Qualcomm acquires Arduino

[EEVblog]

In a press release, Qualcomm mentioned that Arduino would become an “independent” subsidiary.
Yeah, right. Color me skeptical.
Arduino will cease to be independent, when the Qualcomm C-suite dictates to Arduino to meet EBDITA levels which are aligned with the corporate expectations. If the Arduino guys can’t, they will then receive the “assistance” from corporate black belts that will be focused on “extracting value”.
It is all downhill from there.

Yep, my money is on that. Might take a few years, but it's inevitable.

[EEVblog]

The only reasonable answer I can think of is that they were having too much money, and bought a name instead of risking that money to inflation or to other causes. 

Cash is negative return in the current market.

[EEVblog]

This monstrosity has no chance. Why even bother with STM32? Just make a decent integrated SBC, and it will have a place and won't need an awkward IDE.

The only real value Arduino has IS in the IDE and the community support.

[EEVblog]

Adafruit have an article "Turning maker dreams into shareholder value":
https://blog.adafruit.com/2025/10/07/qualcomms-latest-ai-play-turning-maker-dreams-into-shareholder-value/

How does this create value exactly?
Arduino has raised $54M in three seed rounds, with the last one in 2022 at $22M.
These investors weren't going to sell for anything less than the 9 digit $100M+ range, probably $200M+.
Arduino's core business is hardware, at what, 1-2M units a year?
Run the profit numbers on that, minus hardware cost and company staff overhead (Crunchbase says 250 employees?) and other costs. That's not exactly raking it in.

And how much of this business is still goodwill buying of the "genuine" boards? And how much of this goodwill will remain now that it's just another corporate monolith? How many will just switch to the clone boards?

I give this 12 months tops before corporate get the financial KPI's and they bring in the hatchet swingers and subscription goblins.

[rteodor]

I don't get it.  Why did they buy it?  What for?
Also, from whom did they buy it?

It might have something to do with with the fact that not long ago Arduino started in the PLC market. Creeping AI along with Opta. Will see.

I can see something out of this: a handicapped Terminator in a wheelchair that will try to kill you but you are legally obligated to give it the parking place.

[westfw]

Will there be deterministic IO response times ?

Arduino has never provided deterministic IO response times (at least by some definitions.  For example: digitalWrite() takes different amounts of time on different pins.)

Of courses, with the AVRs (at least) you could get closer to deterministic by working outside of the Arduino functions.
(That's liable to be more difficult on a linux AP, or even on the Zepher M33...)

[bingo600]

IMHO Arduino is three things

1: HW  - I suppose most buy in China or use STM32/ESPxxx
    I don't care if they switch to QC boards, i just have ONE original Arduino board.

2: IDE  - Terrible ...
    I always try to use makefiles if possible.

3: Libraries  - The most important part.
    This is the important part for me, and i don't see why it couldn't continue to evolve w. AVR/STM32/ESPxxx

I hope that if QC implements the "Oracle steering way", like Oracle did w. OpenOffice / mysql  (SUN aquisition)
That we'll get LibreDuino  ... mariaduino sounds silly 

Or at least some plain makefiles, that are not PlatformIO

Edit: Dave should register - libreduino.org

[Marco]

Qualcomm probably wants to push their ICs more than make direct profit, a response to Raspberry Pi getting into selling micros.

[EEVblog]

A private fork of linux?

https://x.com/geerlingguy/status/1975675619414294983

[EEVblog]

Edit: Dave should register - libreduino.org

Done. 

[RoGeorge]

Maybe Qualcomm plan is to manufacture small/cheap AI accelerator chips, and make them become (for AI chips) what ATmega family was for microcontrollers.  So, to take the concept of microcontrollers (all included in a single chip), and make it AI driven.  Instead of blinking LEDs with a microcontroller, one would blink them with AI. 

Not only that you won't need to know all the technical details, but you can simply ask the AI to blink the LED for you. 

Sure, for today maker's standards, that will also kill all the fun, but who knows.  Same as one can enjoy blinking a LED without having to write a C compiler from scratch, same one might enjoy blinking the LED by simply telling an AI to blink that LED. 

