Latest / Elon Musk Podcast / Tesla Q2 2025 Financial Results and Q&A Webcast

Transcript

Tesla Q2 2025 Financial Results and Q&A Webcast · Elon Musk Podcast

0:01
Hey everybody, welcome back to the Elon Musk Podcast.
0:05
This is a show where we discuss the critical crossroads, the
0:08
Shape, SpaceX, Tesla X, The Boring Company, and Neurolink.
0:13
I'm your. Host Will Walden.
0:19
Good afternoon, everyone, and welcome to Tesla's second
0:21
quarter 2025 Q&A webcast. My name is Travis Axelrod, Head
0:25
Investor Relations and I'm joined today by Elon Musk,
0:29
Vivatineja and a number of other executives.
0:32
Our Q2 results were announced at about 3:00 PM Central Time in
0:35
the update deck be published at the same link as this webcast.
0:39
During this call, we will discuss our business outlook and
0:42
make forward-looking statements. These comments are based on our
0:45
predictions and expectations as of today.
0:47
Actual events or results could differ materially due to a
0:50
number of risks and uncertainties, including those
0:53
mentioned in our most recent filings with the SEC.
0:56
During the question and answer portion of today's call, please
0:59
limit yourself to one question and one follow up.
1:02
Please use the raise hand button to join the question queue.
1:05
Before we jump to Q&A, Elon has some opening remarks.
1:09
Elon, thanks drives. So we've had a very exciting
1:13
quarter. The we're able to successfully
1:16
launch robo taxis, so providing our first drives with no one in
1:20
privacy with paying customers and Austin and as some may have
1:27
noted, we've already expanded our service area in Austin.
1:32
It's bigger and longer and and it's going to get even bigger
1:37
and longer. We were expecting to really
1:43
greatly increase the motion service area to well in excess
1:49
of what competitors are doing and that's hopefully in a week
1:54
or so, two weeks. Yeah, A.
1:56
Couple of weeks, a couple of weeks or so and and then we're,
2:03
we're getting the regulatory permission to launch in the Bay
2:08
Area, Nevada, Arizona and a number of and Florida, a number
2:13
of other places. So as soon as we get the
2:17
approvals and we improve our safety, then we'll be launching
2:21
the autonomous ride hailing and most of the country.
2:26
And I think we'll probably have autonomous ride hailing in
2:32
probably half, half the population of the US by the end
2:35
of the year. That's, that's at least our goal
2:37
subject to regulatory approvals. I, I, I think we'll technically
2:43
be able to do it. So assuming regulatory
2:46
approvals, it's probably addressing half the population
2:49
of the US by the end of the year.
2:52
But we, we, we are being very cautious.
2:54
We don't want to take any chances.
2:57
And so we're good, we're going to yeah, go, go cautiously.
3:02
But but it's the service areas and the number of vehicles in,
3:08
in operation will increase at a hyper exponential rate.
3:16
So, you know some other notable things.
3:20
Model Y in June became the best selling car in Turkey,
3:23
Netherlands, Switzerland and Austria.
3:25
It is, I believe the best selling car of any kind in the
3:29
world still. And autonomy is a is a big
3:36
factor there. So even without even accepted,
3:41
even without supervised, even with just supervised self self
3:46
driving, it's a huge selling point.
3:50
And it's worth noting that we do not actually yet have approval
3:54
or supervised FSD in in Europe. So our sales in Europe we think
4:03
will improve significantly once we are able to give customers
4:07
the same experience that they have in the US.
4:09
This is my, this is I think a very important point to convey.
4:14
And we've been working with main country regulator, which is the
4:18
Netherlands. And I think we're close to
4:22
getting approval of Netherlands. Then it's got to go to the EU.
4:27
It's quite a, you know, Kafkaesque.
4:30
In fact, Kafka had no idea that's something like the EU
4:33
could exist beyond Kafkaesque challenges with bureaucracy.
4:38
But it will, we will get the approvals and I think we'll get,
4:44
you know some people in Europe will have experience, some of
4:47
them have the US in most of Europe, you know this year
4:53
hopefully at least partially in this quarter.
4:58
And then we also have some regulatory, regulatory
5:02
challenges in China which we're hoping to unblock shortly where
5:06
we because we also cannot provide supervised investee in
5:10
China currently. Are we going to unblock that
5:13
suit And that's also that's another major, it's really is
5:17
the single biggest demand drive. And then within the US, as we
5:24
get confident about safety in different geographic areas, the
5:30
will be, it will loosen up on how much somebody has to be
5:39
laser focused, have their eyes laser focused on the road.
5:42
You know, it's, that's, that's been a common complaint.
5:47
In fact, it it does create an open odd, odd safety issue where
5:52
people will sometimes disengage autopilot, then then do
5:58
something, change the radio or maybe look at the phone drive
6:02
with your knee and then reengage autopilot, which obviously is
6:05
not is less safe than simply keeping autopilot on.
6:10
So anyway, we're get that that that experience will will
6:14
improve in in the next several weeks.
6:18
The the because of our focus on Austin with the one driver's
6:25
seat, the production release of Autopilot is and actually
6:29
several months behind what people experienced on a robotax
6:33
in Austin. So now we have the robotax in
6:37
Austin launched will be provided adding back those elements so
6:41
that there will be a step change improvement in the water pilot
6:47
experience for people outside of Austin.
6:52
This is really, as you can tell, this is very much sort of
6:56
autonomy is, is the story like we need the we need the physical
7:01
product, without which you cannot have autonomy.
7:02
But once you have a physical product, you need the autonomy
7:05
is what amplifies the value to stratospheric levels.
7:14
We also launched the Tesla Diner, which has been a huge
7:16
hit. It actually got worldwide
7:19
attention, which is unusual for a diner.
