Raspberry Pi started out being for kids, schools and hobbyists.
During the chip shortage Eben Upton made the decision to focus all their manufacturing output on their business and industrial customers.
Upton justification was that those companies income depended on Raspberry Pi so that’s where they prioritized their supply.
Realistically those companies could have found other technology solutions, designed their systems around other tech. Upton could have given them notice to find other designs so Raspberry Pi could get back to kids and schools.
But that never happened.
So the realistic outcome is that for years now raspberry pi has been a dedicated supplier of industrial computing systems. Schools, kids and hobbyists are long forgotten.
Nothing has changed and raspberry pi continues to pretend it’s about school, kids and hobbyists but it’s just and industrial computer systems company.
Raspberry pi is a strange sort of ghost company. It exists, but you haven’t been able to buy their core products for many years now. The maintain a sort of phony facade continuing to market to community and schools and kids which is very strange since the community response is always “great but we can’t buy your stuff”.
> So the realistic outcome is that for years now raspberry pi has been a dedicated supplier of industrial computing systems. Schools, kids and hobbyists are long forgotten.
The rp2040 is cheaper and probably better for teaching embedded electronics. (and super easy to get access too). It's documentation shits all over the ESP* tool chain, and its a nice, smooth coherently documented ecosystem.
I understand that you want the pi-4, as do I. But. The money from pi sales goes to education, you can see it in the company accounts. The documentation, and family of hardware is unparalleled anywhere (outside maybe adafruit)
The world is complex, and sometimes doesn't do what you want. Don't shit on something because you're upset that they aren't doing exactly what you want.
I don't see the Zero (2 or 2W) in stock often, but they do come and go.
The model 3 A+ has been in stock regularly now (as predicted in Raspberry Pi's December 2022 blog post), and supposedly we'll get back to regular stock by the end of the 3rd quarter this year for Pi 4 products. Not sure about Compute Module 4... but I hope that too, because it's arguably way more useful.
The RP2040 and Pico have been widely available almost since launch, and in some ways are even more intriguing for educational purposes, especially since they cost $1 and $5, respectively. And the Pico W has been an excellent alternative to ESP devices for ESPHome and other automation needs. The built-in PIO has been a pretty interesting feature, enabling some use cases I didn't initially see.
In the end, I think this year's stock availability will make or break the SBC side of things for the maker / hobbyist community.
There are way faster and more feature-rich boards with chips like the RK3588 at this point, but the software support and available libraries are still leaps and bounds better on the Pi, which is why the nearly 5-year-old hardware is selling at a premium :(
Kids in the UK (and elsewhere?) can access the Micro:bit computer[0], while not the same and powerful/extendable as R Pi - it is cheap, good and plenty available. It includes a LED display and motion sensor. Kids can program it using "block coding", or write Python code that runs with the help of MicroPython[1].
Feel like MicroBit v2 is a drastically underutilised tool. There's the potential to make completely self contained gadgets that anyone could use by copying a file into a directory.
Falls into an awkward position where the people who could utilise it can do much more for much cheaper with an esp8266
Some pretty neat UI design limitation challenges involved too with the 5x5 display. e.g. I made a pomodoro timer with it; iirc I used used left-to-right top-to-bottom for normal time, reversed for breaks, a number for how far into the cycle it was and flashing figures for when setting inputs and it beeped when a break started/ended with a longer beep for the big break
I'm still not sure why the BBC went down the route of making its own platform when adapting the arduino world would have been simpler and more effective. Still its good to see that the BBC is doing computing again.
It’s not possible. My job depends solely on the Raspberry Pi and the software stack is built on top of Linux on Raspberry Pi 3 or newer.
When the supply of Raspberry Pis was drying up, we started recycling returns to make them work again, but even that would not be sustainable.
There is no real alternative, nothing with a functioning Linux kernel and hardware that doesn’t die because of low/high temperatures or humidity or the like. The Raspberry Pi is engineered to a higher standard.
Only since December 2022, a next-generation Rockchip ARM board is shipping that’s based on Cortex A-76 and has 8K video out, but it is expected to get mainline Linux support only in 6.3.
Raspberry Pi works, the rest is just tinkering. Raspberry Pi supports some important niche economies, whose dependents would also break apart, if they didn’t get them. Medical sector, laboratory analysis, etc, etc.
—
What I want to add is that the Raspberry Pi 400 always stayed in stock, and I’ve bought one for my child. So you’re also wrong on that aspect as well.
