Tesla (TSLA) 23 Apr 242024 Q1 Earnings call transcript

[Audio Gap] Tesla's First Quarter 2024 Q&A Webcast. My name is Martin Viecha, VP of Investor Relations, and I'm joined today by Elon Musk, Vaibhav Taneja and a number of other executives.

Our Q1 results were announced at about 3:00 p.m. Central Time in the update deck we published at the same link as this webcast.

During this call, we will discuss our business outlook and make forward-looking statements. on and based today. as are comments our expectations predictions of These our with events materially uncertainties, risks in differ results including recent to mentioned Actual could of due and SEC. the a number those most filings and

Q&A, [Operator opening has remarks. jump some Elon Instructions] we But before into Elon?

Thanks, Martin.

and globally as we EV pursuing Model recap, We pressure Fremont. X in right challenges EVs is There navigated all manufacturers will ramp the of the is was several we to as are pulling rate order of under vehicles the a other plug-in ultimately lot as have back not market. in dominate adoption strategy electric on believe seen, So and and unforeseen -- QX, instead. well the the updated hybrids this

the that faster leading continue team in increase. We And profitability likely as of Megapack, great will that job will, it challenges, environment business. to energy know the all-time these It in significantly car so an to QX, reached to the and a we Tesla high business storage ahead. executing actually and did in looks energy than quarters years expected. record in deployments increase tough a the Despite for particular,

training also training more compute We capacity sequentially. our to continue than expand AI QX, in doubling our

half several previously upcoming second in the accelerate models ahead. updated product We've to our the of of of mentioned of line new vehicle a vehicle XXXX. the weeks. start production about road of new talk in past terms -- lot launch In map, there's I our the of been the lineup next future in

early to not be if this So we late XXXX, it like year. more the expect

be well our the to use produce the we'll more manufacturing vehicles, lineup. new platform our same of on current platforms, current as aspects affordable including These and vehicle lines aspects as as able will of models, next-generation

our made It when current more efficiently. production to X any should of it's And factory So over production new this be get we realized full not or much new will vehicles to extent. massive think on million line. contingent upon the us lines to allow capacity

is It's try Regarding the haven't strongly computer, and Hardware now you FSD improvement rate I out. we've cameras to it for urge the in America. rapid. North of the self-driving, inference that VXX, if from profound and you experienced turned cars, And all pure AI-based which on X on, everything with this, is

I increasing and about people think, So we're it's with million is seeing of around vehicles, to, passing that half percentage week. each been far so it pushed and out X.X use

over VXX. So XXX have been we now have driven billion FSD miles that with

drive. become Since humans -- Supervised clear the Self-Driving, how it's end-to-end autonomy. launch vision-based networks And is scalable Full it's Self-Driving approach really the the right very Full with for solution of that neural

nets eyes. Our entire road biological for network neural is and designed

robotaxi or So out. we've road To nets naturally, it solution to our announced, easy Cybercab reduced and a try make it's digital be to cameras are month, purpose-built system. more August. showcasing neural will the to the so $XX And our we've price subscription we as in current accessible,

Regarding AI compute.

on been Over past expanding actively few months, AI working we've core the infrastructure. Tesla's

For in were a we our while there, training-constrained progress.

we're point, this progress. so at longer training-constrained, making and rapid no are, We

a the or new say G computers thereabouts commissioned, stands when be Roughly by GPU or is word, word. actually probably are stands -- end that XX,XXX not. year [ XX,XXX and it's We've a installed XX,XXX GPU ] in training, HXXXS this for of expect working, we meaning just graphics. active, because training. to and wrong for always I they wentz I HXXX they're need a GPUs. like GPU feel

used. possible as not HXXXs, about how just the being we're efficiently in that they're sure making number but our of It's are efficient We as training.

