Because despite the a slight shift in political tone, making profits on HN is still hugely unpopular. Even if it is a for profit under a Non-profit Foundation.
As somebody with very little understanding of the business world: why would a seemingly successful organisation like this become a public company? What benefits are there?
It allows the company to raise money and existing shareholders to cash out or at least to turn their shares into liquid ones (that they can sell easily), with a "real" market value.
In this case, the article says that they've raised $200 million, which means they sold new shares as part of the IPO.
I wonder what the impact would be to us (consumers) if there would be pressure from shareholders to increase profits. Will future Raspberry Pies be completely bare-bones unless you pay a hefty premium?
Either prices start to increase, product quality lowers or the end-users will become the product. Or maybe all of them. Or maybe nothing changes, but unfortunately history proves otherwise.
Funding is one reason but the real one a lot of times is you hit hard legal limits on how many people are on your cap table before it’s just easier to go public. Meaning if you want to keep giving equity to employees you are almost forced to go public.
Name recognition? You know exactly what you get, which isn't necessarily true for the bajilion clones out there. You know that if you buy an rPi, you can buy a replacement tomorrow and it will work the same. In a world where I can't even rebuy a laptop of the same make and model a few months later that's quite the advantage.
There are a hundred outfits out of Shenzen doing what they do, but Raspberry Pi's move a lot of units because they have backwards compatibility, good support and that has fostered a healthy ecosystem over the years. If even a single Chinese outfit got the memo, they'd be out of business quickly - but they don't.
I don't understand why customer support is so hard to get right for these Chinese companies. The pervasive mindset seems to be beating others on price is enough. For the maker market Pi's are aiming for, it really isn't.
For the likes of us there isn’t really much of a moat but RPi puts a lot of effort into accessibility for beginners. Look at orangepi.org:
“We offer one-stop customized service based on open source products to shorten the path from technology to application”
vs raspberrypi.org:
“Empowering young people to use computing technologies to shape the world”
The underlying tech is very similar but the marketing and community aims are very different. If RPi can capture a solid chunk of the education market they’ll be able to dine out on that, and it doesn’t seem like OPi is equipped to do the same.
Raspberry's moat largely comes to support and compatibility. They've fostered such an ecosystem around their SBCs that they're usually the safer choice if you want to build something with any expectation of long term stability.
Now that they're public, there's no guarantee that they'll maintain this commitment. They may feel pressure to sacrifice long term success for short term profitability. We'll have to see how this plays out.
Maybe, but I think part of the fear behind the IPO is that some bully type investor like Icahn will do a takeover/weigh-in and mess it up, towards a profit driven enterprise.
With major ownership held by someone else, this becomes less likely.
> second-hand 1L pc's (tinyminimicro) is a much better value-for-money.
For initial purchase they certainly offer much better bang-for-buck compared to the Pi4 & Pi5, but do you have any useful links to information about relative power consumption when mostly idle?
I currently have a Pi400 (back when I bought that the Pi proper was out of stock everywhere except the scalpers charging >£150/unit for even the low RAM versions, but I found a decent deal on a Pi400 kit) running as a router/firewall with a few other small bits & bobs in containers. I keep considering replacing it with a small PC like that – my main concern being running a core bit of the home network off an SSD, though also having a bit more performance (both IO and CPU) might at times be useful too, and the Pi400 I'd like to reuse somewhere it is visible for “this is what home PCs used to be like when I were a lad” nostalgia reasons. The reasons I haven't so far (other than momentum: the 400 is sat there working quietly needing no interaction aside from the occasional update and paranoia checks of the config/log/other backups) are possible greater electricity use, and great noise if the mini-PC is not entirely passively cooled.
So just the CPU at idle pulling more than twice the Pi, about the same as a Pi4 running a CPU stress-test, depending on what figures you take.
I had a look around and there are many values out there, just under 3W idle and about 7W spinning the CPU seem like reasonable values (to get better data I might have to try properly testing mine – I've got an overly fancy USB3 cable that displays the power passing through, so I could do that at some point, at least approximately).
Figures for mini PCs obviously vary by model, but also you see a lot of different figures for the same model (some of the difference could be due to differences in storage power draw, some don't say if they have an SSD or traditional disk, and some could be differences in ambient heat affecting draw for cooling? age may also have an effect on the efficiency of fan motors) but between 9W and 15W (mostly below 12) at idle feels like a reasonable guess based on those quick searches. This is lower than I was expecting. The official rating for an Optiplex 3070 (picking a random model from the popular ones) is “up to 1.2W” which is far lower than people have measured in their own tests so I'll not be trusting official values.
Of course the PiZero or similar would draw less than anything else I've already mentioned, though to be a router I'd need to add at least some NICs to that, some extra storage (given one of my concerns is running the router off an SD card), and would lose some CPU resource (the Zero 2 is roughly equivalent to a Pi3 there?) and a goodly chunk of RAM.
I dumped them after they fought against competition by selling very limited stocks of the Zero at very low prices, possibly at a loss. They released just enough of them so that someone somewhere could always tell that they were the cheapest board available, but in fact they were pretty much unobtanium everywhere, except of course in "friendly" sites and forums where every post about the shortage was immediately replied with "I could buy one at XYZ...", where xyz inevitably turned out to be a shop with none in stock. This happened much before Covid19 and chips shortage hit the market, in fact competition had cheap alternatives in stock and I could buy some NanoPi and OrangePi, but the RPi had much better press. I'm considering to make an exception for the Pico for being a really interesting product, but my last Linux RPi has been the 4 and I don't plan to go further.
Before the pandemic, I picked them up for $5 regularly at our local friendly computer store. They even had 0w's for $5 as a gimmick (they were supposed to be $10) with purchase limit of 2 or 5 (can't remember). I'd just pick up a couple every time I went to that store. I've got a sandwich bag full of them now.
I think it was very much the supply chain snags of pandemic that made them get weird high prices everywhere.
