You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
I can't see TSM not planning for the quantum computing game. Any thoughts, comments or observations welcome. Also, if not TSM, what companies will dominate quantum computing? What timeline do readers feel the quantum computing trend will follow? Will it end up being a specialized tech or a dominating one? Thanks
I can't see TSM not planning for the quantum computing game. Any thoughts, comments or observations welcome. Also, if not TSM, what companies will dominate quantum computing? What timeline do readers feel the quantum computing trend will follow? Will it end up being a specialized tech or a dominating one? Thanks
Right now TSMC is still focusing on AI. TSMC was early to the AI train so they will continue to lead the charge for years to come. Quantum computing will be big, maybe not as big as AI, but it will be 10-20 years from now, my opinion. TSMC will need to see some quantum chip design starts from big customers before they start beating that drum.
Quantum computing won’t “explode” like AI or the Internet did but it will creep into importance as hardware and algorithms mature. Right now, it’s comparable to where classical computing was in the 1950s: promising, experimental, and expensive, but destined to become essential. Just my opinion of course.
I can't see TSM not planning for the quantum computing game. Any thoughts, comments or observations welcome. Also, if not TSM, what companies will dominate quantum computing? What timeline do readers feel the quantum computing trend will follow? Will it end up being a specialized tech or a dominating one? Thanks
TSMC does platforms for specific market segments. Automotive, Mobile, HPC, etc... Quantum might be under HPC like AI or it may have it's own platform. Either way TSMC will play an important role in the quantum wave.
Come on guys. Quantum computing is a fun research topic, a great one to inspire a generation of bright PhD's. But it has no serious use case with exponential speed up, other than Shor's algorithm (50+ years before we get there?). I guess Tsmc's research group could use it as a way to attract smart physicists/engineers.
It's not yet clear what the technology platform for quantum computing will be. IBM/Google are all going for superconducting qubits for now (see here for review: https://pubs.aip.org/aip/apr/article/6/2/021318/570326 ), which resemble semiconducting chips but with huge critical dimensions except for the ~two Josephson junctions/qubit, which is ~10s of nm in size. These are typically made using ebeam lithography + aluminum ebeam evaporation. At one extreme, silicon quantum dots (Intel) are ~10s of nm in size, and may be well aligned with what fabs typcally deal with. But companies like IonQ propose using trapped ions, and some other groups are working on neutral atoms too. There are also proposals using photons as qubits. These systems won't at all resemble today's semiconducting logic or memory.
That said, if you zoom out a bit, there's going to be a lot of ultra-low-noise/cryogenic electronics needed to support these experiments. There is a role for semiconducting fabs to play here. For example, to control each superconducting qubit, you need one or two RF coax lines going into a dilution refrigerator. Your sample is at < 20 mK, where your dil fridge has about 10-20 uW of cooling power. So your coax cables need to connect room temperature (~300 K) to 20 mK without conducting any heat, including through microwaves. That's extremely cumbersome even at 100 qubits. Some kind of cryogenic RF control electronics inside the dil fridge (possibly on the 4K plate) is necessary as systems scale up.