So in conclusion, map. autonomy road we're super excited about our

driving who's humans anyone time to is obvious I be we've that. exceed and should reliability ago cars that boat back several a a we that to And with a said will a not of future, VXX Tesla [ we're and and like electric in something the obvious gasoline before the years only autonomous And hindsight. of I very in in time autonomous it be will that that headed an it for horse vehicle using go much matter think are I future. And ]. riding flip not become really

profits during that thank forward investments this will to Tesla look I continue make for to future, we incredible to Tesla and We everything and that planned have period execution the in for drive the wanted necessary growth the and Thanks. team ahead.

Vaibhav and some as much, very you comments well. Thank has

offset see to earlier. Auto these margins XX.X% that excluding per by important reductions did vehicles declined from of did acknowledge to auto in business seasonality, actions had feature we primarily largely and previously uncertain revenues impact environment in other recognition Elon a impact functionality. the and perspective, and our unit quarter-over-quarter the what many reasons, It's Cybertruck. Autopark for XX.X%, certain pricing have the not mentioned Elon costs in were was macroeconomic revenue that because decline of which of from of said, U.S. from The

and due while costs were cost the by in offset experience did initiatives. disruptions reduction cost we ramp in Berlin, of Additionally, Model largely higher to Fremont X these

revenue of vehicle Currently, exclude and slightly. to we improved that Fremont. costs, ramp Fremont and Y very auto close Model Austin X already in are the normalized Autopark, Berlin margins cost Cybertruck if from fact, Model In per

business. efficiency Our to efforts quality ability across Tesla's without the executing costs of amazing on sacrificing the pursuit team in to was reduce the due relentless of

as We've We went credits, other others stream in. would us. seeing for means future that also of industry revenue for in from appetite back we the is that the Obviously, and increasing investments EV is a prefer want. OEMs EV, their there all steady on whole pull not back are that pulling witnessed

options month Model in and FSD, U.S. extremely demand a the expanding for in specials the certain $XXX purchase products. including of options financing initiatives, attractive in undertaken we've our launching paired both subscription On driving financing, for front, that reach believe attractive of and our way the price our X and activities, leasing a awareness the the variety long a markets. We lowering demand for will for with attractive offering go

Our a margins progress, record of Energy continues with meaningful business reaching XX.X%. to make

the energy And for deployments expect business to We XX% XXXX. significantly our will contributing to XXXX least this storage accordingly, from grow overall at begin profitability. higher

lumpiness of control, our may that a that fluctuate variety deployments there due a quarter-over-quarter. our is to of so Note in outside factors are of bit storage deployments

had and flow of in other AI increase On the the operating $X.X free sequential a initiatives, investment projects, activities. in negative saw first cash billion We front, our expense future from continued marketing we primary and elevated in initiatives, this deliveries spend various a mismatch AI CapEx across inventory The between on was of quarter. driver and discussed compute. an before, builds as including increase our from

to the to to quarter in cash return We expect build free the and again. second inventory flow positive reverse

As annual necessary excess we for decision over reduce an the of generated phase had to on XX%. next we of count company rate The the be our growth, savings by hard $X prepare the head are run but well in basis. billion to to make expected

manner. hyper getting installed our CapEx are We efficient in focused utilizing more a capacity also efficiency and on

AI. ultimately our The overall and extremely savings improve In cost the including the from increase will these bright initiatives, is scale us conclusion, help to enable investments profitability our of future reductions in rewarding. again, whole to great would And thank to journey be Tesla there get the can while challenging team will Once I results. like Q&A. extremely we it up for and to the delivering open our

Okay.

with investor Q&A. start Let's

the What is question status is, what is XXXX? The the current of first Lars? output?

a week Sure. XXXX X. production about from about increased greater X Cybertruck, than for we which is XX% gigawatt per over as hours posted -- on XX%, QX, year X,XXX reaching

ramp we're cell in of throughout we We Phase make ramp we're increases. driven the as lines and production COGS expect Because by X to multiple of drop QX to X, cell with stay third the improvements sure ahead of rapidly yield throughout and continues ahead lines maintaining the to volume ramping, while week-over-week, ramp. the inventory production weeks the Cybertruck

of of year. to by our beat is the and expect cost goal, this, So supplier cells we the nickel-based do end to

Thank you.

