[GianFabien]

Unlike software product development, even when using a foundries chip manufacture requires 7-8 figure (USD) development budgets. That is per iteration. Unlike JS development there isn't massive volumes of internet resources to train LLMs on to produce usable RTL, etc code.

[mithro]

Things like my https://wafer.space ($7k USD), TinyTapeout.com (<$200 USD) and ChipFoundry.io (~$15k USD) are making it much cheaper to do IC design.

There are a huge number of designs from Tiny Tapeout which are all public - see https://tinytapeout.com/runs/

The designs are still more in the MCU size, but you have to start somewhere!

The Google open MPW program also had 10 runs with 40 projects published at http://foss-eda-tools.googlesource.com/third_party/shuttle/ -- All the submissions had to be open source and there were 1000+ of those. I did try pitching to multiple Google Research groups that continuing the open MPW funding would grow the available designs which have been manufactured and that was useful for AI training but didn't get any bites.

The now defunct Efabless also ran a number of challenges in this space which got pretty good results, see https://efabless.com/challenges

[FirmwareBurner]

>Unlike JS development there isn't massive volumes of internet resources to train LLMs on to produce usable RTL, etc code.

There is local data on each major manufacturer/designer that they can use to train their LLMs. I'm sure Synopsys/Mentor/Siemens are also working towards providing solutions to their customers can use to train their own LLMs.

[pedro_caetano]

This is pretty much the business model of Adobe in creating their 'copyright compliant' image generation features.

The companies with the biggest treasure trove of non-public training data will likely have a technical advantage. _If_ they can purposefully use that data to create better models for their niche problem domain.

[webdevver]

> chip manufacture requires 7-8 figure (USD) development budgets

im quietly holding out hope for atomicsemi.com or someone like that