7:22
Datas don't typically get headline news around Earth, but
7:29
this is a pretty special diner and if you're in the LA area,
7:33
it's worth visiting. It's sort of a shiny beacon of
7:38
hope and but not otherwise sort of bleak urban landscape,
7:42
frankly. So it's really quite a lovely
7:46
experience. Great job by the team there on
7:52
the full stop driving front continue to worry about that we
7:57
have we're continuing to make a significant improvements just
8:01
with the sock rate. So the we're expecting to
8:06
increase the parameter count. Actually, at this point, I think
8:16
we think we can probably 10X that one was 10X the power
8:19
front. Yeah, roughly, roughly 10X the
8:22
parameter count. So this is actually a very
8:27
tricky thing to do because if you as you increase the
8:32
parameter count, you get, you get choked on memory bandwidth.
8:36
But we currently think we can 10X the parameter count from
8:43
what people are currently experiencing.
8:45
That's, that's so not, not just the Forex actually 10X increase
8:49
in private. And yeah, so, so there's still a
8:59
lot of improvement on the existing hardware to, to, to
9:01
happen. Energy is growing really well
9:06
despite headwinds from tariffs and spare supply chain
9:09
challenges. The Megapack is expanding
9:14
capacity quickly and and we have upgrades the mega pack that
9:19
would make it even better and we had record Powerwall deployment
9:23
gain in Q2. So I think batteries are going
9:29
to just going to be a massive thing that the scale of
9:34
batteries, battery demand is. I think not that many people
9:38
appreciate just how gigantic the scale of battery demand is.
9:45
The way you think about it is that the US sustained power
9:49
output for the US grid is around 1 terawatt, but average usage is
9:56
less than half a terawatt. If you add batteries to the mix,
10:01
you can run the power plants 24/7 at full capacity, thus
10:07
doubling, more than doubling the energy output per year of the
10:11
United States just with batteries.
10:15
But that's a good big deal. It's a really big deal, Optimus.
10:25
So we're revolving the Optimus design and really getting
10:31
Optimus to the point where it is a phenomenal design.
10:38
We're an Optimus version to right now sort of 2 1/2.
10:42
Optimus 3 is it is an exquisite design in my opinion and will be
10:49
an incredible as I've said many times before, it predicted it
10:54
will be the biggest product ever.
10:57
It's a very hard problem to solve.
10:59
You have to design every part of it from physics, first
11:03
principles. There's nothing that's off the
11:06
shelf that actually works. So you can design every motor,
11:09
gearbox, power electronics, control electronics, sensors,
11:16
the mechanical elements. We've also got to train
11:21
Optimists to use its lens and its sensors with a neural net.
11:30
But we'll be applying the same techniques we're applied for a
11:34
car, which is essentially A4 wheel robot and Optimist is a
11:39
robot with arms and legs. So the the same principles that
11:43
apply to optimizing AI inference are the car applied to optimates
11:49
because they're both really robots in different forms.
11:53
And and Tesla, it is important to you know that Tesla is by far
11:57
the best in the world at real world AI like like a clear proof
12:04
point for that be if you compare it to Tesla to Weymouth,
12:07
Weymouth's got you know, the car is best doing with God knows how
12:11
many sensors. And yet isn't Google good at AI?
12:15
Yes, but they're not good at real world AI.
12:18
Thus far they have Tesla is actually much better than Google
12:22
by far and any much better than anyone at real world AI.
12:27
And and by far Tesla has the best the best inference
12:32
efficiency. Like the key figure for AI is
12:38
what is the intelligence per GB? And people talk about
12:42
parameters, blah, blah, blah. But I think we'll just talk
12:45
about stop talking about parameters and talk about
12:47
gigabytes because with the parameters, you can have 4 bit
12:50
parameters, 8 bit parameters, 16 bit parameters.
12:53
But the actual constraints in the hardware are how many
12:57
gigabytes of RAM and how many gigabytes per second can you
13:01
transfer from RAM. Therefore, it is not a parameter
13:03
constraint. There's it is a, there's a byte
13:06
constraint and Tesla has the highest intelligence density of
13:13
any AI by far. And I have a lot of insight into
13:18
this with XAIXAI is, you know, Grok is the smartest AI overall,
13:23
but it's a, you know, Grok 4 is a giant beast actually with the
13:30
TB level. And so kind of a important to
13:38
note, Tesla has the best intelligence density.
13:41
Intelligence density will be a very big deal in the future.
13:44
It is now. So with Optimus 3, which is
13:55
really the right design, It's like it doesn't have at this
13:57
point, there's no significant flaws with the Optimus 3 design,
14:05
but it it, we are going to retool a bunch of things.
14:07
So it's it'll probably be prototypes of Optimist 3 end of
14:14
this year and then scale production next year.
14:17
We're going to try to scale Optimist production as fast as
14:21
it's seemingly possible to do. So try to get to 1,000,000 units
14:24
a year as quickly as possible. You think we can get there in
14:34
less than five years? It's nice to sort of I guess
14:40
that that's a reasonable aspiration is 1,000,000 units in
14:43
your five years, it seems like an achievable target.
14:52
In conclusion, so far 2025 has been a very excited year, a lot
15:03
of major milestones. We've made clear demonstrable
15:07
progress in autonomy that's a lot of naysayers said we would
15:11
not achieve. And it's worth noting that the
15:16
we have done what we said we're going to do.
15:19
I mean, we're always on time, but we get it done and our
15:23
naysayers are sitting there with egg on their face.
15:29
So great, great, great progress by this as a team.
15:40
Yeah, I do think if Tesla continues to execute well with
15:45
big autonomy and and like humanoid robot autonomy, it will
15:51
be the most valuable company in the world.
15:54
A lot of execution between here and there doesn't just happen,
15:59
but provided we execute very well, I think Tesla has a shot
16:08
at being the most valuable company in the world.