> What I want to add is that the Raspberry Pi 400 always stayed in stock, and I’ve bought one for my child. So you’re also wrong on that aspect as well.
Not in my part of the world. In the US, Chicago area, the Raspberry Pi is carried by Microcenter. They seem to have a steady supply of the 3 Pico variants. It ends there. I think I've seen the 400 on average one or two days/month for at least a year. 3A+ same. A bit less for the Zero/Zero W and even less for the 4B variants. I haven't seen a CM4 at Microcenter ever that I can recall. I found 3 US vendors who will take a backorder for the CM4. I've had one on backorder with all three since 3rd quarter 2022.
edit: flippant jokes aside, they suck. They are the opposite of tough. I have a dead one, it was easy to kill, and I heard the same story from many others who got all hot and bothered by the specs.
What sort of abuse has it withstood? I vaguely recall a blog I was reading at the time had a bin full of dead ones. Maybe we both were incompetent, but you certainly don't hear much about em these days, do you?
It has unprotected 3V inputs. Naked, it shouldn't be subjected to any (electrical) abuse. It's a SoM and a few other devices, not a PLC. That said, I'm not particularly kind to silicon and I have overvolted the pins pretty frequently since I keep forgetting that they're not spec'd for 5V.
The reason you don't hear much about them in the consumer space is that Raspberry Pi's are far more popular there. However, for commercial/industrial uses, SoM's are used everywhere.
Fair enough, but ubiquity of a single device creates a froth of creative energy that generally does not originate from purely industrial spaces, which is why so many devices have coalesced around the pi.
I imagine that from an objective standpoint, the pi is actually not any more inherently robust than a beaglebone, except for the critical mass of users doing development of its os, synergizing with the hardware devices that also appear. Beta was the "right" choice from a technical standpoint, but then I spent years not being able to watch anything that was on VHS, which was everything of course.
I completely agree that it's the ecosystem that gives the Pi its value. No argument from me there. I've shipped Pi's on commercial projects and recommended them for others (and been outvoted: old habits die hard!). I just don't want anyone thinking that they're any better, electrically, than the many other options out there.
> What I want to add is that the Raspberry Pi 400 always stayed in stock, and I’ve bought one for my child. So you’re also wrong on that aspect as well.
It's out of stock everywhere in the UK right now and it has been that way for as long as I can remember.
Out of interest, was it possible to apply for access to the OEM pi stock? I don't know how the whole thing worked but wasn't their strategy explicitly designed to help people in positions like yours?
I see that Nvidia are selling these again, but are they providing anything newer than Ubuntu 18.04 (now unsupported) to go with them? Are industrial customers just rolling their own distributions?
It's why I much prefer the Arduino platform. It doesn't depend on using a very specific chip that can suffer a shortage or be monopolized by the manufacturer.
It's an ecosystem that can use several families of chips, from 8 bit AVRs to 32 bit ARMs. The Arduino code you find online will most likely work on either one, as the most used low level functionality (timers, GPIO, etc.) are abstracted by a high level API that will work on any microcontroller supported by the Arduino IDE.
The RaspbeeryPi ecosystem is too tightly coupled to that very specific Broadcom chip family, which gives them and the RaspberryPI foundation, a powerful monopoly.
Well the Raspberry Pi and Arduino are two different kinds of computers. When you need a full Linux system Raspberry Pi is great. For embedded applications Arduino is handy.
But there is now the RP2040, a processor from Raspberry Pi that works well in most Arduino-like applications. It has a lot of very powerful features! And what’s interesting is that last time I checked this, it was the most in stock processor at Digikey (to pick one vendor as an example). The stock numbers for the RP2040 are very high, and I suspect it will never experience a shortage.
The RP2040 and its dev board the Raspberry Pi Pico don’t depend on Broadcom and they’re wonderful chips. If you do a lot of Arduino I highly recommend checking it out!
>When you need a full Linux system Raspberry Pi is great.
Then you're much better off getting one of the endlessly available low-power X86 small form factor platforms either new or off the second hand market, available with Intel or AMD chips.
And unlike the Broadcom based R-PI, you'll have a much better time as you won't get locked into any specific chip vendor so you can easily migrate to whatever chip offers you the best price or performance or energy efficiency at the time of purchase.
The standardization of the X86 platforms is unbeatable by the current crop of ARM SBCs with their proprietary firmware blobs and weird boot methods.
