Eben Upton expands on Raspberry Pi shortage

Raspberry Pi Shortage

As you may have noticed, Raspberry Pi boards may get hard to get and/or be sold at excessive prices by some resellers. Eben Upton explains the reasons behind the Raspberry Pi Shortage and provides some recommendations to work around the shortage and high prices.

Raspberry Pi Trading is still manufacturing around 500,000 Raspberry Pi boards or modules per month despite the global semiconductors shortage, and the issue they face is more of a demand shock than a supply shock with demand for Raspberry Pi products having increased sharply from the start of 2021 and remains unabated as the supply chain can not handle that extra demand at this time. That means distributors have backlogs in almost all products, and it takes time for customers to get their orders.

Bots are also involved in making automatic purchases as soon as stock becomes available often in order to resell those boards at a higher price. So Approved Resellers have implemented single unit limits or even two-factor authentication to prevent bots from scooping the boards. Raspberry Pi Trading also had to make some tough choices and are currently prioritizing commercial and industrial customers at the expense of individuals, because the former depends on the latter for their livelihood. There’s currently enough supply for business customers, which explains why we can still frequently see new products launched with the Raspberry Pi CM4 modules, even though they seem unobtainium for individuals.

Eben’s recommendations will be somewhat disappointing, but it is what it is. First, he recommends purchasing only from approved resellers, listed on the products page, as they are held to a single price, and they won’t sell you a Raspberry Pi 4 SBC for 100 Euros or other inflated prices. Approved resellers should also receive supplies faster, and you can check stocks on rpilocator. The second recommendation is to buy a Raspberry Pi Pico or RP2040 MCU, as there’s plenty of stock as previously mentioned, or a Raspberry Pi 400 Keyboard PC. Obviously, that does not help if all you need is a Linux-capable single board computer.

Another solution that won’t come from Eben’s mouth is that you could buy a similar single board computer from competitors with, for instance, Hardkernel ODROID-C4, Orange Pi 4 LTS, Pine64 RockPro64, FriendlyELEC NanoPi M4V2, RockPi 4, Khadas VIM3, and others. But I reckon that’s not for everybody as it may require some technical knowledge and updates to the software. Finally, since Rasperry Trading prioritizes commercial customers, you could always try to buy through your company, but I could not find any information about the minimum order quantity, which is typically in the hundreds. Companies having trouble finding stock and wanting to order Raspberry Pi products in volume for an industrial or commercial application can contact Raspberry Pi Trading at business [at] raspberrypi.com.

Via Raspberry Pi website

Share this:
FacebookTwitterHacker NewsSlashdotRedditLinkedInPinterestFlipboardMeWeLineEmailShare

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress

ROCK Pi 4C Plus

60 Replies to “Eben Upton expands on Raspberry Pi shortage”

  1. As if this wasn’t obvious that they prioritize their commercial customers… you can think of this how you want but the times when RPi were mostly toys for tinkerers are long gone.

    1. This was my thinking also, they aren’t about making money but prioritize business sales… humph.

      1. As they said, lots of people depends on the Pi hardware to be able to deliver their products and get their livehood.
        Either way, they would sell all those Pis and get lots of money

      2. RPi Trading Ltd. is not a charity.

        Why wouldn’t they ‘make money’ by selling 50k CM4 modules of one kind directly to an industrial customer instead of sending 20 different CM4 SKUs in 200 different packages to various resellers throughout the world having to share margins with those resellers and dealing with annoying consumer issues (e.g. false DOAs since consumer too stupid and all the other unnecessary support stuff)?

    2. Sure they prioritize commercial customers but the Raspberry Pi Foundation operates as a UK-based charity pretending to help digital education. Profit plus tax breaks, the best of both worlds.

    3. Very good point. When you look at one of the statements of the Pi Foundation it states:

      ‘You can empower young people worldwide to learn computing and become confident, creative digital makers.’.

      It is classed as a charity, which makes Eben’s statement even more ludicrous and completely disingenuous.

      1. > Pi Foundation

        Why do SBC users constantly mention this RPi Foundation? They almost never deal with this entity. They deal with RPi Trading Ltd. whose CEO is Eben Upton. It’s a company built to make money, that develops hardware and software and that also pays employees to brag about their success in the RPi forums and to censor there.

        The RPi Foundation has nothing to do with all of this so what’s the point of mentioning it?