So your assertion is that Raspberry Pi was using a Zero as a loss leader AND that it was out of stock. And some how this was enough to stifle competing products? And the competing products you've cited all have "Pi" in the name because they only exist to fill the gaps that Raspberry Pi products don't fill. For example the NanoPi is a Zero competitor that differentiates itself by having Ethernet.
Raspberry Pi dominates the market because it was the first to enter it and provides a limited product catalog that has limited churn with comprehensive support and regular software releases. The same cannot be said for it's competition which historically offered horrible software support land quickly iterated on products meaning that support generally only existed for the very latest.
I know it's an anecdote, but I found out about them bundling the Zero with a Raspberry Pi magazine here in the UK few days after it launched, and I was able to pop down to my local tesco and grab a copy(and it wasn't the last one either). Since then I bought a dozen more for various projects around the house and it's never been a problem, Pimoroni or the Pi Hut always have some.
The "IN" probably refers to in one of the camera modules, or some other module/HAT.
Raspberry Pis don't use any Sony chips on the SBC, Sony image sensors are on the camera modules, and since they're talking about imaging, this makes perfect sense.
Nothing about the Engadget article says to me it'll be on board the Raspberry Pi itself.
The original press releases don't say anything about incorporating it into the base Pi hardware and I don't think they would, given that it would add extra cost and most people don't need it.
You’re assuming that without a patent, the RP2040 wouldn’t immediately have counterfeits flooding the market without any hope of legal action. This has happened with the ATMegas, this has happened with the ESP32s, this has happened with the STM32s, it’s a valid concern. You’re also assuming there wouldn’t be legal-ish pseudo-compatible clones, like the GD32 clone series of the STM32.
> the RP2040 wouldn’t immediately have counterfeits flooding the market without any hope of legal action
That would be great. Here is the mission statement straight from raspberrypi.org:
> The Raspberry Pi Foundation is a UK-based charity with the mission to enable young people to realise their full potential through the power of computing and digital technologies.
Cheaper clones from China would aid that goal, but of course, it flies in the face of a fake non-profit company.
And if you are only interested in preventing counterfeits, just slap the trademarked Raspberry logo on there. No need to inhibit progress with patents.
No, it would be terrible, as technology development is not free. Chinese clones are the epitome of your “enshittification.” They drive prices up for the real product and invade the market with garbage.
This has most recently happened in the 3D printing world, with Prusa versus BambuLab. Who actually develops an open source slicer? Who allows 3rd-party firmware? Who contributes to the community? Who abides by open source licenses? Hint: It is not the Chinese company.
But at least that GPL-violating, closed-source printer was cheaper.
But it was never Prusa's declared mission to get as many people into 3d printing as possible? So I don't see how this is relevant when comparing it to RPi's mission.
> Chinese clones are the epitome of your “enshittification.” They drive prices up for the real product and invade the market with garbage.
The opposite, having no clones makes it easier for a group (like RaspberryPi) to enshittify.
Enshittification is where a group first obtains a large market share with cheap/free services and then pivots to squeeze as much out as possible. Having a competent clone is a strong preventative.
> This has most recently happened in the 3D printing world, with Prusa versus BambuLab. Who actually develops an open source slicer? Who allows 3rd-party firmware? Who contributes to the community? Who abides by open source licenses? Hint: It is not the Chinese company.
Bambu pisses me off too.
Unfortunately your parent is talking about patents, not open source vs closed source, or license violations.
> No, it would be terrible, as technology development is not free.
What about compatibility? Wouldn't it be good for competitors to be able to provide compatible PIO interfaces, so customers can churn from one SBC to another SBC without needing to rewrite their code?
> Having a competent clone is a strong preventative.
This causes a similar problem: people don't buy from the innovator, who needs more money to carry on innovating. They buy from the clone, who mostly only copies the innovator.
I assume that their mission being "accessibility of computing to all", and the org communicating so much on their mission focus, they would be in favor of cheap clones.
I expect most won't care or even notice. Company ownership is probably not that important to the average consumer. They care about great hardware support and a good "ecosystem" of accessories.
> We’re on a mission to put high-performance, low-cost, general-purpose computing platforms in the hands of enthusiasts and engineers all over the world. (https://www.raspberrypi.com/about/)
So what do we believe marketing materials now? Next up: Google is organising the world's information, NOT selling ads! Raspberry Pi used to be a platform for education first and foremost, but you don't get very far with this so they took on some funding to get ahead. Anti-VC whining is quite tangential to the product itself.
> But isn't this a product aimed at hackers and tinkerers?
It was, but The appeal spread outside of that, which is why they are now going public and abandoning the precious niche mission and are instead aiming at mass market appeal.
My understanding that as of a couple of years ago most of the “clones” had various issues from pcb layout bugs causing ddram access issues, to poor documentation and poor Linux kernel support. Also availability was often a challenge. Getting a product that just works and is well supported by a community is worth a lot imo. Especially for hobbyist, less so if it’s one’s full time job, but then why are you using these things?
My impression is that the coming generation of chips are going to be completely integrated RiSCV chips with RAM integrated on one package. Akin to a microcontroller but that can run Linux.
I got an intel SBC for something that required non-ARM and I remember the power raw being surprising after having been spoiled by so many raspberry ARMs.
I've been looking into alternatives to the Zero 2W because I need something in the same form factor with more (processing) power; something closer to the Pi 3B+ or better.
My conclusion is that although there is some nice hardware out there on paper, the poor circuit designs and atrocious software support mean anything I can get that meets my requirements will be an absolute nightmare; and I'm a software engineer of 1.5-decades with an Electronic Engineering degree, so the bar is high for something to be too-much-effort.
Take the Radxa Zero 3W[0] for example; the WiFi doesn't work properly, GPU acceleration doesn't always work, the OS images are out of date and don't always boot, when it does there are power issues, thermal issues, performance issues. No Debian 12 expected until H2 this year, if lucky; Android OS images are Android 9 (we're on what Android 14-15 now?).