Eventually, Qualcomm can sell all kinds of online subscriptions, so to augment the local/onboard Arduino AI accelerators with extra features that will require cloud processing.

[brucehoult]

Will there be deterministic IO response times ?

Arduino has never provided deterministic IO response times (at least by some definitions.  For example: digitalWrite() takes different amounts of time on different pins.)

Not *much* different. On an AVR I think it's around 3-4µs on a 16 MHz AVR no matter which pin. Maybe there's a couple more µs variation.

On Linux it's probably mostly in the same kind of range on a CPU with far higher MHz, but then you'll occasionally see a ms or 10ms instead.

That never happens on Arduino. At least unless you install some interrupt handler yourself. But then you can calculate (or experimentally determine) a pretty tight bound on the execution time of that too.

[brucehoult]

Edit: Dave should register - libreduino.org

Done. 

There's been "Freeduino" for 15 years.

I think westfw was involved?

[Picuino]

Since Arduino is based on free hardware, free software, and a free community that helps and adds value, its real value lies in the brand, the logo, and what that means to thousands of makers.

There is also one aspect that is not free: the USB ID, a unique code owned by Arduino that cannot be used in “Chinese” copies, although it is done nonetheless.

Edit:

USB VID (Vendor ID): 0x2341
USB PID (Product ID): 0x0043 (for Arduino UNO)
USB PID (Product ID): 0x8036 (for Arduino Leonardo)

[Picuino]

By the way, my nickname comes from a project in which I was trying to add a shield with a PIC microcontroller to the Arduino UNO. Hence the combination PIC + Arduino = PICuino.

Here the project with an LDR connected:


In this project, I realized the importance of having a free compiler (GCC) like the one the Atmega microcontroller had, and how none of the PIC microcontrollers I know have one.

[hans]

This monstrosity has no chance. Why even bother with STM32? Just make a decent integrated SBC, and it will have a place and won't need an awkward IDE.

The only real value Arduino has IS in the IDE and the community support.

.. and, you can buy the Uno, program your sketch, pull the chip out of its socket and solder it on your own veroboard.
The chips are <5$, no worries if you smoke some, or if you want to build multiple units of your project.
The through-hole chips can be soldered by absolute beginners in EE. But even the newer boards with ATMEGA4809 in TQFP-32 (0.8mm pitch) should be OK.

And there are TONS of variants that use the Arduino ecosystem, either through Arduino IDE or PlatformIO. From 8-pin ATTINY to 200+ pin BGA STM32H7s.

Qualcomm seems so far from home in all of this. I haven't seen any of their chips live in the hobby/maker space ever.
I surely hope they don't shitcan all of the relatively recent cool Arduino boards (like Portenta or MKR series) because they don't have Qualcomm parts on them..

[ataradov]

The only real value Arduino has IS in the IDE and the community support.

My point was from Qualcomm's point of view. The CPU they have here is decent, so if they just make a good SBC from it and sell that alone for $45, it would have a place. But actually make a real SBC with a real set of interfaces. No Arduino needed.

Arduino community is great, but only for MCU-based boards. If they expect that community would pick up that Linux based device, then they are mistaken. Arduino community never picked up anything outside of the mainstream boards.

This approach ruins two things at once. You end up with something that does not fit Arduino ecosystem at all, and you also end up with half-assed SBC.

[iMo]

..I for example haven't used a single non-clone Arduino board at all, simply because I've found the clones (of SparkFun Pro Micro, ATmega32U4) and non-clone original boards like Teensys, superior.

Me too.. Worked with all atmegas, stm32fxxx chips with arduino.. The "arduino" is about the sw ecosystem, not the original arduino boards, imho..

[SiliconWizard]

Maybe Qualcomm plan is to manufacture small/cheap AI accelerator chips, and make them become (for AI chips) what ATmega family was for microcontrollers.

Seems clear they want to promote their AI solutions via this aquisition (just like they all currently do).
I doubt that will ever be accessible like small MCUs. As others have said, they don't care about hobbyists. What they care about is to promote Qualcomm via universities, which they are probably targetting here.

So that'll be those SoCs with integrated NPUs running Linux and connected to some MCU externally. Meaning systems that will take seconds, if not tens of seconds, to start, and "abstracting" hardware even more than it was already with the Arduino frameworks.

