Comment by magtux
7 days ago
This is a recipe for disaster. Arduino is great for education/tinkering. Qualcomm won't sell you anything even if you are ready to commit to buying 1000s. I tried to source some Qualcomm chips for a startup @ 10k qty and was told there would be no information or support. Qualcomm can have a much bigger market if they simply open up some product lines for distribution like MediaTek do.
China has a way more vibrant, innovative hardware industry simply because you can source everything made by Chinese firms.
This is the exact reason why they bought Arduino... So now startups have a way to buy say 1,000 devices for prototyping. Qualcomm gets used to supporting smaller developers/startups/tinkerers and will hopefully push different types of chips into the Arduino product lines.
The issue is that their corporate culture does not support it. Arduino will be too small to matter. This is the same issue as with Coral, the Google TPU. They are not refreshed as they are too small. They are too small cause they are not updated or supported widely.
People need mainline kernel support and regular refreshes to reliably build projects based on it. This will require some level of building their BSPs in open and providing APIs for people to take advantage of the QCOM specific features. A QCOM that won't talk to anyone without an NDA cannot adapt to this.
This was very much my experience going through an acquisition like that. I was working at a company that served big customers. We bought a smaller company, with one of the goals being to expand to serving smaller customers.
What actually happened was that our management very quickly started telling the people who came along with the acquisition that they were doing everything wrong. The salespeople were selling wrong, the marketing people were marketing wrong, the customer support people were supporting customers wrong. Everything that the company we acquired did differently was seen as a problem.
Within about a year, anyone they hadn't pressured to adopt our practices had left and been replaced with a transplant from the Mothership. Another year later, the customers we picked up in the acquisition were rapidly leaving for other vendors. They simply couldn't work with us in a way that worked for their business anymore. Last I heard, pretty much the only remaining vestiges of the company we acquired were trademarks, and we were back to only having very large customers.
1 reply →
> This is the exact reason why they bought Arduino... So now startups have a way to buy say 1,000 devices for prototyping. Qualcomm gets used to supporting smaller developers/startups/tinkerers
For this, Qualcomm does not have to buy Arduino for a big amount of money: Qualcomm could simply offer this option on their own and save the acquisition cost.
Addendum: For the acquisition cost, Qualcomm could do a lot of marketing of their offering towards makers.
> Qualcomm could simply offer this option on their own and save the acquisition cost.
No they can't. That's like suggesting "the aircraft carrier could simply turn around." The cheap and simple way for a multi-billion-dollar secretive semiconductor manufacturing behemoth that doesn't know how to write a contract for less than a million dollars or to publish documents for the public is not to just change that. It's to write a contract for millions of dollars to buy someone else that can already do that.
3 replies →
Seems like a corporate version of the "buy vs build" question. If it's true that the goal is to become more approachable to students and hobbyists (which personally I think would be a good idea) - then Qualcomm must've evaluated both options and decided "buy".
The problem is that massive semiconductor companies like Qualcomm rarely follow through. They want their lottery ticket to be included in the next smartphone revolution but won't care about random under 5k unit Kickstarter. Everyone knows that those are two sides of the same coin, but they always choose to wait for the bankruptcy trustees to show up.
Seems like a similar play to what Broadcom did with Raspberry Pi— create a new entity/brand that could resell their chips on hobby boards and be stewards of a "community" support framework but largely without distracting the company from its enterprise customers or risk cannibalizing those relationships.
That said, interesting that Qualcomm would buy twenty years of Arduino legacy for this rather than launching something new in the space.
Broadcom wasn't really a driver behind Raspberry Pi. They acquiesced and let them have chips, once the product took off, they continued supplying chips for the Pi. And of course supported the community by refusing to supply non-propriety firmware blobs to this day ;)
Other than that, Broadcom never really had any community involvement, nor any involvement in the Raspberry Pi Foundation that runs it. However, some broadcom engineers were part of the foundation, which isn't quite the same.
> interesting that Qualcomm would buy twenty years of Arduino legacy for this rather than launching something new in the space
I wouldn't minimise the effect of people just googling around and finding the name Arduino all over the place. It would be very hard for an entirely new platform to get critical mass while esp32 is not standing still.