        1. The two are/were? intertwined, as Eben previously said that all profits from Raspberry Pi Trading would go to the Raspberry Pi Foundation.
          I’m not sure whether that is still the case, and in the future, that will certainly not be if they go ahead with the planned IPO.

        2. Simple, because it is relevant to the creation of the enterprise and it’s founding principle in the first place.

          Just look at the statement again.

          The Raspberry Pi Foundation is a UK-based charity that works to put the power of computing and digital making into the hands of people all over the world. 

          It really should be easy to understand and if you look at the Foundation page, it has direct product links,

          The 2 entities of foundation and trading are connected, so the absurdity I pointed out is equally relevant.

          The essence of the Pi was also not, as you state, toys for tinkers, but something to promote and inspire educational investment of the young in future tech and coding.

          1. Marketing BS aside… over a decade ago this ‘education’ thing might have been the motivator and the ‘Foundation’ still seems to focus on this (which might actually be a good thing if it prevents morons from entering IT jobs later – I might be wrong since never paid any attention to the ‘Foundation goals’ or whatever else marketing material they sell).

            But the adoption of those boards almost immediately went in another direction: toys for tinkerers first, then with mature software and large community existing a switch to commercial/industrial consumers that might generate the largest income today or even since years.

            Then it really helps to understand business basics: these devices are not made/sold by some charity but made by a company whose sole purpose is making profits. A significant amount if not the most happens now with commercial/industrial customers who base their own products on RPi Trading Ltd.’s.

            So for RPi Trading Ltd. there is no other way than priotizing those customers over hobbyists/consumers or this ‘educational thing’ if they want to avoid going out of business. Since if you base a product on another one’s product you expect this other party to deliver. And you will never ever again choose any of their products to base your product on if you see them preferring charity actions.

            That’s why I prefer to clearly distinguish between those two. It helps understanding the rationale behind things.

          2. And you will never ever again choose any of their products to base your product on if you see them preferring charity actions.”

            Totally agree, I’ve done similar, for example, I still won’t buy Panasonic servo motors after they dumped their NEMA17 line without a compatible replacement.

  2. Some of the approved resellers also jack up the prices by only selling with unnecessary accessories at inflated prices.

  3. Sure, there are plenty of alternatives SBCs out there, but….all the ARM-based ones are kind of a horrible crapshoot. I mean, if the board isn’t supported by Armbian, then the board is likely stuck with some god damn ancient kernel, buggy/broken software and all that. Even if it was supported by Armbian, it’s likely that there’s no support for all the features of the board, hardware-decoding of video is unlikely to work and hardware-encoding is almost certain to not work.

    I’ve pretty much given up on anything other than RPi or full-blown x86-64 based SBCs simply because I want up-to-date software and kernel and I want to actually be able to use the board’s features and not have to spend months fighting with the thing.

  4. One of the EV charger companies had a comment on their Facebook page about buying 10k Pi4s because they couldn’t get 3s.

    1. Not really. Those pirate copies of SBCs invented by RPi Foundation are usually of inferior quality. Sure some say there’s mainline support but e.g. Orange Pi only ships so called FEX images and legacy BSP 3.x kernels. They’re not even that cheap anymore. When looking at the support table in the amateurish sunxi wiki, a lot of cells show lack of support. Good reasons to stick with RPi 4 8GB when you need dual 4k 60 Hz h265 output for 3d movies and such high def content.

      1. One particular problem with sw support is the lack of encryption accelerator support. If you need serious AES 512 bit XTS support and Argon2 for super safe NAS, RPI 4 8GB is your only choice. So called montjoie hasn’t provided full support for all boards.

          1. Imagine that, benchmark software reporting results for hardware features a SoC, does not have.

            Almost like there not, real world results 🤔🙄😜

          2. And the RPi 4 cryptsetup ‘performance’ is as pathetic as one can imagine: https://forum.odroid.com/viewtopic.php?p=261236#p261236

            Both tests with ARMv6 and ARMv7 userlands. We’ve seen halfed AES crypto ‘performance’ on those Raspberries with an ARMv8/64-bit userland so far. One of those RPi Trading Ltd. guys talked about ‘NEON enhancements’ in the meantime but since everything RPi + AES is such a joke I never tested again.

