Will Tesla/Elmo fulfill their projections/promises?
I reserve the right to modify resolution criteria based on feedback from comments up to Apr 27th.
All production questions only include items produced to be sold on the market, not prototypes or similar.
Robotaxis count only if running:
without a human driver
continuously without any definite end date or larger breaks in operation (no one-off stunts)
At least 5 cars per state
I won't trade on this market.
Specifically:
"We expect volume production of both Cybercab and the Tesla Semi this year,"
Musk signals Optimus optimism. He says he "remains convinced" that Optimus, the humanoid AI bot, is Tesla's most important product, and predicts that the company will begin production later this year.
[Roadster] "May be able to debut that in a month or so," Musk now says. "I think it will be one of the most exciting product unveils ever."
Update 2026-04-24 (PST): Definition of "produced": A unit counts as produced if it is ready to be sold to customers, meaning both hardware and software readiness are required, but regulatory approval is not required.
If Tesla deploys units for use sufficiently similar to their intended customer use (e.g. Cybercabs as robotaxis), that counts as YES
If there is substantial doubt that units have such capacity (e.g. no convincing demonstration from Tesla), the market will tend to resolve NO
The general spirit for Cybercabs and Semis is to check if volume production has started; if it is unclear whether they have 100 usable units I will tend to resolve NO.
Similarly for Optimus - if it is unclear whether they have 10 usable units, it is unlikely that "production" has begun in earnest and I will tend to resolve NO.
1,000Google AI mode:
What is your estimate for how many Optimus v3 Tesla has 1. already and 2. will produce in 2026 excluding those produced before 2026?
1. Already Produced as of Q1 2026
Estimate: Very few (Dozen-level testing units)
What is your estimate for how manyCybercabs Tesla has 1. already and 2. will produce in 2026 excluding those produced before 2026?
1. Cybercabs Already Produced (As of April 2026)
Estimate: Approximately 60–100+ units.
This probably isn't enough for resolution yet so questions are ok as in not invalid as in ready to resolve immediately question is created. Two questions here though
1. Does production in 2026 before question creation count towars the total?
2. If yes to 1 and evidence emerges that the number was higher than the threshold before claim creation does question resolve yes or N/A?
I am more interested in further checking exactly what you mean when you say "produced in 2026" regarding Optimus. Do you do just mean "physical hardware produced"?
i.e. if the physical hardware is "produced" but the AI for them to run well enough to be sold (or otherwise used?) is still in training or some similar stage such that none have been sold nor are they ready to sell (or use them as opposed to training them) then: Are they "produced" and question resolves yes or are they essentially unready and so this is more like 'not really fully produced' and the question resolves no?
While I am at it, I may as well also check re cybercabs:
It also seems possible that Cybercabs will be physically produced but not sold to customers because regulation does not allow steering wheel free car sales. Tesla might produce sufficient for question and is either
1. Using them as robotaxis so not sold but in use, or
2 Perhaps it is more of a remote possibility that in reality isn't going to be a problem but they could be "produced" and not in use but stockpiled for later use as robotaxis or sales when software/regulation/whatever allows.
Do each of these count as "produced"?
Do you want to clarify question whether physical hardware produced is all that is needed or does software and AI and any other intangible needed to make useful a part of what is needed in order to be fully produced?
@ChristopherRandles The spirit is to check if they started "volume production" and I am deliberately setting a somewhat low bar, with the idea being that if they actually start volume production, it will be clearly YES. If there's substantial ambiguity, I will generally tend to resolve NO.
I think a good definition of "produced" is "ready to be sold to customers", this thus includes both hardware and software readiness, but does not include regulatory approval. If those units are actually deployed by Tesla for use sufficiently similar to their intended use by customers (e.g. Cybercabs as robotaxis), it would certainly suffice for a YES. If there's substantial doubt that the units have such capacity (e.g. no demonstration from Tesla), I will tend to resolve NO.
@AIBear Just guessing off the top of my head, it would sound plausible to me to suggest situation at year end could easily be something like
5000 Optimus produced which seems like volume
1000 "in training" at Tesla (at any one time?)
0 sales (as either units or RaaS Robot as a Service subscription dollars revenue)
and it could well be hard to gather data on whether any of the 1000 "in training" at Tesla are strictly training only or whether any have become "deployed" in the sense of the expectation is that Optimus robots will fairly permanently continue doing that useful work job as opposed to it being a training task and then the robots will move on to other training tasks to develop the AI at least until the software is ready for sales or deployments of robots.
Could easily be ambiguous between 0 deployed and so software/AI is not ready so not "produced" and should resolve no or over 10 deployed and should resolve yes. Unfortunately, I am not sure I have much in the way of sensible suggestions to help clarify the question on what you would do in that sort of circumstance. Perhaps only ask if you agree this sort of language distinguishing between 'training' and 'deployment' might be used as a base for what counts? So maybe add to question description something along lines of:
If "in training" is the sort of language used for furthest reached and there is no better information then this might be used as indicative of a likely no to the question of whether ready enough to be considered produced. However with more information the reality of situation might override this. If the language used is deployed or even just test deployment then this might be used as indicative of a likely yes, similarly unless more info reveals that reality of situations reveals that it is more like just 'in training' to help improve the AI or software to help reach point where it is ready for use.
Perhaps worth a mention though above is partly based on this: Suppose there is a "test deployment" heavily discounted or even no sales revenue for a trial period while a corporate customer tries some Optimus robots out to see if they can work well enough to be worth paying the subscriptions for them. Has this reached a stage beyond training such that we think the software is practically ready enough even if not fully perfected sufficiently for large scale sales or is this still at a software not quite there for usage stage? Also is the trial start date the relevant time or is that only if the customer goes on to pay to continue using them after the trial?
Also suppose there was only one such 'test deployment' at year end with 9 or less robots. Does that mean it fails the 10 quantity required or are all 5000 hardware produced robots now ready the instant a single trial deployment starts or the moment a contract is signed/agreed or something else?