Optimus. on is question second The

So Optimus? currently When current any tasks? expect mass to status performing start Are of what do you production? is the they factory

tasks at factory We tasks should I are lab. able or to do say, simple least, the in factory

-- do production think itself, end it in we may year. this of factory then will externally factory have are we of sell be before in in next limited just Optimus -- we And actually think end by of natural the terms year. These guesses. I doing the In useful to the able tasks

do more is before, to humanoid said will that think else reality combined. no I of I've is economy. navigate the request, the be a limit if tasks able sentient size you've than valuable at got Because Optimus to As robots there meaningful and everything

reach be on I And is robot happen. of efficient volume production to So inference with positioned to that's what's to humanoid itself. maker robot going Tesla best think any able the

do in many had have is car. is this, that inference the were we is the We perhaps, choice. constrained I a ways. of the pay because to close in that We've efficiency company. Tesla's AI will company But to dividends vastly than inference There's other Tesla. a -- anyone efficiency mean no hardware any don't by that better inference worth point even emphasizing.

you. Thank

to about question pathway threshold regulatory in The the And human how is, FSD the for of is what U.S.? the third safety think the we drivers? approval unsupervised compared assessment towards should appropriate

states that are adopted guides driver-less I the already for start. states are These operations the vehicles. such can broader of autonomous data have for paving a handful laws. way adoption operations of while There vehicle a

about these U.S., we're in and for work the as I and well pave well. that a way think our data at we states the a Ashok the but will little approval talk ongoing regulatory providing least, bit safety in broad-based countries then can a expect as other methodology, that as

actually been It's car been helpful companies pretty path autonomous have other jungle. regulatory cutting through that Yes. the a

And been helpful. have a obviously quite in they actually that's while. for So Francisco San operating

of approval for I think we L.A. City got

So happening rapidly. are these approvals

a rate that to because say, that's if stopping autonomy people. statistically of a conclusively think at scale that got that half let's think accident human-driven significant has, the killing car I shows autonomous point, means car, at amount of the you've data difficult ignore I

So much that a with view, car. would half operated is guy floors. be regulatory like guy autonomous And there relay barriers be tired, in make just switch. a shows human-driven be safer by significant a my somebody I was that elevators. between actually the used or do mistake, get and than car will not will think this that sometimes there Elevators conclusive or data a But grunt, in to provided

now elevator We a and don't button. think we just press about on it. So get an

And a work. just if standing that fact, will of be relay switch. with is somebody In weird there how cars kind

takes phone, get using out. the in, your destination, a get just you you car summon You it you to

You don't it. think even about

elevator. an It you like You to just floor. think takes don't even it. your about it, That's

You about working how or anything it, the that. don't like elevator is think

And the clarify something I should operating is fleet. Tesla be will that

exclusively can like friends X-star you number in time, user user. them that and So will to and cars think used how Tesla number only Airbnb could and end be Tesla that car come they're the car fleet you to of then owned At Airbnb. meaning by anyone. of only by be if the owns they some can decide operates Uber, or -- there'll to have think can There they they be the That of of and and combination cars an back or where whenever some there or cars will let like a be by and like some car end their itself fleet. any want, of add the -- Tesla by of bunch the want family theirs, they users subtract be

any You time you not want. could guestroom rent out your or

searches of you because in the we wrong, training constantly added is X happening gets has loop, of cars to and compete get Google and It's Search tens to people something fleet So training the doing to goes cars, clicking getting millions time and that cars this have way flywheel Google grows, loop. worldwide. are with that a Google feedback flywheel. our same sort very million every data constant the difficult X as of With going -- million eventually feedback that

