Didn’t Nokia, Nvidia, and IOH announce that AI-RAN (5G-A) field trials will begin in 2026 in Indonesia? In the lab, spectral efficiency improved by 15-25%, and soon we will see if this is replicated in real-world conditions. Total energy efficiency will likely be of interest as well (consumption from GPU chips vs. dynamic optimization achieved through AI). We will also get to test the implementation of inference, as Nokia and IOH have built a fairly substantial AI grid in Indonesia. Whether the commercialization of AI-RAN takes off in 2027 will likely depend on the results. If opex savings and predictable inference revenue streams exceed capex costs, could this indeed happen?
6 Likes
That may be so. Money can run out or faith can falter. People have a herd mentality – if they believe in something (or don’t), fundamentals don’t necessarily decide the outcome. For example, I never believed in Tesla’s success, and I didn’t see Google’s business model. Apple, Amazon, etc. – I watched them in amazement, feeling like economic fundamentals shouldn’t allow for such things.
Are we involved in the same kind of trend now?
2 Likes
I’ve been thinking about this a lot myself; personally, I don’t pay for any AI services yet, as I utilize my workplace’s services for this and everything else possible. However, I believe that in the future—and it’s already partly the case—these will be comparable to streaming services, of which I think I have 5-6 at the moment. According to ChatGPT, consumer use accounts for about 10-15% of AI companies’ revenue/turnover.
What are your thoughts? Do you already have a paid Pro/Premium service in use, or are you waiting for something like Claude Opus or the next version? Perhaps waiting will cost you as much as jumping on board now. 
Have a great weekend!
2 Likes
Yes, and the scale is larger than it was with Tesla, Apple, or Amazon. It won’t vAnIsh (hAIhdu). It is changing our lives.
Nokia is concrete. It produces hardware seasoned with good services. Right now, that seasoning seems to be better than what the competitors offer.
I criticized the activities at Bell Labs, justifiedly (AIheellisesti) in my opinion, a couple of years ago. Now, that same concentration of PhDs is actually helping Nokia.
I might have smirked at these activities before, but perhaps this will finally turn into something commercializable. There are positive signs.
Of course, I don’t know where this will end up. In my opinion, the insider buying at Nokia was a signal. In fact, all signals suggest that now is not the time to back down.
Nokia is Finnish by heritage. The company has been at the top of technology for a long time. Now it is once again time to show it. I am not an emotional investor, but right now everything feels quite good again.
20 Likes
@OldFeki, you have followed the technology sector for a long time. In your opinion, could this be a case of both a partial bubble and a situation from which Nokia still benefits in the short, medium, and long term? These are not necessarily mutually exclusive in a technological disruption where different players can have very different roles.
2 Likes
It would be a pretty crazy scenario if everyone just followed suit without a clear destination. I believe everyone has their own vision and faith in future profitability, as investment decisions have already been made. Nokia has a strong position as an infrastructure provider. The worst risk is that one out of a hundred customers goes bankrupt.
I wrote earlier about how data center infrastructure needs to be updated more frequently than mobile networks. Constant operating profit is guaranteed as long as there is a real foot in the door in the data center business with a large product family. Data centers aren’t going anywhere, as AI is a genuinely useful tool.
8 Likes
While it’s true that CPUs/GPUs and other rapidly evolving product categories are replaced often, the building structures and cooling systems are 30–50 year investments. Nokia’s products likely have a lifespan of 10–15 years before the next generation replaces them?
I don’t have one; I use free services and at work. I’m thinking that for consumers, the business model might shift toward being ad-based, similar to Google? But for companies—for their management, innovation, and their smart products—this will be a lifeline in the competition.
2 Likes
Nokia promises operators HW/SW support for 10 years from the last delivery. This leads to base station sites typically having products from several different generations, ranging from 0-15 years old. When a site gets, for example, a new NVIDIA baseband along with 6G, it is typically limited to 1-2 frequencies, i.e., 3-6 radios. An urban site may have 6-7 frequencies, with the majority using passive antennas for 4G and 5G.
2 Likes
3 Likes
Over on the Kauppalehti forum, there are also attempts to curb the apparent euphoria… Below is my response to a commenter.
saita-itara said:
No one actually knows where the share price will be taken next. This doesn’t even depend solely on Nokia, but more on the entire AI sector, which Nokia has been included in, even though their data center business is small and will remain small for the next few years.
Nokia had a billion euros worth of AI and cloud orders in Q1. Relative to sales, these orders were approximately 20% of Nokia’s Q1 sales, which totaled €4,497 million. We don’t know what the AI and cloud order intake for the full year will be, but Q1 speaks a very positive language. It should be noted that according to Hotard, it is only in Q2 that AI and cloud orders for IP networks will start coming in more significantly, whereas until now, it has primarily been optical networks receiving them.
At the same time, it is worth noting that Nokia’s relevant addressable market related to AI and cloud business is growing at 27% per year, which mathematically means the addressable market will double in three years.
5 Likes