this yourself, but Mobile, along with its various designations, is undergoing a complete restructuring. From a cyclical HW house to a SW house churning out steady, high-margin results. Aligning mobile and optical using AI for the sake of AI.
The little I think about the future beyond the next breakfast certainly hits the right spot in the right way when I contemplate my future ownership. A very, very large amount of unpriced potential. Even on the Mobile side.
14 Likes
Weekend Reading | Up 110% Year-to-Date! The Once-Forgotten Nokia is Returning to the AI Mainstream
The website of Hong Kong-based stockbroker FUTU HK provides an extensive look into Nokia’s transformation. The author ties together Hotard’s appointment, the Infinera acquisition, hyperscaler exposure, the strong growth of AI and cloud orders, and the increasing importance of optical networks in AI clusters. A good point made in the article is that AI infrastructure bottlenecks are no longer just related to GPU chips, but increasingly also to networks and optical connectivity solutions.
8 Likes
I’ll put this here as well:
Nokia is ranked #2-5 in the following applications (excl. China):
- IP core / edge routing (competitors: Cisco, Juniper, Huawei)
- Optical transport (Huawei, Ciena)
- Carrier IP/MPLS (Huawei, Cisco, Juniper)
- DCI (Huawei, Ciena)
- WAN routing (Huawei, Cisco)
- Coherent networking (Huawei, Ciena)
- Mobile (Huawei, Ericsson)
The list is missing players like Coherent and Lumentum. Next year, Nokia will be a direct competitor in component manufacturing for those companies, backed by a strong patent portfolio. I’m not particularly worried even about mobile, as Nokia’s and Ericsson’s paths are clearly diverging. Possibly things agreed upon in a “sauna meeting” (saunassa sovittuja juttuja). There is no sense for either of them to maintain the price levels pushed down by Huawei.
9 Likes