16:15
Obviously, I'm extremely optimistic about future of the
16:18
company. Best way to predict the future
16:22
is to make it happen, and we're making it happen here with Tesla
16:25
team. So I'd just like to say thanks
16:28
to all of our supporters and I think we're going to tell you an
16:31
incredibly exciting future. Great.
16:34
Thank you very much. Elon and Bedlock has remarks as
16:37
well. Tavis, as Elon mentioned and Q2
16:41
was an interesting quarter in a few respects.
16:44
We started ramping up the production of the new Model Y at
16:47
all our factories. We rolled out our robotaxi
16:51
service in Austin and delivered a car completely autonomously
16:54
from directly from the factory to the customer's home.
16:59
It is a Seminole point. Get to this thing.
17:04
I mean, it took a lot of effort and I really want to thank
17:07
everybody at Tesla to make this happen.
17:10
It wasn't an easy thing to do, but we did it.
17:14
It took time, but we've just begun the next phase for the
17:17
company. The one big bill has a lot of
17:21
changes that would affect our business in the near term.
17:24
The first among those changes is the repeal of the higher EV
17:27
credit of 7500 by the end of this world.
17:31
Given the abrupt change, we have limited supply of vehicles in
17:34
the US this quarter as we have already with the lead times to
17:38
auto parts to build cars. We've rolled out all our planned
17:42
incentives already and we'll start paying them back as we
17:45
start to sell. If you're in the US and looking
17:49
to buy a car, this around now as we may not be able to guarantee
17:53
delivery orders placed in the later part of August and beyond.
17:58
The will also make changes to certain emission standards by
18:01
reducing the amount of penalty to 0.
18:04
This in turn will have an impact on the new sales of regulatory
18:08
credits to other OEMs and in turn will lead to lower
18:11
revenues. When we now plan our business
18:16
around such sales, it will nonetheless impact our total
18:19
revenues. Quite false.
18:22
On our automotive product portfolio, the entire line up is
18:25
updated globally, we're seeing an increase in the number of
18:29
test drives. We started the production of the
18:32
lower cost model as planned in the first half of 25.
18:36
However, given our focus on building and delivering as many
18:39
vehicles as possible in the US before the EV credit expires and
18:44
the additional complexity of ramping a new product, the ramp
18:47
will happen next quarter slower than initially expected.
18:52
One thing which is grossly and appreciated and Elon talked
18:55
about it is that all our vehicles in the lineup are
18:59
capable of a dump. This is by far the biggest
19:03
differentiator between US and the competition.
19:06
Our vehicles top safety standard as is, but with FSD they are and
19:11
will continue to set a new standard for safety within
19:15
vehicle transportation. We published our vehicle safety
19:19
report earlier today and you can see the car on FSD is 10X safer
19:25
than the car not on FSD. We've started seeing an uptick
19:29
in FSD adoption in North America in recent months, which is a
19:33
very promising trend. And just to give you
19:36
perspective, you know since the launch of since we moved to
19:42
version 12 of FSD, we've seen the adoption rates really
19:46
increase. We've started seeing the on the
19:51
automated revenue front despite reduction in regulatory credit
19:56
revenue and the total automotive revenue increased by 19%
20:03
sequentially. Even the total deliveries only
20:05
improved 14%. This was primarily due to an
20:09
improved ASASPS because of the new Model Y.
20:13
This had been improving margin sequentially as well.
20:16
Along with improved mix and higher cost fixed cost
20:20
absorption, despite an increase in cost of tariffs, we started
20:25
seeing the impact of tariffs in our PNL.
20:27
Sequentially, the cost of tariffs increased around 300
20:31
million with approximately 2/3 of that impact in automotive and
20:36
rest in energy. However, given the latency in
20:39
manufacturing and sales, the full impacts will come through
20:43
in the following quarters and so cost will increase in the near
20:46
term. While we are doing our best to
20:49
manage these impacts, we are in an unpredictable environment for
20:52
the tire front. The margins for the energy
20:57
generation and storage businesses improved sequentially
21:00
while deployment reduced primarily due to the ramp of
21:04
power deployments. At higher margins, we were able
21:08
to achieve our highest gross profit for the business yet.
21:12
Note that the overall deployments will continue to
21:15
vary quarter to quarter. Thinking on covered this that
21:19
you know industrial storage will make a difference in this drive
21:22
towards AI and data center growth.
21:25
The energy requirements are increasing at a rapid scale as
21:29
application as AI applications are very energy hungry.
21:34
The quickest path to scale up energy is deploying storage.
21:38
This is something that our customers have started
21:40
realizing. And despite this business having
21:42
the largest impacts from tariffs, we're seeing customers
21:46
willing to accept some of the tariff impacts.
21:50
The big bill has certain adverse impacts even from the energy
21:54
business, most notably on the residential storage business due
21:59
to the early expiration of consumer credits by the end of
22:02
this year. The challenges of the storage
22:06
business therefore remain both from the bill and from the
22:09
tariffs and we're doing our best to try and manage through this,
22:13
but it will we will see shifts in demand and profitability.
22:19
The margins for our service and other businesses improves
22:23
sequentially primarily due to higher profits from
22:26
supercharging and improvement in insurance and service center
22:29
profitability. Operating expenses also grew
22:33
sequentially as we continued our investment in AI projects
22:37
including additional expenses related to employee related
22:41
costs including higher stock based compensation and
22:44
depreciation for AI compute. Our operating expenses
22:48
especially R&D related spend will continue to grow.
