Hi Brainchippers,
after all these heated discussions of the last days/weeks I just wanted to join in and summarize my view of the current state of affairs.
I suppose I'm doing this more or less for me personally, as writing it down helps me sorting my thoughts. So here we go:
#1: IP licenses & partnerships:
New IP license agreements obviously are important. If and when they will be announced we will have to see.
I understand the wish for this to happen faster, more often etc. And I assume the reasons why it didn't happen to the extend we all were hoping for until now, might have to do (at least to some extend) with the already much discussed differences between Akida v1.0 vs Akida 1500/v2.0. Whereas the successors seems to have been developed more in line with input from the industry.
So nothing new here - but my personal view is:
It will happen as one or more companies have requested certain changes and they wouldn't request it if they weren't interested in the technology. One could argue now that Brainchip soley acted in this direction as nobody was interested before, but interest is there (at least that was stated multiple times by the company), announcements of partnerships/collaborations did happen, so it sounds more like
"we needed to adjust to better match the requirements of potential customers" to me.
- But as stated - that's just my interpretation. -
Then there's something I see pretty much the same as
@Perhaps, I think partnership announcements with companies like Tata get underestimated by far. Why I think this is not valued enough I'll try to explain below:
The major (IoT) chip-players (ARM, Qualcomm, Nvidia and now also Intel) are currently trying to change parts of their business model.
Previously they just sold their "low end" IoT chips or IP. Revenue was there but if you follow the whole discussion about ARM's valuation you will notice that it's at least interesting that a company that provides a big part of the technology for almost al!l chip-based devices in the world except classic CPU & GPU dosen't generate more money.
What's happening now (and also the german Intel guy - sorry I forgot his name - mentioned it) is - all these companies are trying to get at least a small percentage of their customers' product sales. I think Nvidia and Qualcomm were first, and the automotive industry (and probably) other customers didn't quite like it very much.
So it's not a perfectly safe thing to bet on that this way of trying to improve revenue will work out fine for these chip-companies, rising RISC-V adoption might already be a hint of their customers resistance to it (in addition to avoiding licensing costs in general and other advantages). But I think this will develope in two ways:
Some companies will start to design their own stuff/chips and they still will need some IP for it. Other customers who don't have the knowledge, money or time to develop their own chips will just have to accept to drop some % of their revenue to the chip-companies. You could compare it to the software subscription models. Nobody really likes them but if most of the relevant providers do it you (as a customer) don't have much choice (in the near term). And if it's priced fair, you like the company or their service - it dosen't have to be a negative.
What I was trying to say is: IP licensens are important (and royalties afterwards also of course), but beeing rewarded by product revenue of a partner/customer (e.g. in the medical tech area) is underappreaciated by most of us here in my view.
#2: Announcements & timing:
Of course I also wish Brainchip would announce relevant stuff on the ASX more frequently and regarding the availability of version 2.0 (and hopefully some licence agreements) it couldn't come soon enough.
But what I really don't care about at all is:
Will it be it during the last days of September, somewhen in October or in the beginning of November?
My reasoning is the following:
1) get all bad news/sentiment - company related and macro economically (as far as possible regarding timing) out of the way first
2) get your product out
3) and then keep pushing the newsflow.
So what do we have:
- We have dropped out of the ASX200
- We probaly will have a Q3 report with sub-optimal earnings
- Macro economics are tough but these times seem either almost at the end of the downtrend trend or at least not far away before capitulation (at least stockmarket wise).
- Interest rates won't get higher much more or even start to stagnate before the obvious reduction will kick in (will this happen in the first half of 2024, the second half or in the beginning of 2025? I don't know but it will beginn somewhen and stock market will adjust accordingly in advance)
- August and September seem not to be a perfect timeslot regarding stock markets. If there's a move upwards it doesn't seem to happen during northern hemisphere summer time very often - more likely towards the end of the year.
- Maybe bad performing stock even will be reduced/dropped out of the portfolios before the final quarter (end-of year-rally, if there's one).
If you had the possibility to decide when to announce something important would you have done it during the last two months (August/September) or would you do it - let's say mid of October?
#3: Questions:
Some things I have almost no clue about is this (or at least I'm not yet sure what I would expect to happen as I didn't follow the devlopment of shareholder-structure of tech-startups long enough yet):
If I remember correctly Brainchip has a contract with LDA capital to at least sell a final bulk of shares until years end.
I think it was something around the $10.000.000 mark.
What I don't know is the following:
- Does this have to happen or is it an option?
- Would LDA have to sell the shares to the open market or could they also be sold directly to interested parties?
If I was a big company that recognizes the importance/future potential of Brainchip and want to collaborate for the years to come,
would it be in my interest to at least get some influence on the company so that competitors can't swallow the whole company and I'll get cut off (without my knowledge in advance)?
If I was somehow in the defence sector (company or institution) how would I secure my influence on a potentially relevant tech/IP provider (except the currently obvious limitations of selling hight-tech to China)?
Would it be in Brainchips interest to have some kind of anchor investor (let's say at least a 10-15% stake) or do you think they don't care about that?
#4: PS. - kinda off-topic:
I stumbled upon a new podcast series about neuromorphic computing some days ago which I find quite interesting although it's more about the history, current developments etc. in general:
brainsandmachines.net
Another interesting overview about current developments in AI hardware is the youtube channel "The AI Hardware Show" by Ian Cutress and Sally Ward-Foxton:
Especially the "After Show Podcast" where they discuss or present their opinions regarding each company mentioned during the show, I find quite informative.
Probably you have to take Ian's & Sally's statements as "just another opinion" as they (or at least Ian) seem also to be hired for promoting certain companys (Tenstorrent?) or seem to have good realtionships to some of the who-is-who of the semiconductor world. Nonetheless you might get a good view what major parts of the AI hardware (investors) world is thinking or expecting for the future.
#5: PPS. - some news that recently catched my eye:
Did you see the Prophesee announcement from September 19th (SEMICON Europa 2023, Nov 14–17, 2023, Munich)?
As I'm a interested in photography (the german community might remember an old post of me) I'm awaiting event-based solutions for autofocus-tracking on image-sensors for mirrorless cameras. Imagine one-shot learning applied to autofocus-tracking for your current subject of interest (for photography, video/filming, smartphones, AR/VR ...)
Also interesting although no direct relation to Brainchip (more my interpretaion about the timeframe when AIoT will hit the attention of the average Joe):
AMD Hires Intel Exec Who Led Client AI Team For ‘Meteor Lake’ CPUs
AMD has hired Intel executive John Rayfield, who led low-power AI development for the Core Ultra 'Meteor Lake' CPUs.
www.crn.com
Hopefully see you at the $1 party somewhen next year - Good luck