          3. 86,9 / 76,0 MB per sec is actually pretty decent considering that many SATA 3 drives don’t work that efficiently. For example 2.5″ drives may operate a lot slower. Sustained write speed of 50 Megabytes per second.

          4. > One of those RPi Trading Ltd. guys talked about ‘NEON enhancements’

            BS as usual: https://forum.odroid.com/viewtopic.php?p=346464#p346464

            Keeping in mind that my initial tests happened with 1.5 GHz while now running the A72 cores at 1.8 GHz that’s essentially the same numbers. At least 64-bit cryptsetup scores are unlike openssl’s internal benchmark comparable to 32-bit binaries.

      2. There is no invention here. There is marketing. Rk3399 has great mainline support. Far better I/O.

      3. An sbc with 8gb of ram to do something as dumb as to decode 4k, even a radxa zero 1gb, 22 bucks…can decode 4k perfectly fine from Android, lkke mostly any android tv box.

        Also, the vpu acceleration support on rpi, while better than the others, isnt that great neither, at least on aarch64. On armhf is excellent, yes.

      4. The most of (if not all) the Orange Pi boards have full kernel support. I have four Orange Pi PC with Fedora 35 running the last available kernel.

      1. ESP32 could probably run some really cut down Linux (I haven’t figured out the MMU so I’m not sure) but I really don’t think that’s a great use of the ESP32. Linux isn’t great for real-time stuff and that’s something the ESP32 is good for.

    2. Honestly I’d love to see another group compete with Raspberry Pi on the whole package (not just the hardware), but I’m not holding my breath. Too many other groups are laser focused on hardware, and not on ecosystem and support.

      Hardware is hard… software that makes the hardware useful is harder… support is hardest—and it requires huge and ongoing investment, and less time spent just designing the next version of a board based on the next chip from Allwinner or Rockchip.

      1. How should this happen?

        ‘Economy of scale’ is at work at RPi Trading Ltd. who were caught surprised by the success of this ‘board for education’ a decade ago and since then everything has shifted (especially this ‘education’ thing).

        They’re able to pay a bunch of people and can afford to do serious software work (even dealing unnecessarily with this VC crap and masquerading lacking ARMv8 Crypto Extensions and so on). No other SBC maker is in this position.

      2. “I’d love to see another group compete with Raspberry Pi”

        Capitalistic system has it own rules.

        software that makes the hardware useful is harder”.

        Yep. Magnitude harder and HW profit margins can’t cover R&D + support unless you are on the top / Raspberry Pi. Without heavy subsidizing from community projects such as Armbian, most of those competitors would not become usable and never (almost) fully operational.

      3. The SoC makers need to upstream all drivers. It is crazy that none of them do.
        Once that happens, a bunch of SBC makers can ship hardware without having to take on the whole software burden.

        Imagine just booting a generic Fedora, debian, or Ubuntu mainstream .iso. That appears to be where server ARM chips are going. Why not the littler systems?

        If SoC vendors upstreamed detailed specifications, sufficient for the open source communities to develop drivers that might even be enough.

        Reverse engineering is a horrendous waste of programmer time. And user time because of the inherent risk of error. And time-to-market. And yet that’s what we’re left with.

        The fact that no SoC vendor seems to upstream drivers or specs appears to be really dumb.

        Of course some don’t have the rights to do this. For example, most GPU IP vendors don’t allow this. Including ARM.

        I think that the real problem is cultural. The vendors like open source because it gives them free stuff. They don’t realize the synergy of sharing.

        For all the hate of RP, it is pretty open compared with the others.

        1. > The SoC makers need to upstream all drivers. … Once that happens, a bunch of SBC makers

          Do you really believe any of those ‘Android e-waste’ SoC makers gives a sh*t about SBCs?

          Crappy TV boxes, ultra cheap tablets or phones, sound bars, ‘smart’ speakers, surveillance equipment and other Android consumer junk is the stuff their SoCs end up in. None of the device manufacturers cares about software as long as the product somehow works on launch day.

        2. Its because China makers just copy open sauce chips. RPI secret sauce is only they can bulk buy the SoC.
          The software issue is easily fixed, but why help ? The vendors will not thank you or pay you !

          1. RPi has open source support! Emma and Eric Anholt are developing open GPU support. RPi Foundation is developing free mainline kernel support. The userspace tools are open source. Standard ARMv6 kernels and userspace have been available to all boards.