22:53
We believe even in the current environment it is the right
22:56
strategy to keep making investments in these areas to
23:00
position us for the long term. Other income grew sequentially
23:05
primarily from the mark to market adjustment on Bitcoin
23:09
holdings which was 284,000,000 gain in Q2 while being 125
23:15
million loss in Q1. Just wanted to remind people
23:17
that this would keep keep creating volatility based on the
23:22
Bitcoin price while other while operating cash was increased
23:27
sequentially, so did our CapEx resulting in 146 million of free
23:33
cash flow. We continue to make investments
23:36
in various aspects of manufacturing like cybercap,
23:40
semi mines and other manufacturing spend and the
23:42
expansion of our AI initiatives. Our latest expect expectation
23:47
for the year in terms of CapEx is in excess of 9 billion.
23:52
To summarize, we have near term challenges in our business due
23:56
to the negative impacts of the bill and tariffs.
23:59
However, the investments that we have made for AI, robotics and
24:04
our lead in energy sets us up for a bright future.
24:08
I would like to thank the whole Tesla team, our customers, our
24:12
investors and supporters for their continued belief in US.
24:17
Great. Thank you very much.
24:19
So now we're going to move on to say.com questions.
24:22
The first question is, can you give us some insight how
24:24
robotaxis have been performing so far and what rate you expect
24:28
to expand in terms of vehicles, geofence, cities and?
24:31
Supervisors. Yeah, Robotaxi has been doing
24:36
great so far in Austin. Customers really love the
24:38
experience. Like super small, very safe and
24:43
like just a great experience overall.
24:45
And we already did the first day of expansion in Austin and we
24:49
continue to expand in Austin to probably more than 10X our
24:53
current operating region. We're also testing in a lot of
24:56
other cities. As I mentioned, the next thing
25:00
to expand would be in the San Francisco Bay Area.
25:03
We are working with the government to get approval here.
25:06
In the meanwhile, launch the service with the person in the
25:09
driver's seat just to expedite. And while we wait for regulatory
25:13
approval, we're also testing a lot of other cities in the US
25:18
including Florida, Nevada, etcetera.
25:24
Great. Thank you very much, Ashok.
25:26
The next question is what are the key technical and regulatory
25:29
hurdles still remaining for unsupervised FSD to be available
25:33
for personal? Use can.
25:34
You provide a timeline. First, let me getting there.
25:46
I think it's it'll be available for in supervised personal use
25:51
by the end of this year in certain geographies.
25:54
We're just being very careful about it, so.
25:58
This is not something which we want to rush, no, we want to
26:00
make sure that everything is safe before we met it.
26:04
Hopefully available broccoli. Yeah.
26:06
We're just being extremely paranoid, Yes.
26:08
But I'd be, I'm confident that by the end of this year, within
26:14
a number of cities in the US, we'll it'll be available to end
26:19
users like. Yeah.
26:24
Yeah. And for what I suppose is the
26:25
same a hardware in the Austin robot XD vehicles as some
26:30
customer vehicles and they deliver a car autonomously from
26:33
the factory to a customer this quarter, Yes and every test and
26:36
manufacturer in the US and in Europe autonomously drives
26:40
itself from the end of time to the.
26:43
Loading. Docks and it, so it's just the
26:45
software. Beta, yeah, I think we'll be,
26:49
we'll end up delivering cars in the great Watson area and the
26:54
Bay Area by default from the factory by the end of this year.
26:57
Like a car will deliver itself to your, to where you are,
27:01
unless you say you don't want that.
27:04
It'll be super cool, yeah. Great.
27:07
Thank you guys. The next question is what
27:09
specific factory tasks is Optimus currently performing and
27:13
what is the expected timeline for scaling production to enable
27:16
external sales? How does Tesla envision Optimus
27:19
contributing to revenue the next 2-3 years?
27:24
Yeah. So the Optimus 3 design, as I
27:29
mentioned earlier is I think finally the right design, the
27:32
way further optimizations, but there are enough fundamental
27:34
changes to that are needed or the Optimus 3 design, it has all
27:38
the degrees of freedom that you really want or need.
27:43
So it's, you know, we'll have prototypes of that and I don't
27:48
know, three or three months and it's subtly in production.
27:53
We'll certainly start production on that in the beginning of next
27:56
year. The production ramp is simple.
27:59
It's always simple to predict the ES curve of your production
28:02
ramp when something has got an entire, when everything is new
28:08
because the rate of productions will move as fast as the least
28:12
lucky and least confident element of the entire supply
28:16
chain as well as in internal processes.
28:19
So the more new stuff that is in a product, the slower the ramp
28:24
could be because of unexpected supply chain interruptions or
28:29
mistakes made internally especially easy to predict sort
28:36
of the the end of the of the S curve or late in the S curve
28:40
than the beginning of the S curve and the beginning of the S
28:43
curve of the production route is in any case not really material
28:47
for revenue purposes. The beginning of the S curve
28:52
key, if you're usually not usually yours negative gross
28:56
margin and you're debugging a lot of issues.
29:00
So that's why I like it's, it's, I feel like fairly confident
29:07
predicting things or at least medium confident in predicting
29:10
where we are in five years. But it's hard to predict where
29:14
we are in a year or two years. So that's why I think five
29:21
years, I think we could be at the, let me look at this, I'd be
29:28
surprised if at the end of five years, you know, 60 months from
29:31
now, if we are not roughly making 100,000 Optimus robots in
29:37
month, in 60 months, I would be shocked.
29:44
All righty. Thank you very much.
29:46
Next question is, can you provide an update on the
29:48
development and production timeline for Tesla's more
29:50
affordable models? How will these models balance
29:53
cost reduction in profitability and what in fact do you expect
29:56
on demand in the current economy climate?
29:58
Well, I, I think Bailout did a good job of answering this
30:01
question in his opening remarks. As we said, we started
30:03
production in June and we're ripping quality bills and things
30:06
around the corner. And given the that we started in
30:10
North America and our goal is to maximize production with an IRA
30:15
for pending Q3, we're going to keep pushing hard on our current
30:18
models to avoid complexity. Unfortunately that rolls away.
