Record 1 I recently reached a sort of record that I detailed in my post The 500th Breakfast Bytes Post . Record 2 On my way back from CDNLive Boston, I was in Boston Airport (well, what a surprise). The way United works these days is that they divide you into groups by your status. If you are platinum or above (or in first or business class), you are in group 1, gold is group 2, and so on. I am platinum. I actually flew 130,000 miles last year so I should be 1K, but I'm so good at keeping the cost of flights down (you can thank me Jeff Ribar, in the unlikely event you read Breakfast Bytes) and I didn't spend the $12,000 hurdle. On that flight, half the plane seemed to be in group 1. Two people next to me complained at how group 1 seemed to be getting larger and larger. By way of conversation, I talked about the second record. I said that during the flight I would pass 2,000,000 lifetime flown miles on United. The other two said they were over 1,000,000 but not close to 2M. One of them asked me what I did an I said somewhat vaguely that I was in semiconductor. "What company do you work for?" he asked. I could tell from his accent that he was English. "Cadence," I said. All of us know that not that many random people in somewhere like that have heard of it. "What do you guys think about all the consolidation of your customers? What do you think the equipment vendors think? Companies being hoovered up by Hock Tan?" If you don't know who he is, see my post The Consolidator Makes a Rare Appearance: Hock Tan's Top Ten List . I was somewhat stunned that he instantly knew who Cadence was, and some of what is going on in the semiconductor industry. Hock Tan is hardly a household name. "Well, the number of engineers is not going down. The layoffs in this sort of consolidation tend to be in G&A not engineering." If you want some numbers, see my post Merger Mania , in which a Wally Rhines (who else?) keynote is full of data. "Another thing," I continued, "is you can't do design without design tools and equipment." "That's true," the guy said. "Somehow everyone's margins should be higher." "Yeah, well the supermarkets would love to use the same principle. We all have to eat, after all." "I suppose Instead, the customers make all the profit." "The semiconductor industry isn't that profitable either. It is companies like Apple that make all the profit. So are you in some sort of finance?" "These days I'm a partner at Bain capital. We invested in NXP and Broadcom. That's one reason I sort of know the industry. Arm seems to be doing fine." "They are part of Softbank now, of course." "I know. I bought shares in the IPO. Herman Hauser is a good friend of mine." Herman Hauser was the founder of Acorn, the company Arm was spun out of. "Since it was stock I never intended to sell, partly out of patriotism, I took a paper certificate. I think Herman and I were the only two who did. I framed it and put it on my study wall. So I had to take it down and unframe it to cash in at the aquisition. My wife said that if she'd known how much it was worth, she'd have taken it down long ago." "I was present aat the meeting when Arm was crerated," I said. "I was at VLSI Technology, and we gave them all the tools they needed and some prototyping for, I think, 10% of the company." "VLSI, didn't they get acquired by someone?" "Yes, your NXP. Although they were still Philips Semiconductors then." "Has the NXP acquisition by Qualcomm gone through?" I laughed. "Shouldn't I be asking you that?" "Oh, we got out of NXP ages ago. We went in when they were nearly dead and got out when they were thriving again." "I have another Arm story," I said. "When Acorn decided to design the ARM, they used VLSI Technology tools. But this was so early, maybe 1983, that we didn't yet have any offices in Europe. So as the token Brit I was sent to do the software delivery and installation. I wasn't even an application engineer, I was one of the programmers working on the tools. I installed the tools the ARM1 was designed on. And wrote a couple of them." "Herman has a funny story about that era. When Acorn realized that they needed something better than the 65C02, they looked at everything. When someone suggested to Herman that they design their own processor, he thought it was a bad idea. But if you want to take four guys for three months to see if it is even possible, then you have them." At that point, the third person, said "Back before the end of the Soviet Union, I used to design clones of Intel microprocessors." He said something about what crappy process they had in the USSR back then but forget the number. But then when I started doing design it was 3um and 5um. Or 3000nm to 5000nm by today's standards. We entered the plane. The investment guy went up to first class and I never got his name. I went to the back of the plane (well, in what United calls Economy Plus, where there is a seat pitch that makes working easy. But it was an off-the-wall conversation for three random people in line. And since the plane made it to San Francisco, I am now officially a 2 million mile flyer. An Interesting Datapoint I mentioned Wally Rhines above, which reminded me that when I was planning to go to Hot Chips, it occurred to me that a modern 7B processor chip must have as many transistors as the entire semiconductor industry at one point. The only way I could think of to find out easily was to ask Wally. I shot an email off to him and asked, saying my guess was in the mid-1970s. It turned out that the industry was that big a lot earlier. I got a reply from Wally later that day: Paul, 1965 ... 2.45 Billion Transistors, Source - SIA 1966 ... 7.01 Billion Transistors, Source - SIA Wally If you don't know, the SIA is the Semiconductor Industry Association. When fabless semiconductor companies started to appear, the SIA refused to let them join They were not "true" semiconductor companies. In fact it was in that era that Jerry Sanders (CEO of AMD) made his famous statement that "real men have fabs." So the fabless companies created their own association, the FSA or Fabless Semiconductor Association. It changed its name and became GSA. Ironically, AMD was the first major semiconductor company (what we now call an IDM...or just Intel) to sell off its manufacturing and get rid of its fabs. It sold them to Advanced Technology Investment Company (ATIC), the investment arm of the state of Abu Dhabi, to create GLOBALFOUNDRIES. As Wired magazine put it at the time, "so the guys from Abu Dhabi are real men." The Exact Moment Sign up for Sunday Brunch, the weekly Breakfast Bytes email.![]()
↧