Tesla Q1 Earnings: More reality, still much pretend

Stocks A to Z
1

This earnings call was a mix of sober reality and unconvincing, pretend product milestones.

My personal favorite:

  • A maybe accidental acknowledgement that AI4 ( = HW4, their latest hardware for autonomous driving) has not yet achieved autonomy: “we will be able to achieve unsupervised self-driving with AI4” - “will be able” is future tense, meaning they haven’t achieved it yet.
  • But didn’t they roll out autonomy in Texas the last few months?

It was a solid display of goalpost moving.

For people that have faith in the stock performance, congratulations.

But as far as autonomous vehicle products, it was just more broken dreams.

For Full Self-Driving and Robotaxi, version 14.3 was a major architectural update, and we have a whole pipeline of major improvements to Full Self-Driving that we believe will lead to unsupervised Full Self-Driving being available anywhere in the world that it is legal to do so. Then there’s a version 15, hopefully by the end of this year, but certainly by early next year. That will be a complete overhaul of the software architecture, and will run on AI5. At that point, we’re really just increasing the safety level of FSD above human safety level even more, meaning I think even within version 14, we’re significantly safer than human, but V15 will take that to another level. We’ve expanded Robotaxi to Dallas and Houston, using the same software source in the Bay Area.

Sober Reality:

  • Major improvements still needed to achieve unsupervised (one could say they caveat it with “anywhere in the world,” but we all know it currently does not work anywhere with the required safety level).
  • The FSD version V15, the next major software update, will require new hardware, AI5 (sorry current hardware owners with AI4 = HW4). This bodes poorly for HW4 achieving autonomy.

Pretend Product:

  • Unsupervised FSD has better than human safety (I would hope supervised is better than human given that it is human plus machine).

The limiting factor for expansion is really rigorous validation, making sure things are completely safe. We don’t want to have a single accidental injury with the expansion of Robotaxi, and we have, to the credit of the team, not had a single one to date.

Sober Reality:

  • They say that safety is holding back their autonomous driving product, which is true. Previously they claimed it was held back for regulatory reasons. Another claim goes “poof” into the ether.

Well, we certainly hope to have unsupervised FSD or Robotaxi operating in, I don’t know, a dozen or so states by the end of this year. Initially, we’re taking a very cautious approach to the rollout here. We haven’t had any injuries and certainly no fatalities to date with the unsupervised FSD and Robotaxi expansion. We want to keep it that way. I think probably unsupervised FSD or Robotaxi revenue will not be super material this year, but I do think it’ll be material probably in a significant way next year.

Sober Reality:

  • No, there will not be material autonomy revenue this year. Not a difficult forecast: same for next year and I doubt 2028.

Pretend Product:

  • No, there will be no meaningful unsupervised in 12 states in 2026. Hope is an important qualifier - thank you general counsel.

[Regarding consumer-owned autonomous vehicles] I’m just guessing here, but probably in the fourth quarter. It’s difficult to release this to everyone, everywhere, all at once because we do want to make sure that there are not unique situations in a city that particularly complex intersection. Actually, they tend to be places where people get into accidents a lot. Because perhaps there’s, like I said, an unsafe intersection or bad road markings or a lot of weather challenges. I think we would release unsupervised gradually to the customer fleet as we feel like a particular geography is confirmed to be safe.

Sober Reality:

  • Yes, just like Waymo, any deployment will be gradual and incremental, one geofence at a time (there will be no flip of the switch that wakes up the fleet nationwide).

Pretend Product:

  • No, there will be no meaningful customer-owned AVs this year, or 2027. This is far away with Tesla accepting liability for consumer AVs. Guessing is the legal qualifier.

Unfortunately, Hardware 3, I wish it were otherwise, but Hardware 3 simply does not have the capability to achieve unsupervised FSD. … For customers that have bought FSD, what we’re offering is essentially a discounted trade-in for cars that have AI Hardware 4. We’ll also be offering the ability to upgrade the car to replace the computer, and you also need to replace the cameras, unfortunately, to go to Hardware 4.

Sober Reality:

  • No autonomy for HW3 and older. HW4 owners: watch carefully and brace yourself.

do expect that AI 5 will go into Optimus and into the data center, because it’s looking like we’ll be able to achieve unsupervised self-driving with AI 4 that is far greater than human safety levels. Which means it’s certainly not immediately needed in the car.