This “a company is great until it goes public, then it is guaranteed to be crappy” mentality only displays a lack of understanding of how ownership structures work. It is more nuanced than what you are making out. An incredibly over-simplified view bred by the knowitall “fuck MBAs” attitude on HN lately.
I love HN, but this is insufferable behavior. I also believe it’s why, despite our screeching, we accomplish absolutely nothing. You can’t accomplish anything when you’ve burned all bridges and then yelled “screw you.”
It may seem dramatic to worry about the future of Raspberry Pi now that it's public, but I personally have seen enough mission driven companies go public, and have observed what happens when profit becomes the primary goal. You'll have to forgive me for seeing this as the beginning of the end of a beloved hobbyist SBC company. I expect the next few generations to be just fine, but eventually shareholders will start to squeeze, and the "profit above all else" mentality infecting public companies will set in, and then the enshittification begins.
I'll be looking around at other SBCs, but I'm hapoy enough with my current RPi boards, just no longer confident I'll be happy with then in the future.
An IPO is a cashout. It's almost always certain that there are many key players in the company who had stake that no longer do. This is not a "fuck MBAs" thing. It is much harder to keep a business aligned with its initial purpose (and even customer interests) when it is public.
Tech itself has become more negative lately, for a number of reasons. HN remains a marvelous place - I was just last night enjoying the variety of posts and discussion. I’ve not found anything like it elsewhere.
In many cases this sorts itself out in a few hours as downvotes rank the page. Luckily here I suspect the flamewar detector has kicked in and the page is being downranked.
FWIW I don't find this in other tech spaces though they usually have a lot more young people. The media coverage of tech has certainly become focused on negatives, but if you hang out in Discord channels or talk with random Bluesky/X builders (keyword is builders here) there's definitely a culture of just talking about building or the more "financial" market analysis aspect. I agree with GP that HN has hit a pretty strong downward trajectory at this point that probably disincentivizes more rational commentary on topics that HN gets a lot of hate on.
Do they not have a fiduciary responsibility to their public shareholders now that they did not before?
If a random company with a very different mission (say, Intel) comes along and offers to purchase the company at a decent share price premium they pretty much have to take it, right?
Everyone repeating the mantra about how caring for shareholder profits is the fiduciary duty of the corporation leadership, and how it's "a law".
At this point it's a self-propagating meme. I frequently encounter people, learned people, who act like it is a law or moral principle (in fact, corporations aren't even required to be for profit even when operating commercially, they are just an organizational structure). When questioned, nobody can cleanly point where they learnt it, but actual check shows that Friedman Doctrine[1] was never written into law.
Ah - I see. I don't think that's marketing; just a common misconception. But yes, I've seen that too. No idea if it's true or not. Presumably it changes depending on the country.
That said, the misconception can be tracked down to one person in 1980... that sounds more like an idea that got marketed - whether by Friedman teaching in university in certain way, or his students continuing that, or people taking the idea from his paper and teaching that... or newspapers passing it further...
I never get why people think this does exist. If it did, wouldn’t every company be obliged to get out of the business they’re in and start doing what would be most profitable right now? Obviously that would be absurd.
Just as an interesting tidbit while we here are talking about UK companies it doesn't even exist in the US. The SCOTUS said so:
> While it is certainly true that a central objective of for-profit corporations is to make money, modern corporate law does not require for-profit corporations to pursue profit at the expense of everything else, and many do not do so.
What are you talking about, you can easily get all models same or next day at typical reseller markup. I keep seeing these comments but not experience this problem. So what gives? #fakenews
And children should learn to program edge devices, they are entering a world flooded with them...
> And children should learn to program edge devices, they are entering a world flooded with them...
Which means little unless laws substantially change to allow them to do anything with them. Right now, edge devices are protected from "tampering" and "unauthorised use" by both technical and legal means.
Haha. If you can code then you can create your own devices from scratch. People won't whine about a locked down cloud-only doorbell... they'll make their own.
Sorry, I'm depressed about SaaS, IoT, secure computing and remote attestation. And everything being rented instead of owned - from the edge devices in question to apartments they're installed in. Good luck with parents letting kids hack on stuff when it violates multiple lease agreements.
I just recently got my first RPi for my home automation needs, so I have no emotional attachment to the RPi ecosystem of the past. Frankly I don't get your rant at all, nor do I see any reason to reassess the approach.
My apartment has old radiators with manual temperature control, so my goal is to automate that with some temperature sensors + Zigbee radiator controllers. And liberating my Hue lights as they keep nagging that I'll soon need a cloud account to control my lights at home.
Nice I've been contemplating doing some home automation for heating, I have a dual fuel forced air furnaces, with only 1 thermostat and a fairly large house. There's a large temperature gradient between the floors, so I was thinking automate booster fans on the lower floor to pull more heat, and or cycle the recirculate fan when the temp gradient is above X.
> thanks to every US public school deciding the entirety of their pupils must code --and do so on a pi-- the average price surged in the states to nearly $450 a unit
Wow. Messages from an alternate reality. There are at least three assertions in the above message which are untrue.
Anecdotally, my local school district decided to do what the GP post says, and they're the last district I would expect to do that. I've never seen $450/unit, though.
What your critique does not address is the hardware aspect of the pi. GPIO pins engage many people who are interested in seeing tangible results of programming.
For most hardware projects I reach for Arduino first. They are a lot cheaper (especially the knock-offs), actually in stock, and easier to reason about (fewer layers involved in the software stack, more direct hardware access)
Raspberry Pis are great if you want easy internet connectivity in your hardware project, or need an OS or lots of computing power.
If they were readily available and $20 each I'd grant them that they are more convenient. But at real prices and availability that's more difficult to justify
Yeah, honestly I have become pretty disillusioned with the entire ARM SBC market.
For awhile they were super cheap and so it was easier to justify, but when they’re getting into $200+, at some point I feel like you’re going to get a better deal with a mini x86_64 PC, with considerably better performance (if not per watt, at least total performance).