Locking yourself into a very specific chip platform of a single manufacturer, just for running linux and a web server, is a huge mistake and we're now seeing the consequences of not having a competitor/dual source.
It's a full linux system with a lot of IO and easy to embed in other devices. Since this thread is comparing arduino and Raspberry Pi I think most commenters are thinking about embedded/hardware controlling applications. Plus it has multiple form factors useful for prototyping and final product. And it's just the de facto standard for these applications. There are competing products with varying degrees of compatibility but with Raspberry Pi if you want to add a camera, sensor, motor driver etc you can always find a pre-made hardware interface and the code to make it work, at least at the prototype stage.
For a webserver an small x86 system is a good choice but it doesn't work as the brains of an IoT device or 3D printer.
Is a raspberry pi suitable as the brains of a 3d printer? You could operate the printer with far less than that. I suppose if you want a camera you could power a pi and a camera from the printer’s power supply, and it could be useful for receiving jobs to send to the printer, but it strikes me as the wrong tool for actually performing print jobs.
I’m prepared to be corrected, though. Maybe I’ve just used really dumb printers and I’m not familiar with the world of pi-based 3d printers.
As it happens I just got my very basic printer working after breaking it a year ago (fixed it in an afternoon, spent a year reaching that afternoon), and I've been learning all about this.
Fast answer: yes, you can use a Pi, most often running OctoPi, but only a Pi3 at least is fast enough - with a 2 or some Zero models, it will be unable to keep up with faster prints and possibly wreck em.
There's a newer, better firmware which I'm about to install on my printer called Klipper, which replaces the more typical Marlin that you find on most entry-level printers, and it depends on a Pi running alongside the printer's microcontroller. For that, it apparently is a huge boost in both speed and quality, so I'm checking it out.
Nice, thanks. I should dig into this stuff one day. I’ve only got a Zero W handy so I can’t try octoprint, but maybe when it’s possible to buy a pi again I’ll try this stuff out.
It’s so cool that tiny little microcontrollers can handle prints without much trouble (though you can get stuttering if they can’t hold enough instructions at once), but while a Pi, despite being far more powerful, can become unreliable and ruin prints. Given that’s likely because it’s talking over a network, handling a camera feed, running an operating system, etc. I just find it so cool how useful and capable microcontrollers can be in the right settings.
Not just that it's doing more. A small microcontroller will run in realtime with completely repeatable and testable operation, whereas a pi running linux cannot run a program like this.
Even though I claim pi is the de facto standard for getting started, a microcontroller controlling the stepper motors is a technically improved solution.
You will probably be interested to check out the beaglebone series of boards. They are less of a de facto standard but they are popular enough. Their CPUs have two 200Mhz 32bit microcontroller cores running in realtime. They can directly share the same memory as the ARM cores so you can have the best of both worlds on one computer.
I've been interested in the Beaglebones. They seem much better suited to the types of things I build, and while they aren't super affordable, they're much cheaper than a Pi right now.
Do you know of any obvious drawbacks with beagles? I can imagine the community is smaller, so there would be fewer peripherals and libraries to choose from. Generally speaking I don't really need that, though. Sometimes off-the-shelf solutions are even unappealing because I enjoy learning, or I don't want to hit edge-cases where it doesn't quite fit my use case.
You hit the nail on the head. The community and number of useful repositories is smaller. But certainly large enough that you won't be stuck. Aside from that, they don't have MIPI camera interfaces so for anything image related you are limited to USB cameras.
Depends on your needs. The raspberry pi may have poor stock right now, but when available they are cheap, low power, and have excellent documentation. I’m designing an open source autonomous farming robot and we’ve been using the Raspberry Pi CM4 as the main computer on the motherboard for a while now. It’s perfect for us. I feel like x86 would be a mess. The I/O on the CM4 is extremely important to us, and on the custom motherboard I designed we’re using every available pin of the CM4 to communicate to serial, SPI, I2C, and GPIO peripherals on the motherboard.
> Locking yourself into a very specific chip platform of a single manufacturer, just for running linux and a web server
Ah, yeah, I am very much not doing this and I’m surprised you didn’t start your comment with this stipulation. I use raspberry pi for robotics where the standardized documentation and abundant I/O are extremely helpful.
Yeah but what are you _after_ though? if you want a simple host of software, then yeah a VM/docker host on a low powered x86 machine is probably what you want.
but.
