Published on: June 17, 2026
The Chip Market Has Split in Two: MCUs Are Still in Allocation While Analog and Discretes Loosen — Where Spot Buyers Should Move Now
Mid-2026 isn't a blanket shortage — it's a two-speed market. AI-edge logic and automotive MCUs are locked at 26–40 weeks, while analog, discretes, mature-node logic and NAND quietly soften 5–15%. Anyone still sourcing on a single "chip shortage" headline is losing on both ends. Here's the spot-desk read: what to chase, what to wait out, and how to turn dead stock into cash.
Bottom line first, to save you time.
At this point in mid-2026, asking "is there still a chip shortage?" is the wrong question. The market isn't one board moving at one speed. It's two, and they're moving in opposite directions.
One side is locked. AI-edge logic, advanced nodes below N7, automotive and industrial 32-bit MCUs — these are still on allocation, 26 to 40 weeks. MCUs are the stubborn one. The 28nm–40nm automotive parts sit behind AI, comms and consumer in the wafer queue, so automotive and industrial get rationed. NXP, ST, Infineon, TI and Renesas are all shipping on quota.
The other side is loosening. Analog, power management, discretes, mature-node logic and NAND — lead times have dropped back to 8–20 weeks, and pricing is down 5% to 15% year over year. The reason is simple: the capacity poured in over the last two years is now shipping, Chinese mature-node fabs added a wave of supply, and demand cooled. Parts are flowing.
The problem is that most buyers are still running one playbook across two markets.
On the tight side, don't save the days
Automotive MCUs, AI-edge parts, certain specialty power — when you get a spot quote on these, don't grind the distributor for a week over price. In an allocation market, waiting is itself a cost. While you negotiate, the part gets eaten by someone else or the factory pushes the lead time out another notch. The logic here is: lock the part first, talk price second. Sign what you can, take the NCNR, move.
On the loose side, don't chase
General-purpose analog, standard discretes, mature-node commodity parts — this side is softening, and chasing price now means catching a falling knife. If you have buffer demand here, buy in tranches, push lead times short, ask the distributor for spot, not long-term. In a falling market, whoever signs the long contract first loses.
Dead stock is an opportunity right now, not a burden
This is the point spot traders should grab. In a two-speed market, the value of dead stock is dislocated. That automotive MCU or specialty logic gathering dust in your warehouse — nobody asked about it six months ago. In an allocation market with 40-week factory lead times, it can be hard currency, because the factory can't wait and will pay the spot premium. Flip it around: the general-purpose analog and standard discretes you're holding are trending down. The longer you sit on them, the more you bleed.
So the move right now is to run your inventory list back through one filter — which speed is this part in, today:
- Automotive/industrial MCUs, AI-edge logic, specialty power → the tight lane → list it, work the match actively, take the premium if it's there.
- General-purpose analog, standard discretes, mature-node commodity, NAND → the loose lane → don't sit on it, the trend is down, move it fast.
- The fuzzy middle → check the actual manufacturer lead time and the last four weeks of spot prints, don't go on impression.
Three signals tell you which lane a part is in
One: manufacturer lead time. Still above 26 weeks and not pulling back over the last two months — tight lane. Two: process node. 28nm–40nm automotive, industrial and specialty parts jam more easily; pure mature-node commodity loosens more easily. Three: end market. Feeding AI servers and automotive electronics — tight. Feeding consumer, white goods, general industrial — loose.
The most expensive mistake is reading one "2026 chip shortage" headline and holding every part as if it's tight, waiting for a price that isn't coming. The analog and discrete lane is heading down. Bet the wrong way and your inventory bleeds out slowly.
In a two-speed market, the money isn't in calling "up or down." It's in calling which lane each part is in. In the same warehouse, some parts you should be racing to list and some you should be racing to clear — and right now you have to do both at once.
Whoever re-sorts their stock by speed first sees soonest which parts are cash and which are baggage.