            I think only the GPU’s internal OS is closed, which isn’t really a problem because even Intel/AMD have ME/PSP for monitoring illegal use. I’ve also read Allwinner (favorite pirate board for many) has a OpenRISC spy processor built-in and the boot code is closed. Samsung’s boot code is closed for old Odroid.

            RPi is the most promising platform when it comes to free software support.

    3. Does it need debunking? Educational users and hobbyists get a better cheaper product this way. For example the car charger maker I mentioned earlier fixed a bug in the kernel after updating to the Pi4.

  5. Well, I can sympathize, and hope they catch up sooner or later. I wonder if the original zero and zero w are discontinued: have they made any announcement? Both of those had some attractions over the 2w, and the idea of a $5 linux board was great.

  6. As a side note, yesterday (before this post), I had a company ask me for an alternative processor that could be used as Compute Module 4 replacement with support for 8GB RAM and Linux 5.4 or greater.

    This morning I chatted with a SoM manufacturer saying Raspberry Pi’s claim of “enough supply for business customers” was probably not true, as they got plenty of requests from Raspberry Pi customers who could not purchase CM4 modules.

    1. This could be a huge opportunity for a company like rockchip to get their code mainlined to allow some of their partners to design a CM4-compatible board.

      1. To illustrate your point, the company considered Rockchip RK3566/RK3568, but had to rule it out because of the Linux 4.19 BSP. So instead, I had to tell them to check out Amlogic A311D2.

        Radxa CM3 or SoQuartz (with 8GB RAM) would have been a perfect alternative were it not for the older kernel.

        1. They should run on mainline, without vpu, withou npu, without many things. I know people that works with those carrier boards with plenty of success for many projects involved npu. But, yeah, no mainline, puajjj

        2. We have to be honest, rockchip isnt our paladin, there is no open source soc vendor paladin.
          But compared to broadcom, it’s mostly revolutionary.

        3. > I had to tell them to check out Amlogic A311D2

          Since a BSP based on 4.9 kernel is better than 4.19?

          1. Ah, I missed that comment. Thank you!

            BTW: this is RK3568 running 5.16 and this was RK3566 last summer running 5.13. So depending on the feature set they need (no video at all? Hey, mainline Linux might be an option!) maybe skipping a potentially crappy vendor BSP might also be an idea. Especially when throwing some money at BayLibre, Bootlin and the likes is considered.

    2. What are they doing that actually needs 8gb . Also why do they needs the latest experimental Kernel?

      Its often overlooked that a lot of Linux OS versions do not need 8gb . Not taking servers here.

      There are a lot of products doing plenty of solid work, running stable, on older hardware.

        1. Rk3566/68 does output hw accelerated graphics on mainline nowadays.
          I have a video demonstrating the might power of a mali G52 mp1 2EE…. it’s not amazing hahaha

  7. Don’t buy rock pi 4. The board has the worst power design I have ever and support is horrible. Even verified support engineers in their forum don’t know the components and specs well.

    1. The power design on rockpi 4 is miles ahead from rpi4. It’s far more stable, with PD, from 5v to 20 IIRC. Never had a power issue like on rpi4.

    2. Rpi is a joke on that department , with a very long history on power issues. If you ever had a power issue on rockpi4 it’s bc you didnt use a PD PSU, accessible all around the globe, not like rpi PSU

  8. I must disagree, looking from a specific niche market where the RPI4 is used (powered mostly by 12-14V batteries on wet conditions).

    There are quite some advantages attributed to the RPI4:
    +/+ cheap
    -/+ lots of expansion possibilities
    +/+ lost of expansion boards (shields) available, with example sources
    -/- fast (compared to competitors)
    -/- energy-efficient (compared to competitors)
    -/- possibility to use NVMe drives
    -/- lot’s of users and a very active community (for uptodate distributions).

    The main usage as eductional platform is almost solely obtained by a low price. And in a clean environment with clean power there is no problem with a bad power supply, cooling requirements and quirks about power on USB before the system starts (with a powered HUB).

    Bottom line is, if you can’t cash on price or the expansion boards it is just old hardware with all disadvantages included. You might be better of with the competitors in that case.

Leave a Reply

Your email address will not be published. Required fields are marked *

Khadas VIM4 SBC
Khadas VIM4 SBC