As soon as you start worrying about power, or GPIO, then it starts to get a bit messy.
There is still a sweetspot for the rpi. but as soon as you start wanting to do pure software, then you might want to think again. Its very much a spectrum.
>As soon as you start worrying about power, or GPIO, then it starts to get a bit messy.
7nm-10nm X86 CPUs from Intel and AMD are nearly as efficient as a R-PI, IPC for IPC, and GPIOs can be extended on them via USB in a standardized manner if desired.
Most people buy R-Pis to run linux, python, a web server or turn on some LEDs. Linux, python and web servers can be run on low power X86 machines, and LEDs can be turned on by an Arduino or R-PI 2040 microcontroller.
People stuck to using R-Pis out of comfort and inertia because it was the only thing they heard about and it was the best thing at the time when X86 was not competing in the space, but today it's no longer the only option or the best option. People need to start exploring the other options.
Again, I am designing a farming robot. We absolutely do not want to use USB for anything critical (which is pretty much everything). On our raspberry pi motherboard I’m using SPI to talk to CAN bus chips to do motor control. I’ve had so many weird control USB issues in my life I would absolutely not want to use USB for motor control!
I've been looking to replace my current intel 10watt server with a AMD ~10 watt device for a >3x performance boost, but they are all going to cost >$250.
My current ARM server runs samba, transmission, icecast2 +deefuzzer and nginx for less than 4 W, including the power for a 1 TB Sata3 SSD. No idea how its performances compare to a 10 W x86 machine but it idles all the time and 4 W < 10 W. It costs about 100 Euro.
I think the the main issue is software support.
ARM chips manufacturers are providing bad software. Kernel is outdated, hardware acceleration is undocumented, etc.
And when there is good software support, it is oriented towards smartphones and never lasts more than 3 years (hello Qualcomm).
So lots of industries are willing to compromise the performances (Raspi are not top-of-the-line SBC) over a very stable software
>ARM chips manufacturers are providing bad software. Kernel is outdated, hardware acceleration is undocumented, etc. And when there is good software support, it is oriented towards smartphones and never lasts more than 3 years (hello Qualcomm).
It's why I'm sticking with X86 for my Linuxy projects. Migrating your entire SW stack from one ARM SBC to another is virtually impossible, while on X86 you never even think about it.
Not anymore. Intel sells Celeron quad-core chips with 5W-10W TDP with integrated graphics. AMD also.
R-Pi was leading 8 years ago when it was readily available, dirt cheap and x86 was only gas guzzlers. Now the market has shifted against it.
People should really look into recent developments instead of defaulting to R-Pi out of habit.
I can buy a Intel N5095 based NUC for 170 Euros on Amazon with 8GB of RAM and 256GB SSD[1]. Sure it's more expensive than a PI, but that thing is readily available, will sip power and blow any R-Pi out of the water performance wise, plus it comes with 2 Gigabit Ethernet ports and 3 display outputs. It can be a great home server, it can be a great PC replacement, it can be anything you want. R-Pi is still far to weak in comparison.
If that's too expensive and overkill for you, and you just want some low-end ARM SBC to blink some LEDs from the internet, there's still the Beagle Bone Black with a single core Cortex A8 and 512MB of RAM, for around 60-70 Euros.
They are loud-ish, especially compared to a Pi. They are a bit faster and run windows though (which was required by the vendor that I had no say in choosing).
That said, for my personal projects, I still kinda default to the Pi model. A bit more power efficient and there are some solid home projects built around the Pi.
I now have like 3-4 of them in the house, ranging from a Dakboard above the fridge, to a couple pi-kvm's, including one with a multiport attachment, to remotely manage some of my stuff at home so I dont have to lug monitors and keyboard about as much. Heck I strongly considered a pi-kvm for work, even if just used as a remote crashcart of sorts.
10 W is the the TDP, a fudged estimate of the maximum power, for x86 chip. Idle power is much, much lower and that's really the important part that the Raspberry Pi can't match because anecdotally it has minimal power management capabilities and therefore runs at full power (~4-5 W) whether idle or active. An x86 system can beat that.
[andrewstuart] is exactly on target with his critique. The Raspberry Pi foundation is only using the educational market as a false cover. When the pi-zero family was introduced, it looked like they were being true to their stated goals, but that didn't last long.
Furthermore, Sony is the last company I would ever expect to share those goals. Sony has always targeted the high end of the market. They are profit driven, and have NEVER been interested in the low cost educational or hobbyist value of their investments.