At some point, I think it will make sense for us to switch to AI 5 in the car, but there’s not a pressing issue to do so. At some point the AI 4 hardware is going to get so old that it’s like, okay, the only reason they’re keeping the factory open is for AI 4. We are planning an AI 4 upgrade to use a newer generation RAM. It’ll go from 16 gigabytes to, I think, 32 gigabytes per SoC, so a total of 64 gigabytes. Probably a 10% increase in compute in trillions of operations per second and in memory bandwidth. That’s AI 4.1 or AI 4+ probably goes into production middle of next year, I think. Samsung’s doing the modifications for us, so it sort of depends on when they’re able to finish those modifications and bring it to production.

Sober Reality:

  • A maybe accidental acknowledgement that AI4 = HW4 (their latest hardware for autonomous driving) has not yet achieved autonomy: “we will be able to achieve unsupervised self-driving with AI4” - “will be able” is future tense, meaning they haven’t achieved it yet.
  • But didn’t they roll out autonomy in Texas the last few months?
  • Oh, but they are going to double the memory of AI4 - but then it’s not the same hardware, is it?

Also scaling up the amount of QA fleet that we have across the entire U.S. to accelerate our safety validation

Sober Reality:
They actually do need to measure safety in a standardized, controlled way with company drivers. That’ll take awhile as 20 million miles are needed to statistically validate human or better safety.

I think 14.3 is the last piece of the puzzle for unsupervised FSD. Now, the question is degrees of safety and convenience, I suppose. We have a lot of known improvements, major architectural improvements that we know would improve the probability of safety significantly. I think it’s not going to make sense for us to deploy unsupervised FSD or Robotaxi at large scale when we know that there are major architectural improvements to the software that can improve safety. I think we’re going to want to finish writing that software, validate it, and release it before going to large-scale unsupervised FSD, depending on what large scale means.

We are, of course, as I mentioned earlier, doing unsupervised FSD in three cities, and we’ll expand on it to, like I said, probably a dozen states or more later this year. Depends on what your definition of large scale is.

Sober Reality:

  • V14.3 is the last piece of the puzzle, except for the safety part.
  • Well, safety is important.

Pretend Product:

  • No, there’s no meaningful deployment of unsupervised Teslas, just small demos that serve as marketing stunts.
  • Legal escape: depends on your definition of large.
  • How about 50% of the US?
3 Likes
2

That’s odd, because I have read several reports that there has been at least one with injuries.

As with every previous Tesla crash in the database, all five new incident narratives are fully redacted as “confidential business information.” Tesla remains [the only ADS operator to systematically hide crash details] from the public through NHTSA’s confidentiality provisions. Waymo, Zoox, and every other company in the database provide full narrative descriptions of their incidents.

Tesla quietly upgraded a July crash to include hospitalization

Buried in the updated data is a revised report for a July 2025 crash (Report ID 13781-11375) that Tesla originally filed as “property damage only.” In December 2025, Tesla submitted a third version of that report upgrading the injury severity to “Minor W/ Hospitalization.”

Tesla 'Robotaxi' adds 5 more crashes in Austin in a month — 4x worse than humans | Electrek.

Weird, that.

3 Likes
3

I heard it reported that the person went to the hospital months after the accident. The cynic in me suspects they went after a lawyer told them to go. In any event, Tesla is not counting that and perhaps it is reasonable.

1 Like
4

I can find no verification of that, and my Google-fui is pretty good.

I did turn up a report of a second accident “with injuries”.

As of April 2026, Tesla’s robotaxi fleet in Austin has reported 13 incidents over seven months (since June 2025), with the majority involving only property damage and no injuries. Reports indicated at least two instances of minor injuries requiring, or resulting from, hospital attention

So either both these reports are wrong, or Musk was wrong when he said “no injuries” in his conference call.

1 Like
5

The history of these calls and statements are full of inconsistencies and goal post moves.

It was regulations holding them back (untrue), but now it is safety (true).

It was geofence is bad and they will have a nationwide rollout, but now it’s very much geofenced and a gradual, city-by-city rollout.

But, $0.5 billion quarterly net income for a $1,400 billion market cap and no material new product revenue in sight.

We have threads on the bear case.

What’s the bull case?

2 Likes
6

That they’re one big software advance from a massive revenue infusion.

The bull case on this side of their business is that Tesla will be able to develop autonomous software that allows their vehicles to get to Level 5 or a very robust Level 4 in a way that allows the existing customer-owned fleet to go self-driving. I think an implicit assumption of the bull case is that there will be “light” requirements for a Tesla network to support all those cars driving themselves (ie. a relatively good ratio of remote staff per autonomous car, whether serving the owner or part of a Tesla network).

This would accomplish two things. It would drive revenue to the Robotaxi side, since Tesla would be able to roll out a much more cost-effective Robotaxi to a host of cities far faster and cheaper than competitors. And it would supercharge revenues and profits on the automotive side, since Tesla would be offering the only AV car available for private purchase.