30:22
We'll be running with the more affordable models available for
30:25
everyone in Q4. And you know the goal of those
30:28
products was not to negatively impact revenue or or gross
30:31
margin, but just to make a car that everyone loves and wants at
30:35
a more affordable price. Great.
30:39
Thank you, Lars. The next question is, can you
30:43
talk about the benefits of Tesla investing in XAI?
30:46
This is not the forum to discuss this topic.
30:49
I mean like if, if there is something which we need to
30:53
discuss, we'll discuss it separately.
30:57
I think obviously you know we're a publicly traded company.
31:00
Shareholders are welcome to put forward any shareholder
31:03
proposals that they'd like. I've recently encouraged that
31:07
and and then have shareholders vote and we'll act in accordance
31:12
with the shareholder wishes. Great.
31:15
Thank you very much. The next question is, can you
31:18
tell us a little bit more about what goes on in the Tesla design
31:21
studio do? You want me to take that one?
31:28
We kind of generally say that what happens in the studio stays
31:32
in the studio and that, you know, earnings calls are not the
31:36
place to disclose new product stuff.
31:38
But we're, you know, we're working to make sure that we
31:41
have an exciting future for for our Tesla and our and the
31:44
product lineup. Yeah, there's a lot of exciting
31:48
things happening in the design studio.
31:50
It's not like static. And really what's going to
31:54
happen over the next several years is a fundamental
31:57
transformation of the company from a pre autonomy wall to a
32:02
post autonomy. And I'm working on a new master
32:10
plan to articulate that as a team.
32:14
And there will be some teething pains as you transition from
32:22
free autonomy to the post autonomy world.
32:24
But I think the future version for Tesla is incredibly exciting
32:30
and will profoundly change the world in a good way.
32:37
So you sound like sort of time for whatever, but I, I think if
32:40
what let's just say if we execute on that plan
32:43
effectively, which is you have to actually do that, Tesla will
32:47
be the most valuable company in the world by far.
32:51
Great. Thank you.
32:53
The next question is actually a duplicate on unsupervised set of
32:56
Steve customer vehicles. We'll escape that.
32:59
And after that is are there any news for hardware 3 users
33:03
getting retrofits or upgrades? Will they get Hardware 4 or some
33:06
future version of hardware? 5.
33:11
I mean, what we want to do is we want to get unsupervised done on
33:15
hardware for first. Once it's done, then we'll go
33:18
back and look at what we need to do with the hardware 3 cars.
33:22
I mean the like I said, the focus is first to get
33:24
unsupervised out and then we'll go back and see what more work
33:29
you need to do. Great.
33:33
The next question is can you give an update on Dojo and could
33:37
xai be a customer for Dojo? Dojo 2.
33:47
We expect to have Dojo 2 operating at scale sometime next
33:51
year with scale being somewhere around 100 KH 100 equivalents
34:00
and and then AI 5, which is really spectacular to use those
34:10
words. Likely spectacular to the AI5
34:14
chip will be hopefully be in buying production around the end
34:19
of next year. But that has a lot of potential.
34:25
I think, you know, thinking about Dojo 3D and the AI6
34:30
inverse chip, it seems like intuitively we want to try to
34:36
find convergence there where it's basically the same chip
34:44
that is used. What we use say two of them in a
34:48
car or optimist and and maybe a larger number on a on a board, a
34:55
kind of 512 on a board or something like that.
34:58
If you want high bandwidth communication between the chips,
35:01
how to force serving during doing it?
35:07
Burn serving. That sort of seems like
35:10
intuitively the sensible way to go.
35:15
Great. The next set of questions have
35:19
all actually been covered. So we'll, we'll end with how
35:23
will the BBB elimination of tax credits for solar projects
35:27
affect your sales pipeline for Megapack?
35:30
Yeah, our, our sales pipeline is, is quite diversifying across
35:33
customers and market segments. So we aren't heavily weighted in
35:36
Megapack projects that are paired with solar.
35:39
And as we talked about in the opening remarks, we're seeing
35:42
storage quickly being recognized for its ability to unlock grid
35:44
efficiency and how quickly it can be deployed to help the
35:47
grid. Additionally, although the
35:50
recent bill was not favorable towards solar, we, we believe
35:52
solar projects will still get built because the energy is
35:55
necessary. The projects are well developed
35:57
and they're ready for execution and there's really no
36:01
alternatives in the near term given gas turbine lead time and
36:04
pricing. We also can continue to see
36:07
growth in the data center segment and in stand alone
36:09
storage projects providing capacity to the grid in several
36:12
markets across the US Overall, we're forecasting a very strong
36:17
second-half of the year as we increase deployments.
36:20
And lastly, we we continue to invest heavily in US
36:23
manufacturing to mitigate policy and tariff impacts, expecting
36:26
our first LFP cell manufacturing facility to be online by the end
36:30
of the year and launching our third mega factory near Houston
36:33
and 2026. Great.
36:36
Thank you, Mike. We will now be moving to analyst
36:39
questions. The first question comes from
36:42
Emmanuel Rosner at Wolf Research.
36:45
Emmanuel, please feel free to unmute yourself whenever you're
36:47
ready. Great.
36:51
Thank you so much. Can you hear me?
36:53
Yeah, thanks. So, Elan, are you able to share
36:58
any KPIs with us in terms of the robo taxi business, how many
37:03
vehicles are you operating, miles driven autonomously or the
37:08
number of safety critical intervention?
37:10
Just curious, you know how to roll out, is it generally is
37:14
going in and any sort of like targets that you could share
37:17
more broadly? Sure, Jonah.
37:21
Yeah, we have. You know, more than 7000 miles
37:25
operating in Austin area. It's you know, just because
37:28
services new via a handful of vehicles right now, but then
37:33
we're trying to expand the service in terms of both the
37:36
area and also the number of vehicles both in Austin and
37:39
other locations. So far, you know, the there's
37:43
like no notable critical influence there.
