Raspberry Pi is preparing for an IPO in London for likely more than $500M
arstechnica.comI'm confused. Weren't they sort of started as an educational/charity/open sort of organization?
It seems to me like if Richard Stallman announced an IPO.
I think this will affect users of the raspberry pi. When it becomes for-profit, we will see prices go up, "open" disappear, decisions become more self-serving.
> When it becomes for-profit, we will see prices go up
The prices already went up. The latest version is expensive.
I mean they already plan to have a Sony-made closed black box "AI" chip that sends telemetry to Sony, and they unapologetically work with ex-surveillance people, they're already there imo.
Can you provide sources on this? I’ve never heard of their AI chip.
Yup: https://interestingengineering.com/innovation/raspberry-pi-a...
This article also mentions transferring metadata to the cloud.
That article seems beyond awful, consistently misspelling raspberry, claiming the 4 is the latest model, and the sentence "According to Sony, it protects privacy by simply transferring metadata to the cloud and conducting all data analysis entirely on-chip" is quite contradictory.
Also, the Sony press release they reference as a source (https://www.prnewswire.com/news-releases/raspberry-pi-receiv...) has 0 mentions of cloud.
> claiming the 4 is the latest model,
It was at the time I believe, since that's a detail Engadget also mentions. In fact here, since you don't like that article: https://www.engadget.com/sony-investment-will-put-ai-chips-i...
> the sentence "According to Sony, it protects privacy by simply transferring metadata to the cloud and conducting all data analysis entirely on-chip" is quite contradictory.
How is that contradictory at all? The chip will indeed be more private than doing AI work in the cloud like most people do, because the actual data work stays on the chip, and only metadata is sent to the cloud. Obviously a really privacy conscious person won't be okay with sending only metadata to the cloud either, but the claim makes perfect sense from the perspective of a big company. It's not clearly written in the original article I provided, but the Engadget article backs up the core claim and phrases it more clearly, so here:
"Sony says it preserves privacy by analyzing data strictly on-chip and only sending metadata to the cloud."
> Also, the Sony press release they reference as a source (https://www.prnewswire.com/news-releases/raspberry-pi-receiv...) has 0 mentions of cloud.
Yes but the [literal official website](https://aitrios-promo.wpp.developer.sony.com/en/services/edg...) for the product Sony is pushing says it:
> Build a privacy conscious system by utilizing edge side analysis of data and images and sending only the metadata output to the cloud.
I probably should've led with better sources for sure, but it was late when I responded and I was tired. I did my research on this thoroughly last year when all this came out originally and I was just trying to pull up something without thinking about it too hard. After all the facts of the case are literally an internet search away for those not too lazy.
I fully understand the financial incentive but I feel it's going to be the bean counters who make calls from now.
Or is it already?
...why?
There are like a zillion Chinese single board computers out there. What's so special about raspberry pi that it's worth a half a billion dollar valuation?
Well, I suppose some of that suggests optimism about selling boards to commercial customers (in spite of all the assurances about still having enough to sell to hobbyists). And commercial customers, just like hobbyists, want SBCs that have decent hw and working, maintained sw. Which is sadly not the case for pretty much any of the competitors, not reliably.
Because they actually provide updates and support to the platform.
All the Chinese boards release a single OS, Linux packages maintained on some strange .cn domain. Something is wrong with the wifi chip driver but they don't care, they've moved on to the next iteration.
This. There are a billion SBCs on the market and most of them are irrelevant because you're stuck with some janky image from an out-of-date kernel that probably even the original company can't recreate from source. Hobbyists who enjoy tinkering with them for a few months after they buy them might not mind, but it's not gonna work for the big buyers and it's not great for educational purposes either.
I'd argue no not if they go public they can't.
Some could say that on a European pov, it's because it's not Chinese.
Nothing. This is an attempt to cash out before the RISC-V stuff makes them irrelevant.
It has been clear for years that Broadcom no longer cares about RPI. Without that subsidy, RPi is dead.
Surely, it is the year of RISC-V.
Well, the fact that both Microchip and Nordic Semiconductor are shipping actual chips based on RISC-V says that if it's not this year, the year is really close.
Actual RISC-V chips have been shipping for half a decade, so the year has been close for 5 years.
Not on Digikey or Mouser, they haven't.
Yes, there have been an assortment of dodgy "RISC-V" chips from dodgy Chinese suppliers that would appear and disappear like spectral invaders from another dimension.
Until last year, though, I couldn't plan around being able to buy and put a RISC-V chip in a product. Now I can.
rp2040 is pretty great
Earlier discussion on the topic: https://news.ycombinator.com/item?id=39187817
Geerlings summary:
[dupe]
More discussion yesterday: https://news.ycombinator.com/item?id=39187817
Is it even possible that they can live up to that evaluation?
Are they just doomed to fail at that point?
$500 million is pretty small IMO. They're probably fairly evaluated there.
Are they roughly 1/100th of a Microchip? Maybe that's a bad comparison. But $500 million is in-and-around the market cap of u-blox. SiTime is $2.5 Billion right now (5x larger than the $500 million IPO). So I think its a fair comparison to make.
Military contracts.
Are they doing a lot of military contracts now?
This seems like a product that people already selling the military would have sourced their parts for ... before this company is even in the running.
Something to be said for using mass-produced, very cheap, locally produced electronics in military stuff. Even moreso when the electronics engineers you hire stand a good chance of owning a few at home.
My experience is that government / military contracts are difficult to deal with / get if you're not established and those synergies like "engineer has some at home" are non existent.
Was more thinking that if an engineer has some at home, they're probably already gonna know how it works versus some arcane chipset and driver set that they need to learn from scratch.
If true, "engineer has some at home" means the home is about to blow up. :(