I don't give this venture much time either before it's dismantled, maybe a couple years at best. Then they are likely to resell it to Softbank.
Rinse and repeat.

And if they don't release documentation of their SoCs (at least the ones they'll use for this platform) to the public, that'll achieve no traction in terms of support and you'll just get what Qualcomm will give you, probably all pretty well obfuscated. The RPi have used Broadcom SoCs for ages now and official documentation was never publicly released, so don't expect anything else here.

[Picuino]

They are different businesses.
A small 8-bit prototype board is not the same as a powerful 32-bit system.
I don't think this is good news for Arduino.
It may end up abandoning its traditional business (Arduino UNO), which is what many teachers of secondary school base technology training on.
For younger students, the micro:bit or Arduino boards are simple enough for them to learn the basics. The Arduino Q is too advanced for that level.
I just hope they keep the traditional small boards, which are the core of the Arduino world.
For 32-bit powerful boards, we already have the Raspberry Pi and similar devices.

[coromonadalix]

is the new Uno Q  is the results of the acquisition ?

https://www.arduino.cc/product-uno-q

[SiliconWizard]

is the new Uno Q  is the results of the acquisition ?

https://www.arduino.cc/product-uno-q

It certainly is. Also shows, as expected, that the acquisition had been in the works for a good while already and that they were pretty sure it would work out, otherwise they wouldn't have invested this time in developing this new board.

This board looks like a SBC and so they are competing with the RPi here. Both chips are pretty good, to be fair. It doesn't look like anything that the average Arduino user would go for, though.

[tszaboo]

https://www.arduino.cc/product-uno-q  - $44, €47

Quad core 2GHz cpu, 16GB storage, wow, as you say this is just a SBC with no ethernet or HDMI port.
Its also clear they don't understand their market at all.

Maybe thats the cheapest processor they even make so they had to find a way to shoehorn it in.

They are probably selling the board at a loss. Someone should make a BOM analysis.
Which is not great news. Typically you could buy the Arduino boards always because they were sold at 500-1000% markup, it was good for the hardware selling business.
The raspberry pi guys really must have pissed them off, when they started selling boards for 4 bucks, when the comparable Arduino micro is 25.

UNO Q’s hybrid design makes it the perfect dual-brain platform for your next innovation. It combines a Linux Debian-capable Qualcomm Dragonwing™ QRB2210 microprocessor (MPU) interconnected with a real-time STM32U585 microcontroller (MCU).

I bet you could make a board with only the STM32U585 that would fit perfectly into the Arudino ecosystem.
So I agree with the ppl here, bad move. I can smell the "synergy" from a kilometer that was the sales pitch.

Qualcomm probably wants to push their ICs more than make direct profit, a response to Raspberry Pi getting into selling micros.

I don't see how this would sell their IC. I searched for the QRB2210. Arrow has it listed, 0 stock, RFQ, MOQ is 168pcs. Nobody who uses Arduinos will design with this IC. Literally nobody. You need a hundred bypassing capacitor and DDR memory length matching to make it work. Instead of a DIP package and maybe a capacitor somewhere.

And Qualcomm actually has some interesting ICs that would fit on this board. They have chipsets for BLE headsets, that have a Cortex M33 or similar core, 80 MHz. Only it's usually a smaller package. But they can offer it in a different package, disable BLE, and have a microcontroller ready with relatively small amount of work and time.

This monstrosity has no chance. Why even bother with STM32? Just make a decent integrated SBC, and it will have a place and won't need an awkward IDE.

The only real value Arduino has IS in the IDE and the community support.

I would say the only real value is their github repo, where you can pull a library that works most of the time.

[John B]

is the new Uno Q  is the results of the acquisition ?

https://www.arduino.cc/product-uno-q

This might be worrying about the curtain colours while the house is on fire, but are they still using that weird pin head offset mistake? Bruh.

[langwadt]

is the new Uno Q  is the results of the acquisition ?

https://www.arduino.cc/product-uno-q

This might be worrying about the curtain colours while the house is on fire, but are they still using that weird pin head offset mistake? Bruh.

far too late to change, changing it would make every shield ever made incompatible