For me personally, I am not factoring FSD into the SP for several more years. Certainly not going to do so just because FSDb is about to get wide-released to everyone who purchased it. Here's why:
1) FSDb just isn't robust enough yet. I was in the first safety-score cohort that got FSDb last year on Oct 11 2021. We were on v10.2 back then. First impressions: wow, the car is doing it's own thing! That feeling cannot be understated. But it was heavily tempered by, wow this thing is incredibly jerky and very uncomfortable. Around the end of the year, Musk suggested (to Lex Fridman maybe) that FSDb would be ready for wide release in a year. Of course he's said this so many times in years past, but this time, I could experience the progress firsthand rather than try to sift through the shills and the naysayers opinions. Between each version, it was difficult to tell if there was meaningful progress being made. But after 4 or 5 releases, it was clear FSDb was definitely improving. But is it improving at a rate fast enough to be ready for the end of the year mass deployment?
2) What will be delivered at the end of 2022 won't be what the masses might expect. What will be released is a feature-incomplete, L2 system, where the driver must keep their hands on the wheel along with cabin camera enforcement of looking forward. Yes, Elon did say during the call last night that you wouldn't need to touch the wheel. IMO he was trying to convey capability of the system. But the system is still going to enforce periodic torque like it has since forever. Because in reality, there will still be lots of situations where the car does something completely wrong and unexpected, and the driver is responsible to prevent it from being unsafe. Also note how Elon dodged the question specifically asking about L4/5 capability. He's basically saying that by EoY, FSDb will be a decent L2 system. We're almost at the EoY, and based on my observations since 10.2, we're going to end up with an L2 system where paying attention is very, very mandatory. The car still exhibits unnatural jerkiness on some very basic turns, and it also has no idea about your local driving culture (speed limits, aggressive passing, etc).
I also said it was feature-incomplete because there's still a fairly long list of things the car can't do, like U-turns, 3-pt turns, reverse maneuvers, interpreting more complex signage, school zones, school buses, etc. These aren't things the car does poorly; it flat out can't handle them yet. therefore, feature-incomplete.
3) Maybe the EoY wide release will allow Tesla to recognize revenue from FSD, and that might make for a nice earnings report, but I think there's a fundamental disconnect between the capability of FSDb and the asking price, currently $15,000 as of this writing. I'm not going to dictate what is a good value for FSDb, but based on how it performs today, my intuition tells me many people would be disappointed with FSDb performance after paying that amount. Already there are lots of vocal people on TMC polluting every FSD thread about how they were duped, and FSD is a scam. This problem could escalate more once FSDb is wide-released. Once FSDb is available to all, I would advise anyone thinking of purchasing it to do a 1-month subscription first. You might love it; great! Or you might hate it, at which point I just saved you $14801 that you can put into TSLA. Assuming informed buyers and current FSD pricing, I don't see take rates being high enough to be material to the SP. I feel that FSDb still needs a year or two to become polished enough to justify TODAY's $15,000 pricing. At that point, it would be feature complete and a very reliable and comfortable L2 system.
I love FSDb, and I love being part of the testing process. I send snapshots and send Tesla detailed emails describing problems and suggesting feature improvements, even beyond FSD. But I cringe a bit when I hear the bulls get overenthusiastic about FSD and robotaxi potential. It will need more time. And one more observation: for any robotaxi that intends to operate without a driver, all doors and trunk lids need an ability to auto-close. Imagine passengers getting out and leaving the doors open. Would the car just drive off like that? Not safely. And in the S3XY lineup, only the Model X has auto-closing doors. So unless 3/Y get autodoor retrofits, don't count on them being the driverless robotaxi fleet of the future.