5 ms·

What has damaged the US chip industry is the overeliance upon old IP for "new" products, combined with lack of investment in industries that would benefit from

What has damaged the US chip industry is the overeliance upon old IP for "new" products, combined with lack of investment in industries that would benefit from more cutting edge tech, thus enabling state owned manufactures in china to eat their lunch in asia in this upcoming decade. You'd have to be really on the pipe to ignore the declining sales volume of derivative products the past couple of years (masked quite nicely by corporates willing to buy back shares and employ a host of reporting tricks) before this trade war™ nonsense.

benj111 7y ago
"overeliance upon old IP", "lack of investment". What are you referring to here? The amount of capital required to develop new nodes means theres only 2 companies left actually competing at the cutting edge. And I would suspect the half life for chip IP is shorter than just about any other manufacturing industry.
- cinquemb 7y ago
  I'm trying to find an interview someone from Horseman Global gave with Real Vision sometime in the past couple of months that went into the percentage of profits from semiconductor companies that came from lines released pre-2000 vs after (not sure on the exact dates) but it followed the 80-20 rule. And I don't think there is a lack of investment in general, just a lack of (successful) investment in companies that could use some of the newer stuff. Obviously they still have their R&D, but it means very little if they don't have successful (profits/revenue, not VC investment zombies like Magic Leap) companies buying their latest and greatest. It would also have been extremely negligent of such companies to have ignored the monetary/economic security aspects regarding Chinese incentives. Last year, China imported about $270 billion worth of foreign semi-conductors, a lot of which came from the US or at least on US patents. Them solving that alone with a domestic industry by any means would increase their asset-liability spread to the USD by nearly 30%, and free up USD to put to work stabilizing their banking system.
  - benj111 7y ago
    "profits from semiconductor companies that came from lines released pre-2000 vs after" If you look at industrial applications though. The machine you start designing today probably will use chips designed at least 5 years ago, if its another 5 years before your machine hits the market, that's 10 years before the end consumer even gets their hands on the product. Then you're going to be selling that machine for 10 years, and that machine is going to be running and getting maintained for what? 10 years minimum. But those pre 2000 aren't competitive from a performance pov. I had a Pentium 333 with 32mb of ram in 2000. That's basically microcontroller territory now, something equivalent should cost me pence to buy now, if I were doing it clean slate. Regarding 'successful' investment. Perhaps? But then there's plenty of capital sloshing around at the moment, I would question why they haven't attracted investment in such a situation. To put it more directly, what do you know that people with billions to invest, and access the experts in the field don't?
    - cinquemb 7y ago
      >I would question why they haven't attracted investment in such a situation. I wouldn't say there isn't investment in new industries/tech upstream, but that it isn't in anything at a point that's driving consumer demand/profits. >To put it more directly, what do you know that people with billions to invest, and access the experts in the field don't? I'm interested in brain computer interfaces (worked/consulted in multiple research labs, published in journals, familiarity with all the politicking in between academia-industry/etc). I have hands on experience with medical imaging hardware/software and design of systems/sensors, and have seen a couple of great but really expensive systems/companies only target researchers, and many more trash systems/companies with even more funding that continued to get peddled to the public (and research labs sadly), and the only VC firm I've seen that even seems right up the ally with interest in something like this (some of the people behind it), nearly 15-20% of their portfolio companies are filled with corporate hiring startups (and another chuck companies with gov contracts for bread n butter [i.e. easily financed with 4-52 week notes]). If one follows Robin Hanson's take on our economic system (prestige bias wrt to experts, and how it can go horribly wrong in certain cases outside of wanting to associate with such people), its not hard to see how $10-100 billion or more of other peoples money mostly being burned…

benj111 7y ago

"overeliance upon old IP", "lack of investment". What are you referring to here? The amount of capital required to develop new nodes means theres only 2 companies left actually competing at the cutting edge. And I would suspect the half life for chip IP is shorter than just about any other manufacturing industry.

cinquemb 7y ago
I'm trying to find an interview someone from Horseman Global gave with Real Vision sometime in the past couple of months that went into the percentage of profits from semiconductor companies that came from lines released pre-2000 vs after (not sure on the exact dates) but it followed the 80-20 rule. And I don't think there is a lack of investment in general, just a lack of (successful) investment in companies that could use some of the newer stuff. Obviously they still have their R&D, but it means very little if they don't have successful (profits/revenue, not VC investment zombies like Magic Leap) companies buying their latest and greatest. It would also have been extremely negligent of such companies to have ignored the monetary/economic security aspects regarding Chinese incentives. Last year, China imported about $270 billion worth of foreign semi-conductors, a lot of which came from the US or at least on US patents. Them solving that alone with a domestic industry by any means would increase their asset-liability spread to the USD by nearly 30%, and free up USD to put to work stabilizing their banking system.
- benj111 7y ago
  "profits from semiconductor companies that came from lines released pre-2000 vs after" If you look at industrial applications though. The machine you start designing today probably will use chips designed at least 5 years ago, if its another 5 years before your machine hits the market, that's 10 years before the end consumer even gets their hands on the product. Then you're going to be selling that machine for 10 years, and that machine is going to be running and getting maintained for what? 10 years minimum. But those pre 2000 aren't competitive from a performance pov. I had a Pentium 333 with 32mb of ram in 2000. That's basically microcontroller territory now, something equivalent should cost me pence to buy now, if I were doing it clean slate. Regarding 'successful' investment. Perhaps? But then there's plenty of capital sloshing around at the moment, I would question why they haven't attracted investment in such a situation. To put it more directly, what do you know that people with billions to invest, and access the experts in the field don't?
  - cinquemb 7y ago
    >I would question why they haven't attracted investment in such a situation. I wouldn't say there isn't investment in new industries/tech upstream, but that it isn't in anything at a point that's driving consumer demand/profits. >To put it more directly, what do you know that people with billions to invest, and access the experts in the field don't? I'm interested in brain computer interfaces (worked/consulted in multiple research labs, published in journals, familiarity with all the politicking in between academia-industry/etc). I have hands on experience with medical imaging hardware/software and design of systems/sensors, and have seen a couple of great but really expensive systems/companies only target researchers, and many more trash systems/companies with even more funding that continued to get peddled to the public (and research labs sadly), and the only VC firm I've seen that even seems right up the ally with interest in something like this (some of the people behind it), nearly 15-20% of their portfolio companies are filled with corporate hiring startups (and another chuck companies with gov contracts for bread n butter [i.e. easily financed with 4-52 week notes]). If one follows Robin Hanson's take on our economic system (prestige bias wrt to experts, and how it can go horribly wrong in certain cases outside of wanting to associate with such people), its not hard to see how $10-100 billion or more of other peoples money mostly being burned…