But with a few million investment you can design and build as a startup. It won't compete with Nvidia or AMD to begin with but the potential is there.

It's looking more and more like Nvidia and amd are going to chase they data center/AI processing cash cow. With an initial investment of 10 million a company could start up and within 20 years be a market leader if they dedicate catering to gamers and the gaming landscape would shift drastically to function with their hardware even if the ai/processing centric hardware is removed.

You don’t need $500 million to start a GPU company; you need $500 million to mass-manufacture a 3nm desktop card. If you use a few million dollars to design a brilliant, streamlined graphics architecture, you can prove it works on virtual hardware, license it to console makers, or use an older, affordable manufacturing node to serve the millions of gamers that NVIDIA is abandoning to chase AI profits.

Nvidia is chasing the data center hardware market and will continue to do so to make their investors happy. In doing so they will create a vacuum in the consumer market with AMD and Intel likely to follow. History has shown that when a vacuum is created a new company will step in to fill the void created and a lean team with an efficient design and the drive to serve the people can succeed.