Tesla abandons quest for one-piece 'gigacasting' | Hacker News Reader

Tesla abandons quest for one-piece 'gigacasting'

autoblog.com

11 pointsksec9 comments

> But Tesla has since halted the effort, opting to stick with its more proven method of casting vehicle underbodies in three pieces: two gigacasted front and rear sections and a midsection made of aluminum and steel frames to store batteries, according to the two sources familiar with the matter. That is largely the same three-piece method the company has used for its past two new models, the Model Y crossover SUV and the Cybertruck pickup.

I feel like three piece "gigacasting" still has to be advantageous over the normal process that involves hundreds of pieces.

from the article: Such moves also reflect a fundamental strategy shift, with Tesla now focusing more on developing self-driving vehicles than on pushing for huge growth in EV sales volume, which many investors had been counting on.

Sheesh. Insanity. Even if it works 99.9999% of the time it's not good enough for "FSD" let alone "robotaxi" unless you're referring to L4 autonomy.

What a goon.

Sounds like the $47bn Musk wants could better have been spent on superchargers and this.

Or the $44bn he spent on Twitter.

What happens when a gigacasted underbody is damaged? If it's all just one big piece I've got to imagine any repairs are orders of magnitude more expensive than replacing a few interchangeable parts.

Since it's structural, even if it was a bunch of small parts, if those were damaged, it would still be a writeoff in most cases. As the labor cost to tear apart the car and put it back again piece by piece is huge.

Cast parts are pretty strong and typically you'll put some energy absorbing fixtures in places likely to get hit, attached in a way that's cheaper to replace.

Once you get into the innards of the car, if it's dealing with enough energy to deform it, you're looking at writing the vehicle off anyhow.

My car's got molded sheet metal welded together to make the core of the car; if it gets hit hard enough to wad up the shock towers holding the front suspension in the right place, the car's toast. If my car's shock tower is cast, and it gets bent, the car's still toast, but it likely took a lot more energy to be bent/broken.

if your robots are as efficient as you say they are then there is no need to replace them with gigacasting, what's a few extra pieces to weld together?

"gigacasting" is akin to semiconductor lithography vs discrete logic.

A cast part needs a bit of finishing machine work and validation, but is otherwise "liquid metal goes here / finished part comes out" -- very very fast and space efficient.

Welding things together introduces dimensional errors, is slow, messy, and requires tons of space.

But like semiconductor lithography -- the bigger the output the lower the yields. At a certain point it probably makes sense to just make 2-5 parts and bolt them together vs 1 big part. (See also -- chiplets)

Reply on news.ycombinator.com