Should I keep waiting for the STM32V8 or move on?

The short answer is you cannot design around the V8 with confidence yet. I had a board in early prototyping stages for a year, and the parts were not in. The situation has definitely become complicated because of ST's internal fabrication challenges. They are struggling to maintain capacity across thier entire portfolio, and it seems like the V8 has gotten caught in a really long supply chain bottleneck.

Most sources are pointing to a delay that pushes availability well into early 2027. That timeline is not a guess from a marketing team; it's related to complex processes like respinning silicon for different fabrication techniques. Microcontroller manufacturers like ST don't just magically create more chip space, expecially with the low margins they operate on. If you absolutely must lock in a design now to meet your current prototyping window, you need to pick a known, established family. It’s always safer to go with a proven part from a current run, even if it isn't the absolute newest feature set, because you can actually acquire those chips at scale. Waiting for a chip that hasn't made it past the pre-production test cycle is basially gambling with your entire schedule. I strongly recommend designing to an established platform first and only switching to the V8 if the specific benefits it offers justify the two-plus-year wait. til you see verified stock arriving at distributors, the risk of re-spinning boards later is too high for a project deadline.

I agree the delay news is rough. When we were migrating a high-speed sensor interface last year, we had a specific timing issue on the peripheral bus. We had to pull the clock speed back from 200MHz to 180MHz just to meet signal integrity requirements on the layout, which chewed up significant design time. If you are truly anchored to the V8's unique instruction set, you might consider a tiered chip strategy rather than waiting for the final silicon. Design the core logic around a known, available Cortex-M4, but keep the V8 peripheral definitions in place so the code only shifts when stock clears. Have you looked into the bridge chips ST is planning for those transitions?

The whole idea about ST struggling with internal fabrication capacity definitely seems correct. I saw a similar mess when working with a different high-speed controller that was supposed to be readily available, but the silicon revision shifted mid-run. That change completly broke my initial power delivery layout assumptions. You really don't want to bet your timeline on a chip that is so dependent on specific manufacturing runs that the specs could change unexpectedly.

I went this route a few years back. We were trying to switch to a newer version of an older chip just to save power. We figured the new silicon whould be a drop-in replacement, but it wasn't. The pins were the same, but the power consumption profile was totally different. That meant my entire decoupling capacitor layout had to be redesigned from scratch. The whole power delivery net took weeks to fix after the chips finally arrived. If you plan on making that jump, make sure you pull the full datasheet, not just the marketing summary, before you lock in the layout.