Pi Zero unit economics

[jamesh]

I suspect the bottleneck for production at the levels originally suggested will be manufacture of the 2835. Making millions of those per month would take up a lot of fab time. Similar to when Apple produce a new phone, the fabs around the world pretty much just do chips for those! (Slight exaggeration - but try and get a chip made in the months before an Apple launch. Not easy!)

Of course, you can build more final production lines, even new factories just to build Zero's. But all very expensive, putting the price of the device up. And of course, that sort of funding needs to come up front, and with the minimal profit margins on the Zero, that money might be difficult to find (Kickstarter not an option I suspect).

Best bet? Design your own board dedicated to what you want, use an off the shelf Allwinner, or depending on volume, maybe a Broadcom chip like the 2835 or similar. Lot's of ARM devices out there.

SW Support is the main issue. You are not getting it for free like you are with the Zero.


With regard to schools, I think the Pi3 is a better bet for schools. Take the extra price hit, but get a lot more functionality, and ease of use. The Zero is too fiddly for most classrooms.

[curious4872]

I'd like to address two of the posts here.

Regarding the policies mikerr talked about:

IMO its more to do with RP's politics of the zero

Must keep $5 price
Must go it alone without RS/farnell
Must keep production in UK, not china

Now if those were to change we could have a real available product (like the pi3)

I don't object to these goals, I think they're not bad! At the same time that they do the above, if the unit economics makes sense they could always make a single one-time exception to fulfill one single PO, especially if it means they are receiving a $1M donation - baked into my kickstarter idea. That can help them exceedingly well in their educational plans, which I support. It's around the size of the educational fund they released not so long ago, or if they used it as part of their budget it must be a large part of it.

If you look above, my calculations include a $1m bribe to them, and when you're an educational charity that is hard to say no to, especially if all you have to do is fill just one PO to get it. So, as an exception they could choose to fill a single PO like that, while continuing to follow their policy elsewhere. After all even girl scouts (a charity) sell cookies. It can also be a way to get to some distributors, I mean things like smaller computer shops in college towns, since 680 is something they might sell over many months. For small hobbyist shops it it only ties up $5,000 in inventory, the price of a few laptops, and would be probably something they can sell over time. Without a kickstarter program, the foundation don't REALLY have any idea of the total demand, and it's hard to expect them to take a risk sight-unseen, but with a kickstarter they could at least definitively handle a single PO.

I'd like to reply to this other comment, by alexeames:

Adding to what gkreidl said... Also - don't forget Eben has publicly stated (on the Pi Podcast this week) that the Zero is not guaranteed to be a "fixed" form factor. They reserve the right to mess with it. They are not messing with the form factor of the other Pis.

Well, this would be another good thing about a kickstarter campaign that the foundation did as a one-off thing: by having an exact fixed commitment (whatever the end result of the kickstarter is) they would not be required to continue to manufacture those boards later, they can do it as a one-time thing.

I also completely agree with your reasoning for the simple reason that they might want to update and change the boards! After all there is very good reason to think about certain changes: for example multiple people thought of an audio jack, people gave a lot of thoughts around the USB limitations -- and even smaller things such as the speculation in this forum and elsewhere about what the "surprise" was in the new raspberry pi showed a good level of interest in updates soon, some interesting ideas from people, which might be incorporated. So a one-off kickstarter or something would not tie them into anything, including large support contracts since there is a maximum size (1694 units) which is very low. The numbers I listed are more like people who have small hobbyist stores or that sort of thing, or might want to sell to small schools, not for huge manufacturers wanting contracts for mllions of units.

But all this is only possible if the unit economics makes sense, which may not be the case. It might not be possible to accept such a PO even if it came with a $1 million charity contribution and even if it was a one-time thing, if a kickstarter succeeds. I'm curious what they think.

[fruitoftheloom]

Curious I have read your various posts and over the last 4 years many similar ones.

From a layman's point of view I would say you are trying to get RPT to support your product which is not feasible.

There several similar products available and have they made a good profit ?

So I believe you best bet is to base your product around a Compute Module as others have done.

Finally a Forum like this is not a place I would discuss creating a Product IMO

[mthomason]

Another option that comes to mind that might be worth investigating would be approaching the Foundation and asking to license the Pi Zero design for your project. You pay them royalties, and they hand over the specs so you can go and manufacture your own. That way they don't have to worry about the additional manufacturing issues. You'd likely also have to negotiate with Broadcom directly for the processors and support (especially for any documentation not available to the public that you might need).

You could also go to Element 14, who apparently are set up to do production runs of customized Pi boards (including prior generations), so you could take, say, a Pi 1 A and strip it down to the exact specs you want.

I have a feeling the Zero, while not being produced at a loss, is being produced close enough to cost that they don't really have any direct interest in them being used in other commercial or large-scale projects - rather they're being made purely to get cheap $5 computers into the hands of individuals (another reason why I suspect that one-per-order limit is going to stay over the longer term). They could possibly have a vested interest in ensuring anything else is channeled into buying the other Pi models/going via E14 for custom boards, in order that they actually make a profit off those orders.

[Cancelor]

... my calculations include a $1m bribe to them, and when you're an educational charity ......

Charities are not allowed to accept bribes.

... I'm curious what they think.

I don't believe you. If you were serious you would ask them.

In light of the above you look like a troll to me. It's high time this thread was locked.

[fruitoftheloom]

Another option that comes to mind that might be worth investigating would be approaching the Foundation and asking to license the Pi Zero design for your project. You pay them royalties, and they hand over the specs so you can go and manufacture your own. That way they don't have to worry about the additional manufacturing issues. You'd likely also have to negotiate with Broadcom directly for the processors and support (especially for any documentation not available to the public that you might need).

You could also go to Element 14, who apparently are set up to do production runs of customized Pi boards (including prior generations), so you could take, say, a Pi 1 A and strip it down to the exact specs you want.

I have a feeling the Zero, while not being produced at a loss, is being produced close enough to cost that they don't really have any direct interest in them being used in other commercial or large-scale projects - rather they're being made purely to get cheap $5 computers into the hands of individuals.

FYI Odroid produced 5000 W Models and then found Broadcom blocked any further sales of the 2835 so I would say it is unlikely BCM / RPT would changes this stance:

http://www.hardkernel.com/main/products ... 0610189490