4 ms·
There is an opportunity in the market for sometime to come in and make relatively low cost machinery to fab at older process nodes. If someone designed and sold
by thereddaikon 5y ago
There is an opportunity in the market for sometime to come in and make relatively low cost machinery to fab at older process nodes. If someone designed and sold photolithography equipment meant for no greater than say 90nm and it did for pennies on the dollar compared to current EUV equipment then it could be a huge deal for the semi conductor industry.
There are plenty of designs that don't really need to be produced with more modern processes. That are just fine at 180nm, 130nm or 90nm. The natural attrition of old process nodes due to the hardware reaching end of life is unnecessary and I think in part why we are in this situation.
- jopsen 5y agoBut can older processes be made much cheaper than they where when they were launched? Is it an under investment in these processes because the price was too low?
- thereddaikon 5y agoPresumably if a machine can be made with X level of precision for Y cost then one with a lower level of precision could be made for a lower cost. Although it probably doesn't scale linearly. One relevant large advance I know of is in laser. What used to take very expensive chemical lasers in the 90's can now be done with inexpensive solid state lasers with much lower power draw, much higher output and greater reliability.
- jopsen 5y agoMaybe, if someone is making the investment -- which the article suggest isn't happening. After all, if you build such a factory, you won't be able to target the lucrative high-end chips. And if competing factories were built when they were state-of-the-art, and the hardware paid for when they did high-end chips. The problem could be that low-end chips simply are cheaper than the amortized cost of building a factory.