of with millions. same So a but is comprehend. scale difficult But maybe ultimately, to Tesla, it be tens at that will

should -- and we've but distributed of I cars, a X the inference very Hardware here hopefully actually inference. potential much X. an think the And cars next Hardware is run designed year. towards pretty element is there's in when got being because potential down X also road cars where are made car not all be we've a for AWS to there's if powerful in the to now moving, end run some got with Hardware the

a an inference. So takes magnitude of kind model, of of like to compute it. less many AWS, but run computers orders it to but AI train lot distributed Like

a perhaps where not future of used And it the computer car autonomous it is is on And future So It's XXX seems maybe gigawatts distributed doing week. a used could compute. still used it. got of kilowatt a path compute, hours a world. where if fleet can together XXX average, to the million AI even you hours to the use XX imagine instead like there's an like inference they've XX week, leaves compute of XXX all of a where [ in of that's hard ] a put pretty inference something hours Teslas, will being XXX inference car be week and gigawatts around a waste else. it That be over

Ashok, in in do chime on to want your process you safety?

of safety. trajectories can car, many, week, train In like other sure of we critical improving the neural to the Those them or all networks we our already in over critical how we through clips events given that from we replay different someone have any that many produce we through hundreds own QA. our collected Yes, have gathered years we jumping are millions users like drive of the multiple out and that net have replay for validating of and them make to like years, are safety. events, front that database

systems. simulation have we then And

this also recreate in fashion. to this We and test try close to

we to our it validated, networks. is give QA this of some And

like that week New a we They have of to once We net some have are new we initially Austin, locations. what process. QA customers. build. cities, them they a Los in And that that all improvement, compared like critical employees of then this in improvement the they is different hundreds an driving San Angeles, we own in shipping of York, previous And collecting it we like They validated, did then our and feedback build on X,XXX this also net have noting different will the not give external improvement lot only of would estimate to start events, an miles, the it's they're Francisco, early is go users, confidence are if to issues capture and are after builds. real-world or we to

every sorted have when by live we And even like we that's it. event go happening external, of the fleet in monitoring dashboards criticality the of critical

safety we failures data to, So and any the are in pulse the training the the improvement constant then having the on we'll cycle. add the that along a to And Elon next model alluded build quality it and get like the way. improves back,

how and fixes, So be code It's with mostly they without on the issues, like on architecture, we engineers much in learning VXX especially engineering loop this rinse to And don't own based repeat. algorithms. automatically evaluations like have interventions data. sense and this have improving and all of constant new the its creative is that of then feedback requiring the

it's no There's rotate getting. based everything saying like new soft, and you circular. that, like person is drive must data or by intersection neural data it pretty instead neural probabilistic much how learns that, this the see it's oh, that you back. a or this they network, is or probabilistic on hard data trained engineers from that. it something automatically some get this you chooses, If like that, conditions. the That's every steering to that here, and add the failure We this how wheel it's inference distribution something network, So of the is okay,

like, our than we the them things because insight is we some let's good but have have X fix. car, months the be say, to capable will models in X more need with into how advanced have And far or that we that some Yes. do in issues what

are to to step are quirks some of need will in they a it. but be be that addressed order So have release capabilities improvement it there'll that car, in change the the --

As be to very we general. or careful release customers what to we fleet in Ashok saying, was in have the

are neural XX.X retrain in because nets we case, are each close look XX.X, say even VXX, pretty if like at a could different. and to So substantially arguably VXX and -- total it's be the which really of

where well have good into is, car say, we the or in, will model X insight X the months. So how perform,

Yes.