Of course, both of those scenarios depend on Tesla having an AV software product that is vastly better than what they have now. But that’s the bull case - believing that Tesla will be able to do it fairly soon, and the several million HW4 cars will explode in value that Tesla can capture.

3 Likes
7

Consumer-owned AVs with L4 autonomy are far away for Tesla, for any meaningful driving scenario and any meaningful revenue.

Difficult to see this happening in 2027 given where they are with corporate-owned robotaxi/AVs.

At least for in-house developed AI.

They could try to acquire/partner the technology, like what we see discussed with SpaceX/Xai and Cursor for AI and same with Intel for chips.

8

Perhaps. But you asked what the bull case was.

That’s the bull case. Tesla makes a significant enough advance in the capabilities of the AV driver software that: i) they can rapidly expand the corporate-owned robotaxi/AVs to a lot of new service areas; and ii) they can open up full autonomy to customer-owned cars, at least in the areas that they operate the Robotaxi service. Get the software good enough, and the corresponding ratio of remote drivers to “cars in AV service” low enough, and they could make that happen.

The bull case assumes they’ll be able to do that in the near future. The extreme bull case is that the jump in capability gets them close to Level 5, so that they’re flipping that switch all over the country. I think the less extreme version is that there still needs to be a support infrastructure of remote operators and people to physically help some number of cars, but that it’s cheap enough (relative to the number of cars) that Tesla can accommodate customer-owned vehicles alongside the corporate fleet. The weak version is that the software gets good enough that it enables them to expand the corporate fleet, but still requires too much support personnel that it’s cost-prohibitive to let the individual customers use that support without paying for it.

2 Likes
9

Yes.

And I’m just saying meaningful Tesla L4 autonomy and related revenue seems quite unlikely this year and at least 2027.

So I don’t see that bull case supported by their in-house AV capability and reliability.

To get something meaningful in 2027, they better show a higher rate of progress than what we’ve seen the last 12 months and the last decade.

10

You offered the fantasy case. The bull case is that SpaceX IPOs, Musk, having effective control of both companies combines them, and slaps a Starlink antenna on each one so Intel (via SpaceX) can design a video game which controls all the cars from up in the sky, and we skip all this nonsense about Lidar and cameras and “stop signs” and “red lights” and “school zones” and such.

I won’t be surprised if he pitches this on the next conference call, and I will be even less surprised in the fan bois don’t swallow it whole.

1 Like
11

I sure feel very similar to this. It sounds to me like they can still do these minimal test rollouts in other places just for whatever nominal benefit they will gain. They seem to have limited value and mostly window dressing for appearances of something more than is real.
It seems like Musk has pretty much conceded with talk of AI4+ and/or rel 15 that true scaling is still years away. I have seen some seemingly knowledgeable skeptics suggesting 5 or more years away from a bonafide product. Even the 14.3 and 14.3.1 releases appear to be very limited for the first week+ only entries in Community tracker were for Illinois. Now Texas and VA have entries. I would be tempted to say it appears nil improvement but not near enough data yet to actually warrant that.

And the HW3 debacle may have really bad implications. The JPMorgan analyst made following comments based on conf call discussion:

…with regard to FSD and robo-taxi – management backed away from certain maximalist claims about the sufficiency of some of its installed hardware to achieve true full self-driving and hinted at a slower rollout of autonomy-enabling software for the vehicles that are already capable of it from Tesla’s perspective. While we have been noted skeptics here for some time, including given the potential for insufficient sensor redundancy, we expect the sentiment of more bullish investors to be negatively impacted by the more cautious tone taken with regard to the pace of rollout of true full self-driving for customer-owned cars and also the acknowledgement that some millions of Tesla vehicles (those equipped with its Hardware 3 package) – which were advertised as capable of full self-driving from a hardware perspective – are, in fact, not capable and will require extensive rework to make them capable. We wonder also about the potential civil liability as Tesla has often charged hefty premiums for these vehicles, and there are a number of lawsuits already along these lines.

(Note: Bolding by JPM analyst - not me)

2 Likes
12

Nothing illegal about this but does illustrate one of several issues related to why earnings “quality” is being noted by many analysts in recent qtr.

Battery storage undoubtedly would have been bought elsewhere if not Tesla. Certainly true about vehicles too but this outright “unloading” of unsellable Cybertrucks is a real black eye for TSLA

Tesla said it made about $430 million last year by selling Megapack products — a lithium-ion battery system — to xAI, Musk’s artificial intelligence company, and the maker of Grok. It added that it made about $143 million, mostly from selling vehicles to SpaceX, the billionaire’s rocket company, last year.

2 Likes