37:46
You know, sometimes we have our own restrictions as to for
37:49
example we resting towers speedy limit to 40 mph and the vehicle
37:55
wants to go on like higher speed or we can stop the vehicle, but
37:57
those are a lot of convenient as opposed to safe and critical
38:01
nature. So for the service that has been
38:03
really well received and we continue to expand on it.
38:09
Great. And that, yeah, longer term from
38:14
an economics point of view, longer term, you've previously
38:16
talked about working to drive down the cost per mile on the
38:20
robo taxis maybe towards 30 or $0.40 per mile over time.
38:25
Now that your service is live, how should we think about the
38:28
main milestones to getting there?
38:32
Yeah. Well, the the cyber Cab is which
38:36
is really optimized for autonomy that that I think has got to
38:44
probably sub 30 cent per mile potential over time, maybe 25.
38:52
You know, it's, it's really like if you design A car from scratch
38:57
to, to be a cost optimized robotic taxi, like cyber cab,
39:07
you know, you like, for example, we're not trying to make it, you
39:11
know, corner like a like incredibly well, like a model 3
39:15
would or Model S or even model Y like it's got model, you know,
39:20
model all of our, all of our cars that are driven by people
39:24
are super fun to drive. They've got incredible
39:25
acceleration, you know, the incredible cornering capability.
39:31
But we're confident that very few people in a cybercap want to
39:36
be huddling around. So, you know, so we were, we
39:41
reduced the top end speed, which means we can use more efficient
39:45
tires. We don't need as much
39:48
acceleration. We don't need as much, you know,
39:53
big brakes to sort of we want stopping distance, but we're not
39:56
expecting it to be heavily used. It's, it's a gentle ride.
39:59
Essentially, if you design it for a gentle ride and then you,
40:07
you have a. A much more optimized design
40:10
point so that that's why it seems probably we could achieve
40:14
that, especially if optimist is, you know, so it's serving
40:19
cleaning up the car and make doing maintenance and stuff and
40:26
it's, you know, doing automatic charging.
40:31
So I think it's going to the actual cost per mile of Cyber
40:36
Cab will be very low. The cost from Wells, our
40:41
existing fleet will be higher, but still very competitive.
40:46
So yeah, maybe some number of $0.50 in viscosity, Yeah.
40:55
So this really Tails Rd. Taxi fleet will go from tiny to
41:01
gigantic in terms of operations in pretty short period of time
41:08
like my guess is as it has a material impact on our
41:12
financials around the end of next year.
41:17
Great. Thank you very much.
41:19
The next question comes from Adam at Morgan Stanley.
41:23
Adam, please unmute yourself. Great.
41:26
Hello, everybody. So Elon, as Tesla moves into
41:30
this next phase of physical AI, autonomous humanoids, robotaxis,
41:35
etcetera, world changing, civilizational species changing
41:41
technology with dual purpose, are you comfortable moving Tesla
41:45
in this direction while only having a 13% stake in the
41:49
economy, sorry in the company, is that sustainable or does he
41:55
still insist that something needs to happen given, you know,
41:58
your current lack of control on the types of technologies you're
42:01
getting into? Yeah, that that, that is a a
42:06
major concern for me, as I've mentioned in the past and I hope
42:09
that is addressed at the upcoming shareholders meeting.
42:13
But yeah, it is a big deal. Like, you know, I want to find
42:17
that I've got like so little control that I can easily be
42:20
ousted by activist shareholders after having both, you know,
42:23
these these army of humanoid robots.
42:28
I think that, as I've mentioned before, I think my control of a
42:32
Tesla should be enough to ensure that it goes in a good
42:35
direction, but not so much control that I can't be thrown
42:38
out if I go crazy. OK, Elon, and you're not going
42:44
to go crazy. We, we, we, we trust you.
42:48
You can stay a little crazy. A little crazy is OK, Elon,
42:52
though we understand the the board of directors of a major
42:58
U.S. investment bank recently toured Optimist production.
43:01
That's, I don't know if you want to confirm that or not.
43:04
It's just what we've heard. That's cool.
43:07
But when do you think others will be able to get a first hand
43:10
view of Optimists like that? And is the second-half of this
43:15
year too soon to have an AI day? Because it seems like everybody
43:19
else in the world's doing it. And this talent war is getting
43:22
freaking crazy. And I, I know, I know you've
43:26
mentioned for recruiting purposes, this is a very
43:29
important thing that you've done.
43:30
I think people have, have you have copied you on this?
43:34
And I'm wondering if this is if this year's too early for that.
43:37
Thanks, Elaine. Yes, it's a bit of a tough thing
43:41
because like when we're doing AI day, we find that some of our
43:44
competitors have literally done a frame by frame examination of
43:47
our slides and everything we say and then copy us.
43:51
So you know, say like how what's the trade off, which does help
43:57
with recruiting, but then competitors look very closely in
44:01
copy us. I mean, that said, we should
44:07
probably, I mean, I guess we could consider the shareholder
44:12
beginning to be sort of an we can do, we can maybe go into
44:16
depth some, some amount of depth at the, I know, shareholder
44:19
meeting with respect to Optimus and AI and sort of that chip
44:28
stuff perhaps. Yeah, Tesla's also really
44:35
underrated in terms of AI chip design as well as AI software.
44:41
So like there's still not a chip that we would that that exists
44:46
that we would prefer to put in our car, that is an AI chip
44:50
that, that we would prefer to put in the car over our own.
44:54
And even though it's been now out for several years and we're
45:00
confident that the AFI chip will be a profound game changer.
45:07
In fact, it's so it's so powerful that we'll have to nerf
45:11
it or to some degree for markets outside of the US because it
45:16
flows way past the export restrictions.
45:19
So unless the export restrictions change, we actually
45:21
will have to know if our AI 5 show, which looks kind of weird.
