Used to be. With the passage of time on the fab business, the work became more and more difficult. There are a lot of old products all over chip manufacturing, and the problem is that upgrading them to make new chips with better methods and technologies (which allow new chips to work better, and are fundamental to technological progress) is very difficult and expensive. Old fab continues to make chips the old fashioned way while parent companies invest in the latest turnkey manufacturing technology and build new fab manufacturing products.
With costs increasing dramatically, an area that used to have dozens of competitors 30 years ago now has 2 or 3. I think it’s basically Samsung, Intel, TSMC these days. Intel is an excellent example of the loss process in maintaining the latest technology. They are really struggling to take the next step and have been “stuck” for many years now that it doesn’t look like they are going to get their next product producing any big returns. This struggle to stay on the edge of technology has broken many companies before, and in the end they gave up and kept making what they could. AMD used to have the Fab but it seems to have gladly gone out of business and is a huge hit now that it relies on TSMC for its top-tier chips.
Intel might figure it out, but I think time is running out. At some point, cost-for-no-return will catch up with them and they’ll just start subcontracting chip production to someone like TSMC just to be able to make a top-tier product again.
Note that Intel is currently still competitive despite its great problems, but the industry is well aware of the problems it has faced and for how long they will last. Their future is at stake.
A great top-tier piece can cost billions. And then they can produce high-end products for a few years, and then you have to build the next product and this will also cost you billions, but at each level they become more expensive.
Edit: One final point, the component sizes on these new chips, which is what these chips should be doing, are getting so small that we’re tackling quantum mechanics and the fundamental limits of physics. That’s why it costs billions. We have to find solutions to break the physics. I am eagerly watching the progress over the next ten years. I wouldn’t be surprised at all if we had to discover quantum computing in our lifetime to continue this kind of continuous progress. Watching a company like Intel struggle with it is a good indication of how ridiculously difficult it is.