I had some used thin clients with AMD CPUs a few years ago. They were $50 each, and only ate like 10W of power, even under load. No, there wasn’t GPIO built in, but even three years ago there was USB 3.0 and it blew the Raspberry Pi out of the water at the time.
I was running a full rack mount server for a few years after that but since cheap mini gaming PCs have been coming out on Amazon semi recently, I bought one of those to save power usage, and I have extremely happy with them. They have a GPU where I have been able to VAAPI transcoding fully working and they have a considerably faster CPU, and since RPis have gotten so expensive with spiked demand, it wasn’t even that much more expensive for a home lab.
I too worry about enshittification, but let's step back here. When you're talking about a price difference that can be made up in a few hours of work, the extra cost can be easily justified, because of the Raspberry Pi's secret sauce.
The secret sauce allowed me to search the web for guides on what I wanted to do and have a template to work from. Not having to spend potentially hours or days of my free time figuring out some obscure Linux issue or driver support problem made up for the extra cost of the board.
The secret sauce was always the documentation and software support, and if the cheap clones can master that, then the boys in England have a real problem, but that remains to be seen. A lot of that comes not from cost savings inherent to microcomputer technology, but because of the dedication of people who know how to communicate to software engineers and hobbyists.
My hope is that most of the price increases apply more to their industrial computing products than their hobbyist ones.
It sounds like you would be okay with a Pi clone though?
Like the OP, I have avoided the clones in the past because I wanted to reward the company that were "cause-driven" to do this. I too am less inclined now though.
That depends. Is "clone" a synonym for "compatible"?
If so, sure.
My concern is, I don't want to spend any time whatsoever messing with stuff like driver or architecture issues. Needing a few repositories for Debian packages is okay, but pushing it. I'm not here to configure a system; I'm here to work on a project I find fun or useful.
Ideally I take the latest RaspberryOS and slap it in and it "just works".
I buy Pis because of their support, stability and out of the box experience is far superior to other cheaper clones. I have tons of cards which are paperweights without Armbian or self-rolled distributions, and even them some are feature-capped because of kernel-dependent closed source modules or out-of-tree drivers.
I can still download and install a recent version of RaspberryPi OS to my Raspi 1, customize the image even before it touches the SD card, and remove the card and install to a more recent Pi, and continue working.
More importantly, I can upgrade OS from version to version (Armbian doesn't support this, and they say it explicitly for example).
OrangePi, Radxa and Asus failed on all these fronts. I still peek and poke OrangePi 5B's Debian image since I can't trust them blindly, but can't use its AI and video (rendering/encoding) accelerators without its official image.
Oh, and my OrangePi Zero always crashes within a day or two if I don't practically disable its frequency scaling capability via governor. Even a good cable and adapter which can run a Raspi 3/4 without any hiccups can't make this postage stamp sized thing happy and stable.
P.S.: For everyone who thinks RaspberryPi brand and hardware will be enshittified: I'm not saying that it won't happen. What I'm saying is, let's watch how this moat evaporates, and who passes (said moat) and replaces them on their hill.
If that happens, I won't buy them anymore. The boards don't have an expiration date, and I can roll my own distribution "if it comes to that". Have done that in a professional capacity back then, can do it again.
You're right, but I refuse to keep my mind busy with it. I'll look around when I need a new SBC for the task at hand, and if RaspberryPi has enshittified, I'll just pass.
So it's an if from my perspective, since I don't constantly watch them and ask "aretheyenshittifiedyet?".
Support and stability will go down, they will cut support roles to drive profitability and they will release multiple variants of the same board to saturate the market so hardware support will get dicier. They may even roll their own distro as a way to capture the market.
When a company lists they switch from being customer focused to being share-holder focused. They will do anything and everything to make number go up.
Let's watch as it happens, shall we? I'm not saying it won't happen, yet I don't feel gloomy and dark about it. If they fall, another will rise and fill the void.
The market is healthy from that standpoint, but let's watch as it happens and see what happens.
Similarly my Orange Pi 4 (not the 4B or 4 LTS) no longer gets new images from Orange Pi themselves. Armbian's images had an issue with display out on the board for at least a year. I haven't checked if it's been fixed. And that's with a RK3399 which is basically mainlined in the Linux kernel.
I use the Pis because the ecosystem and documentation. I tried using clones but they're often horrible when it comes anything outside the direct hardware. Also they have a ton of competition. Jumping back to STM32 is pretty trivial if you use open toolchains.
Are you also inclined to support and develop kernel and OS level software for them? Because the lack of good support is why they remain woefully noncompetitive despite the lower price.
And once they make one that isn't slower than a Pi 3 I might actually buy it. Their entire lineup is obsolete, they're selling 2GB models like it's 2018 while Orange and Rock Pi are moving onto 16 and 32GB.
The only Raspberry Pi models I regret buying are the sub-4GB versions, since you're always one step away from needing swap.
The Raspberry Pi Foundation put out a statement[1] on the IPO and how it relates to the mission.
In 30 words,
instead of receiving a share of the company’s profits each year, we will convert some of our shareholding into an endowment that we will use to fund our educational programmes.
While it remains to be seen how the IPO will affect future Raspberry Pi products, it's hard for me to argue against the Foundation diversifying its investments, especially in light of growing competition from cheap clones.
The words “public company” often carry a negative tone but ownership rules are complicated. It doesn’t mean much until you know the specifics. For example a non-tech example of a publicly owned corporation that people might not think of is the Green Bay Packers.
Another tech example of a charitable foundation funded by for-profit subsidiaries (both wholly-owned and publicly traded) is the Carl Zeiss Foundation[1].
This model has been a mixed success for Mozilla and OpenAI.
The TechCrunch article is speaking only of it like any other commercial investment opportunity.
Where's a good writeup of charity/non-profit mission concerns about this commercial subsidiary?
(Example from early on in Raspberry Pi history: There were concerns about pragmatic compromises of open source or libre software principles that the RasPi made, so that they could use the closed-software Broadcom SoC, presumably to deliver an otherwise good education-oriented SBC at a price point that made it accessible and got uptake.)