As other have pointed out, that doesn't make any sense. Qualcomm doesn't want anyone buying their products at 1k quantities for prototyping. They want huge customers that place huge orders consistently. The return for supporting those small orders is miserable and doesn't align with their business objectives
Qualcomm has bought plenty of companies that serviced small customers, and what happed is exactly what the person you’re replying to described. You can’t even get a quote many times.
What I expect short term is what happened to Eagle in the PCB space when Autodesk bought it (best thing that happened to kicad).
Longterm Arduino goes into the periphery of the maker market, similarly to beaglebone.
Qualcomm did not need to buy Arduino in order to do that.
It's cheaper and easier to just spin your own boards at that point. Arduinos are not complex or special in any way. Even if you did need a ton of off the shelf boards, there are countless clones that will sell you as many as you want for next to nothing.
Plus the market you're implying exists is so small as to be utterly worthless to Qualcomm. They are in no way interested in individuals or small businesses
> Qualcomm gets used to supporting smaller developers/startups/tinkerers
I'll believe it when I see it
From the presser:
> Entrepreneurs, businesses, tech professionals, students, educators, and hobbyists will be empowered to rapidly prototype and test new solutions, with a clear path to commercialization supported by Qualcomm Technologies’ advanced technologies and extensive partner ecosystem.
At the least the official line is to remedy this situation. Could be embrace/extend/extinguish but tech companies spend all kinds of money on getting students and smaller businesses into their monolithic ecosystems.
The data center AI race was won by nvidia, embedded AI might still be up for grabs and it helps to have developer adoption.
Yeah, I was going to say this is like the worst-case scenario for the average user.
One of the benefits of the main Arduino line is it was very simple to convert to your own design. Companies like Broadcom and Qualcomm won't sell (many of) their chips on normal distributer sites.
Same reason why Raspberry PIs kind of suck in my opinion. Great you've come up with a neat thing you want to build with it; you are forced to utilize either their compute modules which may not be sufficient for your task, or might be out of stock, or XYZ.
Yep. Same here but dramatically lower quantities. Was told basically we'd have to pay for a partner's support. Not that I'm expecting better from Arduino, but the community makes up for it. You Google "dragonwing stackoverflow" and there are 604 results, but even the first few aren't remotely relevant. "Atmega328p stackoverflow" is over 14k and relevant. Arduino is 52 million. It's a nonstarter
Arduino is great for education/tinkering
Arduino has been trying out a new "pro" line for about a year now, making PLC-level devices to be used in automation but hopefully attracting developers by letting them use the same family of tools as the educational line.
https://store-usa.arduino.cc/collections/pro-family
10k? Sorry, pal! We're in QualcommLand, and you need to be at least 100k units tall to ride this ride!
But if you are a small developer, there are options for you! Have you tried to: eat shit? And die? So that you don't insult our PRECIOUS FUCKING TIME by IMPLYING that a MERE 10K would be ENOUGH for THE GREAT QUALCOMM to ACTUALLY CARE?
The optimist in me wants to believe that this acquisition is a sign of Qualcomm actually trying to be better than that. But realistically? Yeah no. It's Qualcomm. They wouldn't have let it get this bad if they ever cared.
afaik, Raspberry Pis move around 7M units annually.
Based on their first announced product (https://www.arduino.cc/product-uno-q), I think Qualcomm is trying to get into that space, and they bought Arduino for the brand name.
You're right that Qualcomm isn't in the business of small business. But maybe they expect that the market is big enough that it's worth their while to pay a subsidiary (Arduino) to do it.
Yet another example of how corporate consolidation in America is hurting the consumer. FTC needs Lina Khan back to break up the oligopolies.
I had the same knee-jerk reaction, but the optimist in me wants to say "isn't that the point of the acquisition?" Another comment linked to the Uno Q, which looks like a Qualcomm dev kit by Arduino. Perhaps Qualcomm is trying to get better at exactly the kind of thing you're talking about.
s/1000/10,000,000/