Nope. BOM cost is well under $25. The price includes a healthy component margin, assembly cost (in the UK), assembly margin, test and OTP programming cost. We are professional electrical engineers; we know how to do this stuff (and in any case, it's not rocket science).

Volume assumption is 10K batches. Anyone think I can't sell 10K of these?

Regarding the GPU (BCM2763?):

Can you share any more details on the user space situation? Will there be Xorg/Mesa/Gallium support? Will you be relying on Broadcom's driver team for updates, if any, or do you guys have driver source under NDA with which you can support the GPU in future?

This seems to be the biggest concern for many who are looking at the platform. Just trying to gauge what kind of useful lifespan this board will have with regards to keeping an up to date Linux distro running on it. There's going to be a lot of flux and ABI changes ahead in the Linux graphics stack. Some more GPU details on your Wiki would be great if you have them available.

I'm personally buying three (one for me, one each for the kids). I manage a team of unix sysadmins, you've got quite a few sales lined up there. 10K should go in about.. 5 minutes?

I wish you the best of luck in the venture. I hope it's very successful.

Would you believe me if I told you we're just nice guys, who used to be kids who couldn't afford computers?

Hi Eben.

Sorry, I don't mean to detract from what you're doing. I think it's great. I'm very very excited for the Raspberry Pi, both personally and to see what it does globally.

I'm mostly just curious why the chip isn't publicly available. That's all. I personally hope to be able to use this chip in other designs.

Comparatively few Broadcom chips are available via channel partners. Unfortunately that's just the way our business model works.

Well, if the RP organization is getting them for their boards, why not resell them to hobbyists for cost or a very small margin to fund further research?

"Your business" here refers to Broadcom, when the internet generally thinks you're Raspberry Pi. That's the rub.

Maybe you have to be an old fart... I was here for this last time, but then I remembered that 'last time' was 1981.

This still looks like the most important thing of any kind for three decades to those of us who came out of the cheap home computer wave of the early 80s. (Me, I'm even closer to the Raspberry Pi heritage since I got my CS O-Level in 1987 with an IRC-alike network chat app for BBC LANs.)

So try not to cock it up, eh? :D

I should throttle back my skepticism as well. I understand the intent of the project and I think it's a noble one.

We've already lost an entire generation of potential programmers by not having something like this in the post Apple ][ / BBC Acorn / C64 era. Maybe I'm an old fart by thinking that we don't need a full desktop running at 1080p to make the kids happy, but what do I know anymore.

It's an exciting and ambitious project that only the scale of current SoC systems can achieve. Unfortunately the environment also means licensing of cores and building of these highly complex chips by massive companies like TI and Broadcom. We're all afraid of obsolescence as things accelerate even more, of course, which is probably why we're so hard on Broadcom here.

Actually I could never afford a C64 nor Apple II. ZX81 and Spectrum where the thing for me.

No.

Commendable skepticism :) It's true though.