This is high time for a group of individuals to bring a RISC-V SBC with a tweaked Linux distribution along with it to fill the gap and... later get to IPO.
RISC-V is still way too slow, and even then, there’s this stupid misunderstanding about RISC-V and open source.
RISC-V means only that the instruction set is open source.
It does NOT mean:
- There are open source drivers
- The chip design is open source
- The chip design is not patented
- The software is not patented
- The chip has no proprietary additional instructions
The NVIDIA firmware blob and drivers could run on RISC-V right now with no practical benefit to you. The ISA being open source does mean that open source CPU cores can exist, but no chip is required to use them.
Being an open ISA is still hugely important though, and a big improvement over the status quo. Don't let perfect be the enemy of good.
The ISA is arguably the most important thing to have open, and the hardest to get open since it is the thing that requires the most coordination between so many disparate and disinterested entities.
Today we celebrate an open ISA, tomorrow, being enabled by the open ISA, we'll be able to celebrate the next iteration of open, until at some point we get fully open.
Yeah, that's the whole point. Have your own CPU cores. As for slow, Rasberry Pi class of machines are not meant to compete with Intel/AMD server grade. That's for enthusiasts (with even if 40nm process) and has to be cheap.
All drivers are not needed, nobody is expecting to do triple 5k displays at 120FPS. Basic WLAN, Ethernet, Bluetooth and basic display capabilities around HD even at 30FPS are good enough for that market.
As for open source, you as a computer architect, can't even make your own SoC that is on x86 or ARM without paying royalties and spending lots of money upfront.
Not so with RISC-V so I guess you'll see lot more RISC-V in future.
This time - 20204 is the 1999 of RISC-V if you want to borrow a precedent from Linux's rise as defacto OS on server side.
arduino is the other name in this space + has been making interesting moves; their industrial / PLC offering has gotten very serious in the last few years and they are ahead of pi on the cloud front (though open question whether either of them will do well in cloud; their customers are too savvy + nimble for a lock-in play)
arduino is approaching the business bottom up (tiny well-known chips), pi top-down (it's just linux with some GPIO pins)
also have competitors at the margin like nvidia (who could legit own 'iot edge AI' if that ever even matters) + waveshare, the off-brand pi manufacturer who remarkably has way more pi-compatible devices than actual pi
excited for IPO if it means the #1 linux board finally supports suspend-to-ram
You’re describing a product with well known stock shortages that have, in floating, unlocked a bunch of capital which can among other things help them scale their production to meet demand.
That sounds like a good thing both as a shareholder and as a consumer.
How am I supposed to know if something is worth investing in if I can’t even use it? A pinky promise that maybe I’ll be able to get one someday isn't a good deal. I want to know what I am getting when I put my money in. Maybe if I was a VC and that’s what they were begging for my money for, but as a retail investor the value proposition is basically nothing.
The company markets itself as an educational tool, but 70% of its sales are industrial? Like I get “bringing low cost computers to everyone” feels and sounds good in press releases but when the business actually focuses on something different - that seems counterproductive to their stated goal - what am I supposed to take from that?
I am concerned that ultimately whatever supply chain and manufacturing improvements they see from raising this capital will benefit their commercial/industrial customers more than their non-profit/educational/enthusiast customers. I'm not asserting that will happen, but I'm concerned it will.
Where I am it has been next to impossible to order a non-kit model of Pi released within the past several years since at least mid-pandemic.
I started buying up SFF Lenovo boxes on eBay, which I've been somewhat happier with since I never used some of the Pi's standout features like GPIO.
I never quite got the appeal of Rpi's. When I needed a silent low power home server box 10 years ago I went with Intel Atom x86 based board which supported virtualization and had expandable memory. When I needed a small low power box for and observer node this year I bought a Dell Wyse 3040 for 40€, also x86. If I wanted a general purpose PC I'd still go with a NuC or Minisforum these days.
It's marketing. For someone like me with little familiarity of the mini computer and microcontroller world, I know about raspberry pi and arduino. I know there are many others out there with likely better value and quality, but the perception of accessibility to a newbie drops off of a cliff.
Yes, but there seems to be a pretty reliable supply of used thin clients. I have no idea who still buys new thin clients, but whoever it is must replace them regularly.
For a lot of stuff it just comes down to convenience, there's community projects that are primarily distributed as raspi sd card images (home assistant and octoprint come to mind).
238 comments
[ 3.4 ms ] story [ 256 ms ] threadI own quite a few Pis (although mini PCs have largely replaced them in my homelab), but I’m not sure why we should be expecting major growth.
> Raspberry Pi Ltd, the public company, is the commercial subsidiary of the Raspberry Pi Foundation.
That’s insane
In this case, the article says that they've raised $200 million, which means they sold new shares as part of the IPO.
But if they company is very profitable already, this should not be necessary. They could still grow with complete mission-oriented way.
They certainly have an edge on ecosystem and software polish but it’s pretty slim.
I’ve got a mixed k8s cluster - half rasps half orange pi…and I’m having more issues on the rasp side.
I don't understand why customer support is so hard to get right for these Chinese companies. The pervasive mindset seems to be beating others on price is enough. For the maker market Pi's are aiming for, it really isn't.
“We offer one-stop customized service based on open source products to shorten the path from technology to application”
vs raspberrypi.org:
“Empowering young people to use computing technologies to shape the world”
The underlying tech is very similar but the marketing and community aims are very different. If RPi can capture a solid chunk of the education market they’ll be able to dine out on that, and it doesn’t seem like OPi is equipped to do the same.
They don't have TLS enabled. Doesn't exactly promise much on the software side of the board either.
Raspberry's moat largely comes to support and compatibility. They've fostered such an ecosystem around their SBCs that they're usually the safer choice if you want to build something with any expectation of long term stability.
