> If you're doing cheap IoT trinkets, you're never going to pay extra for a brand.

Except for the Arduino brand. Arduino boards have margins that traditional hardware vendors can only dream of achieving. The only thing carrying that profit margin is the Arduino brand. The software stack is not tied to their hardware, but they make tons of money on hardware.

> The problem is that there just isn't a whole lot of money to be made in providing hobby hardware for enthusiasts.

With Arduino, the hardware is probably the least interesting/important part. The software side is more important, providing an easy-to-use IDE and a simplified API and platform abstraction layer to make it super-easy to get started. Then there's the documentation, sample code, and community.

> A decade ago, Intel tried that back in the day with Galileo / Edison, and tellingly, they came up with the same "ideas": IoT / AI.

Intel's execution - as usual - was poor and lacking.

Both the Galileo and Edison were much more expensive than their Arduino counterparts, and their x86 cpu's were of little value within that space (especially at the time). Neither made it past 5 years without being killed - which is exactly what people feared. A stunning lack of long-term commitment from Intel to develop and grow a community, leaving anyone that actually built products based on their devices holding a useless bag.

> I suspect there was an internal deck saying how this acquisition is going to give them foothold in the hobby community,

No. This is to give them a foothold in the "IoT hammer" manufacturing business. They looked at how the Raspberry Pi went from cheap hobby computer running Linux to low effort rapid prototyping embedded platform that can run a full web stack. They want to be part of a full dev pipeline from prototype to product.

The real target audience are people building things who don't care how it works as long as it works. So expect 99.9% of these projects to use some sort of Python or JS thing running in a container on the Linux while the microcontroller runs a few lines of c to manipulate IO pin state from the Linux thing. Just like all those abandoned Spin scooters in Seattle that had raspberry Pi's in them. That is the market they are after, not the person who builds a one-off Arduino fish feeder.

I wonder if even inside the hobbyist space, Arduino got obsoleted by the Raspberry Pi and its clones/compatible devices.

Basically, if you already got the skills to work with "bare" microcontrollers, you won't need all the simplification and handholding that Arduino provides and you can just buy the individual chips and fully utilize the tiny form factor and low power requirements.

If you want to learn programming microcontrollers, then locking yourself into Arduino's abstractions is probably counterproductive.

On the other hand, if you do want to just combine different ready-made modules, focus on programming and don't want to worry too much on the low-level stuff, you will probably use a raspberry pi or similar: The form factor is only slightly larger than an arduino, but you get a full-fledged PC instead of a microcontroller.

So I don't really see a niche there.

I agree. The Arduino brand isn't for professionals.

But let's say tomorrow they come together with bundle/partnerships to create a new, great dev environment, very easy, that a mechanical engineer can prototype a great robot for a niche use case,and continue to use that chip and code, with some changes in V1 production ?

Is there value to the Arduino brand and community than ?

There is a whole lot of commerical products built out of what we consider hobby projects (Adruino, Raspberry Pi). Eg: digital displays, industrial equipment controllers etc. All of this is clubbed under the nebulous IoT moniker.

My take: Qualcomm hopes to leverage Adriano adoption to expand their IoT share, and also to grow Adruino's footprint to include more smart IoT devices using Qualcomm's chipsets (Eg: Robotics)

I think it might be related to them charging say $100 instead of $5 for the device and providing "lifetime" (read: "indefinite") access to their IoT Cloud. Except there are no guarantees on the duration of that access.

As a side note, I don't get why they can't find the NPV of actually lifetime cloud compute. Compute costs are decreasing rapidly, so a $5/yr perpetuity has a NPV of $185 assuming 2.7% inflation?

> And if you're doing serious AI, you basically go for a real computer with real computing power, and in that segment, the Arduino brand means nothing.

What about cheap AI for toys and gadgets? Maybe the next Furby or some smart Toaster could run on their chips. AI is spreading, moving into casual corners outside of hobbyists and high professionals, maybe they aim to get a foothold there?

>If you're doing cheap IoT trinkets, you're never going to pay extra for a brand. You're going to buy the cheapest wifi / BT chipset out there and make do with that.

It's the opposite of that. Hobbyist/low volume maker gonna spend extra money to buy a familar tool, instead of going extra miles finding the cheapest available.

Even ESP32 is bad in term of perfomance/features and how much it cost.

Maybe. New people means new perspective. Maybe they see value in an ecosystem of developers who are keen to spend their free time to drum up interesting content, based on their projects and applications. This grassroots interest is what drove Apple to displace Sun Microsystems as the de facto, UNIX system.

Not everything in life is about money.

I mean, if you have seen RasPi prices lately, I'm not so sure this is true. Seems like a really profitable biz..granted, I wouldn't pay their absurd prices for such underpowered hardware. Virtually nobody should buy their $200 CM5 product for example.