What do you make of the AIBO line, the Net Yaroze and the PS2 Linux kit then? (Just really curious about your assertion they have no interest in the hobbyist and education markets, since those are some admittedly outdated examples)
IIRC it was the PS3 that advertised Linux on the box (of the launch model at least) and supported the "Other OS" feature until Sony disabled it with a firmware update. ;-(
> So the realistic outcome is that for years now raspberry pi has been a dedicated supplier of industrial computing systems. Schools, kids and hobbyists are long forgotten.
When I backed the Turing Pi 2 project, I hoped that I'd be able to source some CM4s by the time they were done with the project. Well, the boards arrived last week and I still haven't been able to get my hands on some CM4s, so those boards are just sitting on my shelf waiting.
You made me realise that it's time to drop the Raspberry Pi for new projects, much like I decided to drop Android after 13 years of use in favour of an iPhone.
Unless Raspberry Pi comes back with a roar in the next year or two, the best hardware platforms will be decommissioned x86 low-power SFF PCs, followed by cheap-ish special-purpose ARM devices if a specific application requires that. Raspberry Pi was the only well-supported, well-priced ARM machine that was truly general-purpose in my eyes.
The Pi isnt bad for certain things. And can still be had for fairly cheap.
I put one in my parents house for remote support, some small postfix and NUT/UPS shutdown goodness.
I usually just get a canakit on amazon for under 200 bucks and it comes with most of the utilities, cases etc needed to flesh it out.
Still thinking of making a Sonos style speaker system in the house with some Pi's and some JBL charge speakers or something which can always be used as a BT speaker and even setup in tandem for stereo. Those things are solid. I use them on the boat, when playing along to songs, even when setting up a projector in the yard for the kids to watch a movie in a "special" setup.
The second hand and refurbished market remains the best option, regardless of the form factor. You can get a more powerful, multicore, perfectly functionnal laptop, desktop or minipc with 4GB of memory or more for 100usd/€ while the pi price do not drop as much. In most case battery will be dead on the laptops but many people stick with old laptops with dead batteries for years and just use them as portable plugged desktop because they don't want to spend the money on a new computer.
Raspberry pi are really only interesting if you want to stick to the small form factor and the usually come without screens and keyboards.
EDIT: raised price from 50 to 100€ as a quick search only showed netbooks and laptop with 1 or 2GB of ram for 50€ in a 10km radius. In 2 minutes I found a Thinkpad T420 with crappy screen resolution 1366x768.
Yeah the pros of having access to a common architecture cannot be understated. You'd think ARM would be widely supported by now given that nearly every phone runs it, but every time one tries to do something vaguely nonstandard you run into walls that simply aren't there on x64 (especially with an AMD/Nvidia/Intel GPU and common net card instead of <insert SoC manufacturer's barely supported custom homebrew card>). There's a reason one cannot simply update Android phones to the latest OS without explicit vendor support and the reason why is not pretty.
Goes doubly so for the Pi specifically, since they stayed on 32 bit armhf OSes for the longest time which have even less package support and just straight up cannot run a wide range of software. And the new arm64 builds are laughably bad, having an almost 10-50% performance overhead on certain processes. And the Pi 4 is already painfully slow even compared to cheap android phones.
It is in most aspects an absolutely terrible way to get started with computers.
Conversely, for anything embedded or with hardware control aspect to it that needs more than an arduino or needs a GUI, Pi is the common architecture that has widely available hardware interfaces, code and community support available.
Yeah, if you're using it for web development or something it will suck to just try and install all the dependencies you need. But if you're making a 3D printer with a web interface, touch screen and maybe a camera, or a robotics project, it is by far the easiest thing to get started with.
Originally Pi was created to promote CS learning and I get that it has some shortcomings there but it sure as hell has enabled a lot of people to learn about embedded computing and robotics.
That's definitely true and why the Pi is so popular at all, but if you don't need to explicitly interact with hardware there are generally better options.
Cannot comment on that location but I've noticed a few sellers on ebay seem too just take older 4GB ram chromebook and install linux on them before selling them for $50-75. If you're not as lazy as I am you can do the installation yourself of course.
I think the raspberry pi is mainly interesting for its small size, GPIO pins and very low idle power.
And I have experienced enough times that actual durable storage medium would be a good thing. I think there should be better options if you just want small cheap box to run linux. Though the cheap is slight challenge.