Now that they're public, there's no guarantee that they'll maintain this commitment. They may feel pressure to sacrifice long term success for short term profitability. We'll have to see how this plays out.
Like nvidia.
Now that they are in stock-price-go-up mode, I am inclined to buy and promote cheap clones, and I expect most will feel the same.
Turns out that doesn't mean shit.
Definitely won't be abused /s
With major ownership held by someone else, this becomes less likely.
Hell, where did the foundation get its money before?
I've a ton of Pi4s and one Pi5, but the ease of use, the simplicity of 1L pc's is - for the stated applications - a no-brainer to me.
Then two used/refurbished Dell Optiplex's, great deals!
Curious because for me the affordability of a hobbyist PC is offset by the cost of Windows.
Why not just use the windows trial. Or buy a license key.
Windows license keys are absolutely everywhere. Maybe check the bios from some old laptops you own?
At my university we have this massive ewaste bin full of old pcs stickered with probably valid license keys.
Though you might be better off sticking with Linux anyway.
For initial purchase they certainly offer much better bang-for-buck compared to the Pi4 & Pi5, but do you have any useful links to information about relative power consumption when mostly idle?
I currently have a Pi400 (back when I bought that the Pi proper was out of stock everywhere except the scalpers charging >£150/unit for even the low RAM versions, but I found a decent deal on a Pi400 kit) running as a router/firewall with a few other small bits & bobs in containers. I keep considering replacing it with a small PC like that – my main concern being running a core bit of the home network off an SSD, though also having a bit more performance (both IO and CPU) might at times be useful too, and the Pi400 I'd like to reuse somewhere it is visible for “this is what home PCs used to be like when I were a lad” nostalgia reasons. The reasons I haven't so far (other than momentum: the 400 is sat there working quietly needing no interaction aside from the occasional update and paranoia checks of the config/log/other backups) are possible greater electricity use, and great noise if the mini-PC is not entirely passively cooled.
I had a look around and there are many values out there, just under 3W idle and about 7W spinning the CPU seem like reasonable values (to get better data I might have to try properly testing mine – I've got an overly fancy USB3 cable that displays the power passing through, so I could do that at some point, at least approximately).
Figures for mini PCs obviously vary by model, but also you see a lot of different figures for the same model (some of the difference could be due to differences in storage power draw, some don't say if they have an SSD or traditional disk, and some could be differences in ambient heat affecting draw for cooling? age may also have an effect on the efficiency of fan motors) but between 9W and 15W (mostly below 12) at idle feels like a reasonable guess based on those quick searches. This is lower than I was expecting. The official rating for an Optiplex 3070 (picking a random model from the popular ones) is “up to 1.2W” which is far lower than people have measured in their own tests so I'll not be trusting official values.
Of course the PiZero or similar would draw less than anything else I've already mentioned, though to be a router I'd need to add at least some NICs to that, some extra storage (given one of my concerns is running the router off an SD card), and would lose some CPU resource (the Zero 2 is roughly equivalent to a Pi3 there?) and a goodly chunk of RAM.
Well, yes, of course?
Why burn billable hours on a design I can send to production?
I think it was very much the supply chain snags of pandemic that made them get weird high prices everywhere.
Raspberry Pi dominates the market because it was the first to enter it and provides a limited product catalog that has limited churn with comprehensive support and regular software releases. The same cannot be said for it's competition which historically offered horrible software support land quickly iterated on products meaning that support generally only existed for the very latest.
Metadata of some processing will start leaking to external cloud whether you want it or not.
https://www.engadget.com/sony-investment-will-put-ai-chips-i...
https://aitrios-promo.wpp.developer.sony.com/en/services/edg...
Based on what? Even the linked article title says IN. Maybe they have different models, but based on all the wordings, it is very odd if it is HAT.
> Complaining about this shows ignorance.
If it is external HAT, then maybe. But until then, maybe not.
Edit: There seems to be AI kit available now, but it does not include this:
https://www.raspberrypi.com/documentation/accessories/ai-kit...
Raspberry Pis don't use any Sony chips on the SBC, Sony image sensors are on the camera modules, and since they're talking about imaging, this makes perfect sense.
Nothing about the Engadget article says to me it'll be on board the Raspberry Pi itself.
https://www.elektormagazine.com/news/raspberry-pi-monitor-te...
The original press releases don't say anything about incorporating it into the base Pi hardware and I don't think they would, given that it would add extra cost and most people don't need it.
That would be great. Here is the mission statement straight from raspberrypi.org:
> The Raspberry Pi Foundation is a UK-based charity with the mission to enable young people to realise their full potential through the power of computing and digital technologies.
Cheaper clones from China would aid that goal, but of course, it flies in the face of a fake non-profit company.
And if you are only interested in preventing counterfeits, just slap the trademarked Raspberry logo on there. No need to inhibit progress with patents.
This has most recently happened in the 3D printing world, with Prusa versus BambuLab. Who actually develops an open source slicer? Who allows 3rd-party firmware? Who contributes to the community? Who abides by open source licenses? Hint: It is not the Chinese company.
But at least that GPL-violating, closed-source printer was cheaper.
The opposite, having no clones makes it easier for a group (like RaspberryPi) to enshittify.
Enshittification is where a group first obtains a large market share with cheap/free services and then pivots to squeeze as much out as possible. Having a competent clone is a strong preventative.
> This has most recently happened in the 3D printing world, with Prusa versus BambuLab. Who actually develops an open source slicer? Who allows 3rd-party firmware? Who contributes to the community? Who abides by open source licenses? Hint: It is not the Chinese company.
Bambu pisses me off too.
Unfortunately your parent is talking about patents, not open source vs closed source, or license violations.
> No, it would be terrible, as technology development is not free.
What about compatibility? Wouldn't it be good for competitors to be able to provide compatible PIO interfaces, so customers can churn from one SBC to another SBC without needing to rewrite their code?
This causes a similar problem: people don't buy from the innovator, who needs more money to carry on innovating. They buy from the clone, who mostly only copies the innovator.
that's not how supply and demand works...
