I'v been thinking for a while now and decided I would share my thoughts on the Spresense. I think its great that Sony has created it and provided so much documentation and support, and for what is actually a 'hard core' MCU. I do worry though that it won't gain as much traction as it could, and go out of production sooner rather than later, for the following reasons:
Cost. For almost all of us, the Spresence is not much use without the Expansion board. The 1.8 volt main board has too high a barrier for connecting even 3.3 volt devices without level converters. But of course, that's what the expansion board is for. If I was designing the Spresence product line I would have released a board, as per the expansion board but with a CXD5602 on board. The cost of this could be somewhere between the price of main and expansion boards. Then release a "main board" designed specifically for integrating with custom projects, as plug in to something I'd be able to make with an EDA on a through hole board.
The current main board is not especially usable for anything but the expansion board. I have been contemplating designing a custom board for a project and "plug in" the main board. Most of the I/O I need is only available on the board-interconnect, but these connectors are too fine for anything I could make for a through hole board. Connecting power with USB is not so practical for an integrated solution. There are the battery terminals, but this requires soldering first.
I was originally happy that I could use the Spresense to learn NuttX, improve my ARM development in general, but then would have to target other ARM processors for anything I wish to produce, in small but then hopefully larger quantities. But now I'm getting great results with the GNSS receiver. And I don't want to write something for 6 cores and then have to scale back to one to actually make something which doesn't live on my workbench.
The invitation to contact sales for "large quantity orders" will exclude just about everyone except another large manufacturers.
It seems the Spresense is great for one off maker projects but you have to understand, that's as far as you can take it.
I'm developing a marine robot based on the Raspberry Pi http://robotics.catchpole.net/ . On principal, the Pi is overkill. I could replace most of its functionality, including camera and GNSS with a Spresense. But I worry about locking myself into another platform which has its quirks, rather than just making a fully custom board with an STM32 or something.
My dream CXD5602 board would be something similar to the main board, exposed more of its I/O on standard DIP, all at 3.3v and had 3.3V power input on the DIP.
If you guys make that, I'll happy help you test it.
Anyway, I hope this doesn't sound too much like complaining. I like the CXD5602 and hope to keep using it for long time and heaps of different ways.