Would Tesla roll a new MCU for a $25K car or possibly reuse the one from the 3/Y? The CPU in the current MCU is custom, if they wanted to limit software development cost on it, using the same MCU might make sense as well as economies of scale and what not.

The cameras on a Tesla are used for Front Collision Alert, Lane Departure Warning and Lane Keep Assist as well as Sentry Mode. The backup camera is required by law in many places now. Removing those would limit the number of advanced safety features they could offer.

I wonder how much cost savings would be realized by removing the nonessential cameras? I know automotive grade parts tend to cost more but are we talking about hundreds or thousands of dollars in savings?

Removing the cameras may be a good idea, but note that their extensive fleet of camera-equipped vehicles is an important part of their potentially VERY lucrative self driving software development.

EDIT: I highly recommend watching head of AI at Tesla Andrej Karpathy's talk on their self driving architecture. "The Fleet" as a source of data is a vital functional part of their self driving software development. They may be able to calculate that including cameras and a computer on a car that will ship in high volumes would increase sales of their self driving option across all Tesla vehicles (including that low cost model) enough to justify the inclusion of those parts.

https://www.youtube.com/watch?v=hx7BXih7zx8

Cameras are pretty cheap, the custom board is probably decently expensive, but if FSD is a 10-15k upgrade then it doesn't take much to sell that portion of the hardware at a loss to make it back in software options.

The cameras and machine learning thing is a money maker for Tesla. It will allow them to gather a lot of data, I mean, which Silicon Valley company will consider having thousands of eyes on the road at all times a bad thing?

Plus, expect these cameras to be the basis of expensive options and subscriptions.

Tesla has over 2 billion miles of autopilot driving data. I am skeptical of the suggestion that more driving data will accelerate or open the door to significant progress. I could easily see more data reducing efficiency and increasing overall costs.

"which Silicon Valley company will consider having thousands of eyes on the road at all times a bad thing?"

Any company that doesn't want it or need it? I'm not sure I see the obviousness. Just because something is cool doesn't mean I want my company spending money on it. What am I missing?

You're missing a wide array of applications, from dystopic surveillance to checking the state of a place (damage, garbage).

Sounds potentially profitable to me.

Of course owners should be able to opt out of that or even be paid for opting in.

Insurance..