Issue link: https://iconnect007.uberflip.com/i/1532052
68 SMT007 MAGAZINE I FEBRUARY 2025 board. e "golden screw" phenomenon was very real, but at no point through the up cycle was the narrative that inventories are lean. Some of the $270 billion in data that we track—which does skew to the publicly traded companies—show that total inventory days and dollars are still 30–35% above before COVID. Significant margin progress has been made, but there's still a lot more to do. From a demand standpoint, a couple of sub-segments are driving any real consumption. Are there sector-specific trends? e most obvious one is AI, which has been a rising tide that only lis the NVIDIA ship. NVIDIA's data center business was, I believe, $5 billion in quarterly revenues, and it will likely be $30 billion two years later. It's sub- stantial. When you get outside that halo effect, there are pockets: small power management suppliers or interconnect. We focus on the business because they have hundreds of thou- sands of part numbers and tens of thousands of customers they run through the channel. No customer makes up more than 3% of sales, so everyone benefits. AI is real, but it's still not enough to move the broader needle. Looking at industrial, we're in the seventh inning of an inventory drawdown. We're obvi- ously closer to the bottom. But again, inventory is too high. e middle part of next year is the base case for any recovery. at's just the cyclicality that still reigns in robotics and automation. ose megatrends are real, but we built too much inventory. In the medical market- place, it's supplier-program specific. It's still healthy, but we're all getting older and less healthy. We're seeing a con- vergence between consumer devices like health monitors, Whoop® straps, and such, which can collect and share data with a physician. ere's the medical con- sumables business, too, such as RFID tags, and there's the cyclical CapEx-driven business. Aerospace and defense had a couple of really strong years behind a less safe world. We've seen some near-term moderation; it's still growing but coming off the double-digit type growth. I wouldn't read too much into that, though. Defense, in general, is a long-term secular grower. Pentagon budgets emphasize planning for future wars, which are believed to be cyber. You just deal with the near-term effects of budget discussion, fiscal year ends, and things like that. To me, the last of the big North American markets is automotive, which is more prob- lematic. ere is cyclicality in inventory. We are clearly past the early adopter phase for elec- tric vehicles (EVs). at's not to say hybrids as a mix aren't growing, but those growth rates are down-ticking. is reveals a lack of sound roadmap strategy from some Western OEMs that have jumped from the internal combus- tion engine straight to EVs, skipping the hybrid altogether and without building products that consumers want. ere's no doubt that China leads the globe from an electrification of the fleet standpoint— not to say it will be all clean sailing. Right now, there are 170 to 175 OEMs. It's naive to think that five years from now, there will still be 175 OEMs. ere will be some level of fall- out, but those are the two real markets driving it. Outside of that, if you look to the sec- ond part for the North Amer- ican market, you could run through a couple of things. Before COVID, the SAAR (Seasonally Adjusted Annual Rate) was 17.5–18 million in the U.S. It fell into the 14 mil- lion range, which we attrib- uted to the supply chain. e supply chain is now fixed and the SAAR has stayed stub- bornly between 15–16 million. " AI is real, but it's still not enough to move the broader needle. "