I assume that their mission being "accessibility of computing to all", and the org communicating so much on their mission focus, they would be in favor of cheap clones.
I expect most won't care or even notice. Company ownership is probably not that important to the average consumer. They care about great hardware support and a good "ecosystem" of accessories.
I know the first thought that came to my mind was: damn, I'll have to find new hardware when my Pi dies.
Who told you that?
Regardless, I posed it as a question exactly because I don't know.
Maybe they got it from the marketing material?
It was, but The appeal spread outside of that, which is why they are now going public and abandoning the precious niche mission and are instead aiming at mass market appeal.
Something like the SG2002
Ex: LicheeRV Nano
https://www.cnx-software.com/2024/02/08/licheerv-nano-low-co...
This is sufficient for "automation" hobby projects. You get the full Linux/OS creature comforts without the hassle of doing Arduino or Mircopython.
If you want a full desktop experience (or a project with video output) on a SBC it's probably easiest to go for a Intel SBC
RPi is getting squeezed from top and bottom price brackets
and so is a $2 ESPHome..
For actually working with hardware Arduino is much more convenient than whatever abomination you would write in Python.
For anything that needs video output you should use a cheap Android. Anything else is just wasting money.
Display out, built in amp, and 1GHz riscv for $10 USD [2]. Incredible.
1. https://wiki.sipeed.com/hardware/en/lichee/RV_Nano/1_intro.h...
2. https://a.aliexpress.com/_mPnKu26
https://a.aliexpress.com/_mtj0TmG
My conclusion is that although there is some nice hardware out there on paper, the poor circuit designs and atrocious software support mean anything I can get that meets my requirements will be an absolute nightmare; and I'm a software engineer of 1.5-decades with an Electronic Engineering degree, so the bar is high for something to be too-much-effort.
Take the Radxa Zero 3W[0] for example; the WiFi doesn't work properly, GPU acceleration doesn't always work, the OS images are out of date and don't always boot, when it does there are power issues, thermal issues, performance issues. No Debian 12 expected until H2 this year, if lucky; Android OS images are Android 9 (we're on what Android 14-15 now?).
[0]https://forum.radxa.com/c/zero
I love HN, but this is insufferable behavior. I also believe it’s why, despite our screeching, we accomplish absolutely nothing. You can’t accomplish anything when you’ve burned all bridges and then yelled “screw you.”
I'll be looking around at other SBCs, but I'm hapoy enough with my current RPi boards, just no longer confident I'll be happy with then in the future.
FWIW I don't find this in other tech spaces though they usually have a lot more young people. The media coverage of tech has certainly become focused on negatives, but if you hang out in Discord channels or talk with random Bluesky/X builders (keyword is builders here) there's definitely a culture of just talking about building or the more "financial" market analysis aspect. I agree with GP that HN has hit a pretty strong downward trajectory at this point that probably disincentivizes more rational commentary on topics that HN gets a lot of hate on.
If a random company with a very different mission (say, Intel) comes along and offers to purchase the company at a decent share price premium they pretty much have to take it, right?
... but it's how it's marketed to people so they believe it and follow Friedman's doctrine like gospel.
At this point it's a self-propagating meme. I frequently encounter people, learned people, who act like it is a law or moral principle (in fact, corporations aren't even required to be for profit even when operating commercially, they are just an organizational structure). When questioned, nobody can cleanly point where they learnt it, but actual check shows that Friedman Doctrine[1] was never written into law.
[1] https://en.wikipedia.org/wiki/Friedman_doctrine
That said, the misconception can be tracked down to one person in 1980... that sounds more like an idea that got marketed - whether by Friedman teaching in university in certain way, or his students continuing that, or people taking the idea from his paper and teaching that... or newspapers passing it further...
At some point you start to wonder cui bono?
This automatically prevents them from mindlessly switching.
> While it is certainly true that a central objective of for-profit corporations is to make money, modern corporate law does not require for-profit corporations to pursue profit at the expense of everything else, and many do not do so.
https://www.law.cornell.edu/supremecourt/text/13-354
And children should learn to program edge devices, they are entering a world flooded with them...
Which means little unless laws substantially change to allow them to do anything with them. Right now, edge devices are protected from "tampering" and "unauthorised use" by both technical and legal means.
Or you could do that now.
Protected from tampering? What?! "Attempting to prevent tampering".
Wow. Messages from an alternate reality. There are at least three assertions in the above message which are untrue.
Raspberry Pis are great if you want easy internet connectivity in your hardware project, or need an OS or lots of computing power.
If they were readily available and $20 each I'd grant them that they are more convenient. But at real prices and availability that's more difficult to justify
For awhile they were super cheap and so it was easier to justify, but when they’re getting into $200+, at some point I feel like you’re going to get a better deal with a mini x86_64 PC, with considerably better performance (if not per watt, at least total performance).
I had some used thin clients with AMD CPUs a few years ago. They were $50 each, and only ate like 10W of power, even under load. No, there wasn’t GPIO built in, but even three years ago there was USB 3.0 and it blew the Raspberry Pi out of the water at the time.
I was running a full rack mount server for a few years after that but since cheap mini gaming PCs have been coming out on Amazon semi recently, I bought one of those to save power usage, and I have extremely happy with them. They have a GPU where I have been able to VAAPI transcoding fully working and they have a considerably faster CPU, and since RPis have gotten so expensive with spiked demand, it wasn’t even that much more expensive for a home lab.
I'd rather not. Working with real hardware is much more fun and SBC's barely draw any power.
I just wouldn't buy Pi's.
Buy a Pine64 or RockPi or Milk-V or something similar.
The price some of these smaller boards go by is staggering.
The secret sauce allowed me to search the web for guides on what I wanted to do and have a template to work from. Not having to spend potentially hours or days of my free time figuring out some obscure Linux issue or driver support problem made up for the extra cost of the board.