45:25
Hopefully those hopefully we keep, you know, raising the bar
45:28
on export restrictions. Otherwise it gets a bit silly.
45:34
But we'll have a bunch of optimist robots on stage at the
45:39
shareholder meeting. The optimist lab is cool to see.
45:44
It looks like it basically looks like the set of Westworld.
45:48
You could robots in various stages, some of them you know,
45:53
in various stages of prepare, like, I don't know, some some
45:57
combination of like the tattooing junkyard and the and
46:01
West walls. That's very cool.
46:05
An optimist is walking around the office here in Palo Alto.
46:10
So 24/7 is just walking around like a small, I think.
46:17
And we're so optimistic, you know, the Tesla Diner.
46:21
So we both won't. Yeah.
46:23
So it's we'll go from a world where robots are rare to where
46:28
they're so common that you don't even look up.
46:33
Great. Thank you so much.
46:34
The next question comes from Edison at Deutsche Bank.
46:38
Edison, please unmute yourself. And then please go ahead and
46:47
unmute yourself, right? Well, Edison figures that out.
46:54
We will go to the next question, which is going to come from Dan
46:57
Levi at Barclays. Dan, please go ahead and unmute
47:02
yourself. Great.
47:05
Thank you. Yuan, you've talked about the
47:09
opportunity to put non Tesla owned vehicles into the robo
47:14
taxi network. Just talk about the gating
47:16
factors to enabling that and what timeline we should expect
47:20
on personally owned vehicles in the robo taxi network.
47:23
We. Haven't really thought hard
47:28
about that, but we need to make sure it works when the vehicles
47:33
are fully under our control. And yeah, it's kind of one step
47:38
at a time here. We don't want to jump the gun.
47:41
As I said, we're being paranoid about safety.
47:43
So, so it's like, but I, I guess, I guess it would like
47:54
next year is, I'd say confidently next year.
47:57
I'm not sure when next year, but confidently next year.
48:01
People who would be able to add or subtract their part to the
48:04
Tesla fleet. I mean, one thing to keep in
48:07
mind is that we will have some criteria because like even when
48:11
you put your car in an Uber or lift and go through a more
48:16
checklist process of making sure things are working.
48:18
Second Airbnb, Yes. Yeah.
48:21
So we will do something like that.
48:23
It's. Kind of better process.
48:24
Yes, we want, like Elon said, we are, we want to be paralleled
48:30
about security and assets. All things like tread on the
48:35
tire can have an impact on safety.
48:38
So that's why we would want to do some proper validation before
48:42
we let other cops come in. Can you follow up?
48:47
Yes, thank you. Could you just unpack the
48:51
different costs associated with scaling the robo taxi business
48:56
and how you think about funding those costs?
49:00
Are the cash flows in the auto business sufficient to fund it?
49:03
And if not, what other funding sources do you think you'd use?
49:07
You know, would you just fund it off the balance sheet?
49:13
Well, as soon as there is a clear cash flow stream
49:17
associated with any any product, you can debt finance it.
49:25
And in the internal. When in the intern, we will use
49:30
our balance sheet, but like once we get to a certain scale in
49:34
terms of the current revenues, like Elon said, we could get
49:38
into a easily earned kind of transaction to try and get
49:43
funding. Great.
49:48
Thanks so much. And we will now move on to Mark
49:52
from Goldman Sachs. Mark, please feel free to unmute
49:56
yourself. Yes, good afternoon.
50:00
Thank you very much for taking the questions.
50:02
With the FSD trials that Tesla has been offering to consumers
50:06
and the attention on self driving more generally, are you
50:08
able to comment more specifically on what you're
50:10
seeing with FSD subscription trends and take rates and help
50:15
us better understand how large FSD revenue may be currently?
50:19
So we've definitely, I, I mentioned in my in my opening
50:22
remarks since we've launched version 12 of FSD in North
50:28
America, we've definitely seen a marked improvement in the FSD
50:33
adaption. And it's the other thing which
50:36
we had also done last year is we did bring down the pricing and
50:40
we've made subscription much more affordable.
50:43
So we have seen you know, 25% increase since that time.
50:51
So which is an encouraging trend.
50:53
But honestly we, we've just started the story around
50:59
explaining the benefits of FSD with I like I said before we
51:03
released our vehicle safety report, even if you don't
51:07
believe in this anything else a car on FSD being 10X safer
51:12
should be a motivator. Plus the other thing is people
51:16
don't realise even at $99.00 a month, it's like you're getting
51:21
a personal show for for almost $3.33 a day.
51:26
And this is by far the biggest game changer, which I know we've
51:32
been talking about it because part of it is we live and
51:35
breathe it. But yeah, I feel like.
51:37
Most people still don't know. Yeah, but the best yard people
51:40
don't know it exists. And it's still like half of
51:44
Tesla owners who could use it haven't tried it even once.
51:49
Obviously this is something we want to educate them on.
51:52
So we've got to when they come in for service, we'll reach out
51:57
to them, send them like videos of how to make it work.
52:02
And much, much. It's such a shocking thing.
52:06
They don't, they don't think a car is capable of this.
52:09
So you have to, you have to show them and, and, and get them
52:11
comfortable with turning it on and off.
52:13
It's so trivial, but it's you know, it's like saying you've
52:16
got a cat that can you sing a dance, but it just looks like an
52:19
old cat and you're like, you know, until you see the cat sing
52:23
a dance and talk, you're like, you assume it's just a cat.
52:27
That's that's Tesla, Steve. Our car is intelligent.
52:32
And So what we're going to do to.
52:34
Elan's point, like we've been giving people free time to try
52:39
and try the FSD, but we'll start giving more prompts to say, OK,
52:45
this particular drive try FSD. So that I mean, because it's
52:50
literally seeing as believing, like Elon said, it's think of it
52:54
like a cat. It looks like a normal cat, but
52:56
this cat Can't Sing and dance. Yeah, same thing on great.