There's been an announcement that the supply chain has been retooled, and we can expect lots of stock by summertime. That's a short enough timeline that I say let's wait and see, first off.
Second, kids and schools have not lacked for platform tools to learn about electronics and computing - we are drowning in computers, and any instructor that wants to teach kids about GPIO can get any of about a billion arduino clones for a couple bucks apiece.
Your comment presumes that things will stay this way for good, and that isn't the case regardless; they're very due for a Pi 5 release, and picking a new SoC with a reliable supplier is a good way to sort both issues.
Anyways, let's see what it looks like come Labor Day.
> Realistically those companies could have found other technology solutions, designed their systems around other tech.
There is nothing realistic about this statement. What useful notice period and what alternate designs (that were magically not impacted) would those companies have been able to use? We had, and are still suffering from, a GLOBAL semiconductor shortage. The GLOBAL supply chain fell over.
I would expect the charity arm continues it's work with kids and schools. The focus of this article is Raspberry Pi Ltd, a for-profit computer company.
It is sad they split the company up and the roots are gone but it makes sense if they are to grow and continue producing their own chips/boards/peripherals, hopefully without Broadcom.
> you haven’t been able to buy their core products for many years now.
Did they manage to keep it in the same price envelope? changing CPU/memory to something double the price is going to have caused a whole lot of squealing.
> Manufacturers and brokers are, of course, able to charge a premium, and companies have little choice but to pay.
But the article is irrelevant anyway. It's talking about peripheral 50-cent chips, not the Raspberry Pi processor which is likely the heart of your device.
Well, if the alternative is going out of business, you'll redesign the product around whatever you can buy right now! That's exactly what an untold number of companies have had to do, some multiple times in one year.
I almost quit my day job in 2020 building a product based on Raspberry Pi Compute module 4, it has great form factor. But I didn’t! And it was another proof, that playing safe is not the worst way to do things.
What’s alternative!? As an electrical engineer I designed a Spartan 7 FPGA based board with same form factor as compute module 4 for replacement in my system. Nowadays Spartan 7 chips are available again at Mouser and Digikey.
I ready don’t need a Raspberry Pi with A.I. chip, because I can’t buy even a regular one.
some of the cm4 variants do have flash memory, 8, 16 and 32gb mmc flash, but those are just as hard to obtain as standard raspberry pis. you use the raspberry pi io board or any of the third party ones to connect the cm4.
AI inference engines. The learning part is done on big machines in the cloud. The inference part (giving answers) can be done on edge devices. There's a ton of articles listing the competitors, here's one fairly at random: https://www.aiacceleratorinstitute.com/top-20-chips-choice-2... (Note the article covers both learning and inference chips.) What's interesting is I can't find out much about Sony's chips.
If we're talking about Sony, it's probably vision related A.I. accelerators. Their Alpha cameras depend on them very much. Even my now old A7-III can do amazing things realtime.
Chips optimized to perform the type of calculations used for NN inference at high parallelism. A good example would be the google spinoff https://coral.ai/ (though their usecase is highly limited by sub-par software constraints)
no idea but if you asked me what I would want it to do is tensors for GPT and DL image processing inference and if possible training (so I can at least finetune a pretrained model)
Yeah, assembly is part of hardware which Sony can do just fine.
The current announcement however is about some AI co-processor. Mark my words, if the software side is anything implemented by Sony instead of the community it will be a shitshow.
I can't imagine anything about this thing will be open so whatever the community produces will be at the mercy of the shit-show firmware/driver Sony provides. Fun times ahead!
During the chip shortage Eben Upton made the decision to focus all their manufacturing output on their business and industrial customers.
Upton justification was that those companies income depended on Raspberry Pi so that’s where they prioritized their supply.
Realistically those companies could have found other technology solutions, designed their systems around other tech. Upton could have given them notice to find other designs so Raspberry Pi could get back to kids and schools.
But that never happened.
So the realistic outcome is that for years now raspberry pi has been a dedicated supplier of industrial computing systems. Schools, kids and hobbyists are long forgotten.
Nothing has changed and raspberry pi continues to pretend it’s about school, kids and hobbyists but it’s just and industrial computer systems company.
Raspberry pi is a strange sort of ghost company. It exists, but you haven’t been able to buy their core products for many years now. The maintain a sort of phony facade continuing to market to community and schools and kids which is very strange since the community response is always “great but we can’t buy your stuff”.