The secret sauce was always the documentation and software support, and if the cheap clones can master that, then the boys in England have a real problem, but that remains to be seen. A lot of that comes not from cost savings inherent to microcomputer technology, but because of the dedication of people who know how to communicate to software engineers and hobbyists.
My hope is that most of the price increases apply more to their industrial computing products than their hobbyist ones.
Like the OP, I have avoided the clones in the past because I wanted to reward the company that were "cause-driven" to do this. I too am less inclined now though.
If so, sure.
My concern is, I don't want to spend any time whatsoever messing with stuff like driver or architecture issues. Needing a few repositories for Debian packages is okay, but pushing it. I'm not here to configure a system; I'm here to work on a project I find fun or useful.
Ideally I take the latest RaspberryOS and slap it in and it "just works".
I buy Pis because of their support, stability and out of the box experience is far superior to other cheaper clones. I have tons of cards which are paperweights without Armbian or self-rolled distributions, and even them some are feature-capped because of kernel-dependent closed source modules or out-of-tree drivers.
I can still download and install a recent version of RaspberryPi OS to my Raspi 1, customize the image even before it touches the SD card, and remove the card and install to a more recent Pi, and continue working.
More importantly, I can upgrade OS from version to version (Armbian doesn't support this, and they say it explicitly for example).
OrangePi, Radxa and Asus failed on all these fronts. I still peek and poke OrangePi 5B's Debian image since I can't trust them blindly, but can't use its AI and video (rendering/encoding) accelerators without its official image.
Oh, and my OrangePi Zero always crashes within a day or two if I don't practically disable its frequency scaling capability via governor. Even a good cable and adapter which can run a Raspi 3/4 without any hiccups can't make this postage stamp sized thing happy and stable.
P.S.: For everyone who thinks RaspberryPi brand and hardware will be enshittified: I'm not saying that it won't happen. What I'm saying is, let's watch how this moat evaporates, and who passes (said moat) and replaces them on their hill.
So it's an if from my perspective, since I don't constantly watch them and ask "aretheyenshittifiedyet?".
When a company lists they switch from being customer focused to being share-holder focused. They will do anything and everything to make number go up.
The market is healthy from that standpoint, but let's watch as it happens and see what happens.
Are you also inclined to support and develop kernel and OS level software for them? Because the lack of good support is why they remain woefully noncompetitive despite the lower price.
The only Raspberry Pi models I regret buying are the sub-4GB versions, since you're always one step away from needing swap.
In 30 words,
instead of receiving a share of the company’s profits each year, we will convert some of our shareholding into an endowment that we will use to fund our educational programmes.
While it remains to be seen how the IPO will affect future Raspberry Pi products, it's hard for me to argue against the Foundation diversifying its investments, especially in light of growing competition from cheap clones.
[1] https://www.raspberrypi.org/blog/what-would-an-ipo-mean-for-...
https://www.profootballnetwork.com/who-owns-the-green-bay-pa...
[1] https://www.carl-zeiss-stiftung.de/en/
The TechCrunch article is speaking only of it like any other commercial investment opportunity.
Where's a good writeup of charity/non-profit mission concerns about this commercial subsidiary?
(Example from early on in Raspberry Pi history: There were concerns about pragmatic compromises of open source or libre software principles that the RasPi made, so that they could use the closed-software Broadcom SoC, presumably to deliver an otherwise good education-oriented SBC at a price point that made it accessible and got uptake.)
But yeah, kinda wack.
"Our parent company is a non-profit, but you'll be paying a few taxes!"
Trading since yesterday.
https://www.londonstockexchange.com/stock/RPI/raspberry-pi-h...
RISC-V means only that the instruction set is open source.
It does NOT mean:
- There are open source drivers
- The chip design is open source
- The chip design is not patented
- The software is not patented
- The chip has no proprietary additional instructions
The NVIDIA firmware blob and drivers could run on RISC-V right now with no practical benefit to you. The ISA being open source does mean that open source CPU cores can exist, but no chip is required to use them.
The ISA is arguably the most important thing to have open, and the hardest to get open since it is the thing that requires the most coordination between so many disparate and disinterested entities.
Today we celebrate an open ISA, tomorrow, being enabled by the open ISA, we'll be able to celebrate the next iteration of open, until at some point we get fully open.
Is there any SOC chip that is totally open with no proprietary features??
All drivers are not needed, nobody is expecting to do triple 5k displays at 120FPS. Basic WLAN, Ethernet, Bluetooth and basic display capabilities around HD even at 30FPS are good enough for that market.
As for open source, you as a computer architect, can't even make your own SoC that is on x86 or ARM without paying royalties and spending lots of money upfront.
Not so with RISC-V so I guess you'll see lot more RISC-V in future.
This time - 20204 is the 1999 of RISC-V if you want to borrow a precedent from Linux's rise as defacto OS on server side.
They actually do... Nvidia GSP runs RISC-V: https://www.techpowerup.com/291088/nvidia-unlocks-gpu-system... which furthers prove your point
arduino is approaching the business bottom up (tiny well-known chips), pi top-down (it's just linux with some GPIO pins)
also have competitors at the margin like nvidia (who could legit own 'iot edge AI' if that ever even matters) + waveshare, the off-brand pi manufacturer who remarkably has way more pi-compatible devices than actual pi
excited for IPO if it means the #1 linux board finally supports suspend-to-ram
You’re describing a product with well known stock shortages that have, in floating, unlocked a bunch of capital which can among other things help them scale their production to meet demand.
That sounds like a good thing both as a shareholder and as a consumer.
The company markets itself as an educational tool, but 70% of its sales are industrial? Like I get “bringing low cost computers to everyone” feels and sounds good in press releases but when the business actually focuses on something different - that seems counterproductive to their stated goal - what am I supposed to take from that?
Where I am it has been next to impossible to order a non-kit model of Pi released within the past several years since at least mid-pandemic.
I started buying up SFF Lenovo boxes on eBay, which I've been somewhat happier with since I never used some of the Pi's standout features like GPIO.