53:01
Yeah, and that 25% comment was 25.
53:05
Percent increase in the penetration, yes, since we've
53:07
seen the release of of V12 will be fair to you in North America.
53:10
Great, thank you. Mark, did you have a follow up
53:12
question? Yeah, I think Travis, Tesla has
53:16
historically said it would use pricing as one tool to help
53:18
drive auto vehicle growth as long as free cash flow stayed
53:21
positive, given the ability to monetize products like FSD.
53:24
I'm curious how you're thinking about pricing from here as a
53:27
potential tool to drive increased volumes given where
53:31
you stand with FSD as well as the fact that the IRA purchase
53:34
tax credits are poised to go away in the US started in the
53:37
fourth quarter. So should we expect more
53:40
meaningful price reductions given that monetization
53:42
potential? Or do you envision price
53:44
reductions be more limited compared to cost downs given
53:48
where free cash flow now stands? Thanks.
53:52
Well, we're in this like weird transition period where we'll
53:55
lose a lot of incentives in the US, incentives actually in many
53:58
other parts of the world, but we'll lose them in the US and at
54:04
the relatively early stages of autonomy.
54:09
On the other hand, autonomy is most advanced and most available
54:12
from a regulatory standpoint in the US.
54:15
So I mean, does that mean like we could have a few rough
54:19
quarters? Yeah, we probably could have a
54:22
few rough quarters. I'm not saying it will, but we
54:26
could. Q You know, Q4Q1, maybe Q2, but
54:32
once you get to autonomy at scale in the second-half and
54:36
half of next year, certainly by the end of next year, I think
54:40
the are we surprised if if Tesla's economics are not very
54:45
compelling, great, the next question is going to come from.
54:52
Will free truest at will, please.
54:55
Feel free. To unmute yourself when you're
54:56
ready. Great.
54:58
Thanks. So much for taking my questions.
55:01
1st, I'd like to ask for a little bit more detail about the
55:06
lower cost model that you talked about having I think started
55:11
production in the first half, but you said we'll ramp later.
55:14
At the last analyst days, I recall you talked about some
55:20
aspects of this, like two 2/3 of 3/4 reduction in silicon carbide
55:25
and not using rare earths in the in the motor and perhaps other
55:30
cost downs. You also had this unboxed
55:32
architecture that I think you said would not be part of this
55:35
sort of interim approach. Can you update us on what we
55:39
should expect this thing to actually look like?
55:43
Oh, we. Won't get into the looks.
55:45
Because, you know, let's just go model what just.
55:48
Lift the car out of the back there dancing cat there can
55:50
singing dance, but it can it can talk and sing a dance, though.
55:53
That's the cool part. Yeah.
55:56
I mean, the, the, the fundamentally the biggest
55:59
obstacle remains that people just don't have some people
56:02
don't the, the desire to buy the car is very high.
56:05
Just people don't have enough money in in the bank account to
56:08
buy it. Literally that's the issue.
56:10
Not, not a lack of desire, but a lack of ability.
56:13
So the more affordable we can make the car, the better.
56:18
I think it's going to make it will be a very big deal when
56:20
people can release their car to the fleet and have it earn money
56:25
for them, which I'd like I said, I'm, I feel confident in saying
56:29
that'll happen next year in the US at least.
56:31
And in the US we're, we're legally allowed, you know,
56:35
appropriate disclaimers and that'll make the affordability
56:40
dramatically greater. Just like you know, if you, if
56:44
you have an Airbnb and you rent out your, your home when you're
56:48
not there, rent out a guest room or guest house or something like
56:51
that, your the affordability of your home is much greater.
56:57
OK, trying another. Another topic then, can you
57:01
know, we, we see all these wonderful developments at XAI
57:05
like Grok and you know, obviously Tesla's trying to do
57:10
quite a bit in AI. Elon, how do you manage the
57:14
division of efforts and recruiting and talent and
57:19
capital between these two that that seem like there's a very
57:24
high potential that they could in fact compete?
57:29
Well, they are doing different things here.
57:32
So you know the XAI is doing like you know, TB scale models
57:45
and multi TB scale models. You know, Tesla's 100 times
57:50
smaller models, so one's real world AI and one's kind of, I
57:57
guess a lot of visual super intelligence type of thing.
58:03
The, I mean, the really kind of the, the genesis for XAI was
58:06
that there were, there were certain people who some people
58:08
would not join Tesla AI engineers because they wanted to
58:12
open ASI and they would join Tesla.
58:16
And I was like, well, maybe they'll join a new company.
58:20
And I think the Tesla problem is extremely important, but not
58:23
everyone agrees with me on that. And so rather than have them
58:27
join, you know, open AI or Google or whatever, some other
58:31
company, it's like as well have them create a company in that
58:36
regard for it, which is like, which is like AI.
58:42
So that's, you know, and make a make a decision.
58:48
Do they want to work on like super intelligence data center
58:54
or real world AIS? They're both compelling
58:57
problems, but some people want to work on one and some want to
59:01
work on the other. Great.
59:08
And unfortunately. That is all the time we have
59:11
today. Thank you everyone so much for
59:13
your questions and you'll see you next quarter.
59:15
Hey, thank you so much for listening today.
59:16
I really. Do appreciate your support.
59:18
If you could take a second and hit the subscribe or the follow
59:21
button on whatever podcast platform that you're listening
59:23
on right now, I greatly appreciate it.
59:26
It helps out the show tremendously and you'll never
59:29
miss an episode. And each episode is about 10
59:32
minutes or less to get you caught up quickly.
59:35
And please, if you want to support the show even more, go
59:38
to Atreoncom Stage Zero and please take care of yourselves
59:43
and each other, and I'll see you tomorrow.