4 ms·

Russian University Vows to Build 7nm Chipmaking Tools

ggm 4y ago
Within the bounds of current engineering, nothing is impossible in this space but that said, past attempts to emulate the major fablines at scale have tended to falter. It's one of those "you can't bottle it" things. I'm told most of the tools for VLSI at scale come from a very small circuit of suppliers, many in Europe. This must be what they're trying to emulate. Past efforts to do things like high purity glass for optical stuff worked during wartime and postwar markets but the planned economy ruined it for consistent quality. Heck, from what I read even the Japanese struggle to make lenses these days on the market they have, with all their smarts. Now translate that to atomic scale tech. One bad day with a high cleanliness room and its back to square one. The chemicals for chip production are fearsome. Nine ninths purity. And some are highly reactive. They can probably make small production run stuff, with their own tooling as bootstrap. Volume and rejection rate on die could kill them. Making tooling which can scale up to a commercially viable production house could take them years. Decades even. I have nothing against people who want to build stuff. I don't like the regime they're in, but the effort is interesting. I just think it's a mountain to climb. (I'm not in VLSI. I read stuff. I've read some very interesting private blogs about TSMC and samsung, lost links alas.)
hunglee2 4y ago
Russias manufacturing was completely destroyed by exposure to the free markets during the Yeltsin era, with military manufacturing only being insulated, due to national security considerations. Hard to see how Russia might rebuild back this industry from effective zero, though complete isolation will certainly spur the effort - Iran's ability to produce highly effective combat drones, DRPK with ballistic missiles - good examples of what can be done if the internal market is shielded from foreign competition

ggm 4y ago

Within the bounds of current engineering, nothing is impossible in this space but that said, past attempts to emulate the major fablines at scale have tended to falter. It's one of those "you can't bottle it" things. I'm told most of the tools for VLSI at scale come from a very small circuit of suppliers, many in Europe. This must be what they're trying to emulate. Past efforts to do things like high purity glass for optical stuff worked during wartime and postwar markets but the planned economy ruined it for consistent quality. Heck, from what I read even the Japanese struggle to make lenses these days on the market they have, with all their smarts. Now translate that to atomic scale tech. One bad day with a high cleanliness room and its back to square one. The chemicals for chip production are fearsome. Nine ninths purity. And some are highly reactive. They can probably make small production run stuff, with their own tooling as bootstrap. Volume and rejection rate on die could kill them. Making tooling which can scale up to a commercially viable production house could take them years. Decades even. I have nothing against people who want to build stuff. I don't like the regime they're in, but the effort is interesting. I just think it's a mountain to climb. (I'm not in VLSI. I read stuff. I've read some very interesting private blogs about TSMC and samsung, lost links alas.)

hunglee2 4y ago

Russias manufacturing was completely destroyed by exposure to the free markets during the Yeltsin era, with military manufacturing only being insulated, due to national security considerations. Hard to see how Russia might rebuild back this industry from effective zero, though complete isolation will certainly spur the effort - Iran's ability to produce highly effective combat drones, DRPK with ballistic missiles - good examples of what can be done if the internal market is shielded from foreign competition