Huawei, LG Electronics and Nokia Named as Founder Licensors of New Sisvel POS Pool
Share
2 Likes
Marvell can now be added to the list, the same 2 yards as the others. How one wants to interpret this from Nokia’s perspective remains to be seen; there are pros and cons.
Ron’s interview from OFC, some good small pieces of additional information:
https://www.lightreading.com/optical-networking/nokia-s-optical-roadmap-for-the-ai-era
7 Likes
Ron Johnson explained (starting at 9:25) why the product portfolio, to be launched in mid-2027, was announced now. First, customer planning cycles: by announcing specifications now, Nokia can be included in these plans and tailor its offerings to customer needs. In fact, Nokia has already coordinated its roadmap with at least some customers, and Johnson says there’s good visibility (“zero risk”) regarding the schedule.
Second, a supplier signal: Nokia is already communicating its future volume needs to its component suppliers, so that suppliers can commit to increasing capacity in time.
It’s also worth mentioning that, according to Johnson, Nokia’s and Infinera’s strengths were clearly different, so where Nokia was strong, Infinera was not, and vice versa. This made it easy to eliminate the weaker Nokia/Infinera product from the combined portfolio. I got the impression that the integration of the companies has gone well.
And finally, for your information, Ron Johnson’s background:
Nokia
1 year, 1 month
Senior Vice President and General Manager Optical Networking Division at Nokia
Mar 2025 - Present
Infinera
4 years
Senior Vice President and General Manager Optical Systems and Network Solutions Group
Sept 2021 - Sept 2025
Cisco
21 years, 8 months
Senior Director, Product Management and Architecture
Jun 2000 - Sept 2021
Senior Manager, Product Line Management
Jan 2005 - Dec 2010
Manager, Product Line Management
Jun 2000 - Dec 2005
Sales SE
Jan 2000 - Dec 2001
12 Likes
3 Likes
What’s the upside for Nokia in that Marvell deal? All of Nvidia’s investments in the number two spot so far seem to be mainly at Nokia’s expense, at least in Nvidia’s ecosystem, and why not elsewhere now that helicopter money is available. Nokia aims to be an end-to-end provider, but in Nvidia’s vision, Nokia seems to be just a radio provider and the subsequent network parts are sliced up for others, except perhaps data center interconnects. Does Jensen regret that Nokia only agreed to the number one spot? 
It would certainly be healthy if Nvidia had serious competitors. If someone developed into a true full-stack end-to-end network provider, it could even be a risk to Nvidia’s dominance, and perhaps partly for that reason, data transfer is wanted to be built in a fragmented way for several vendors on a leash.
Not quite the triumph that could have been. It’s not a good time to be second or third best in a little bit of everything. The overall solution is perhaps second only to Huawei, and it may be that this will sometimes be decisive. An optimized end-to-end network solution with individual products that reach second place might be better in the future than what is achieved by combining individual solutions from several vendors with their interfaces.
4 Likes
Here’s a “layman’s terms” summary:
Where Nokia could realistically win
If executed well, Nokia could:
-
Become a neutral, open alternative to NVIDIA networking
-
Win with telecom + hyperscaler relationships
-
Lead in AI networking over Ethernet + optics
This is especially relevant as AI clusters scale to hundreds of thousands of GPUs, where networking becomes the bottleneck.
Bottom line
-
Bad for NVIDIA? → Yes, especially for networking margins and lock-in
-
Game over for NVIDIA? → Not even close
-
Strategic shift? → Potentially very important over time
The real outcome:
-
NVIDIA remains dominant in compute
-
Nokia could become a serious player in AI infrastructure plumbing
Yes—a strong, open networking layer (e.g., from Nokia) would make it easier for AMD and Intel to compete with NVIDIA.
Why it helps (very short)
-
Breaks NVIDIA’s GPU + networking bundle
-
Lets hyperscalers pair:
-
AMD GPUs + Ethernet fabric
-
Intel accelerators + Ethernet fabric
-
-
Weakens reliance on NVIDIA InfiniBand / NVLink
Pretty standard stuff. I believe that it would even suit Nvidia’s interests to prevent Nokia from getting involved, but large customers will surely try to get rid of single vendor lock-in with their own solutions. Many have their own network stack (Amazon, for example), but a reliable, neutral alternative with a universal software interface could gain some traction.
From time to time, when negativity strikes, I try to remember that data center builders will surely try to increase competition among suppliers. The current situation, where the GPU vendor might even decide the location of the maintenance staff’s restroom and take their own margins from everything, is perhaps more the spirit of the times than a vision of the future.
Nokia has good products and a strong contribution to product development on the NW side (too). Customers who care about costs (everyone) will competitively bid even screws, rivets, and nails to the very end with different suppliers in construction projects. The same will surely apply to data centers, but my own tea leaves only told me about the future that it’s better to eat those chocolates before the trick-or-treaters start ringing.
2 Likes
AI jargon is abundant on this forum, in particular. It would be desirable to mention at least the first 3 AI sources in connection with these, as ChatGPT and Copilot at least indicate the source material, and Gemini does so upon request.
9 Likes
A few words on this fascinating Nokia story, this time from the perspective of independent analysts: on how the Finnish company is losing some market share to Ericsson in its own 5G core area, but at the same time is cleverly investing in short- and long-term strategies with new AI-powered telecommunications solutions; and how it may have something up its sleeve in the field of AI-RAN technology – for enterprise use.
2 Likes
Sources for me easing my own speculation with AI? I’m guessing there are very few peer-reviewed studies on, for example, Nvidia’s potential desire to maintain its dominant position in data centers. I have read through the generated AI jargon and it aligns with my current understanding.
I can’t be bothered to check how you participate in the discussion, so I’ll leave spam filtering to the moderators, as perhaps you should too. This thread is certainly of a higher quality than KL’s equivalent, but it’s not yet just posting news from reliable™ sources or pure technical jargon.
If there’s a policy that AI cannot be used in a thread about a stock rising due to AI hype, then it will be followed.
Has anyone managed to explain the sharp rise in the US before Easter? The stock price has sometimes seemed to anticipate major news, but I shouldn’t wear a tinfoil hat. If I got a euro every time I said “interesting stock”…
5 Likes
Nokia Technical Analysis April 3. https://stockinvest.us/stock/NOK
1 Like