In in model we have terms performance, lot of loss out they scaling, scaling, corresponding to scaling their generally models model gains people they figured in making also where coming scaling, increase here and then in also the talk a data addition about the scaling. and side and other access but size

train of training the you also gains data that amount use network scale gives you also the can increase neural compute. similar by to and can up You and

You and better same And you or can more amount one you compute can the scaling train better that better more of architectures it for also gives Dojo much have nodes, architecture results. GPUs also performance. longer where produce with and for count

the combination a of you future what to video scaling can it extrapolate, to model Obviously, is all a train, be can predict scaling at have basically past and takes performance. of going seen them, on they're few in time generally few but can true scaling, and based with estimate do the past, weeks clips and trends experiments tens of we we based weeks scaling, the takes a collect because the of that progress takes compute size millions process based held it okay, we on the data. it ratio, training architecture to data scaling, So to future continue the this

$XX,XXX vehicle? to very line Okay. which get go the announcement official we can the the I'll an next you question, is, much. for Thank time of

in But affordable our mentioned, possible. most accelerate lines mission, we as our to utilize our to of to CapEx. our fast Elon on future all we lineup think capacity cars before These launch vehicles That's opening spend the mentioned more a I that's the efficient shift, major customers and get new open we're vehicles go as to the existing it CapEx our updating way. low-cost vehicle a and he capacity, remarks. as marginal built with high in

Yes. way appreciation I terms the asset We'll being that this And fleet. can value is to autonomy entirely But history happens of full you on about for turn the day when Tesla autonomy solving talk X. of almost when August able be unsupervised more might do self-driving. to gigantic and it in in think a think really, biggest on that

cars? million X

less. little I million of this a million about like of within it's several particular -- talking be the It millions Yes. million then X think. tens before end A the so XX cars, or will then of in cars eventually, and year it's we're cars. Not eventually, and tens of decade,

The is, Cybertruck progress is the of what next ramp? the question

to Cybertruck that a I can This at take too. X since X the late first ago. SOP-ed just week happened one year. months last of couple weeks X,XXX in a we

That's Of and is limitations, efficiency costs quality. drives some are still lot like a continue and volume faces and technologies, focusing matters. just of what so supplier production course, costs, et challenges dropping. to ramp new cetera, many this our But what with cost the year, so on ramp

Okay. question, any FSD possibly about you. automakers the Tesla licensing the Thank legacy future? have of contacted The next in

major one with regarding conversations licensing in We're automaker FSD.

question the robotaxi next is The about already unveil, Elon talked about you. that. Thank

about next-generation The vehicle. We about we'll talked question have till is already that. So the wait following to August.

time the the Semi. is to go let's So scaling for line What Semi?

In a Reno. fleet started year construction high-volume the cost-effective enable the we on So production our fleet we fleet, in pilot learnings like this have expanding and we the shareholders' are showed factory of our with Pepsi's which marginally. Semi as to parallel, in deck, engineering super from and finalizing our we're

XXXX, Our in first customers for external XXXX. starting with are vehicles late planned

Okay. A more couple questions. of

FSD make autonomy? favorite, can transfer is X until Level fully FSD delivered So our permanent we with

No.

Okay.

the Next at of at question. getting run getting Where -- the see end Mike? the What Megapack production the the is do Lathrop? is what ramp year? you rate

ramping as planned. is Lathrop Yes.

That year us -- second to gigawatt per to nothing exit year of hours line of our end by increase ramp. the the year. limiting rate the our really XX the line this allowing is year commissioned. gigawatt per start from GA have XX There's at to We hours

ship sales longer have typically the months we these dates. for to Given to large XX visibility cycles projects, prior order XX

to plan So several we're advance. plan build the able in quarters

to business So ramp this align allows the and factory to with order the us growth.

as customers we'd for our these to Lastly, thank partner trust a like projects. for globally Tesla their incredible in

very Okay. Thank much. you

analyst to go Let's questions.

question The from from Sacconaghi Toni comes first Bernstein.

models frame, talked the on given work Or relative if Highland about a the elaborate models? in have quick Model tweaks lines? I Maybe them are these new we Model of you can wondering models, lot update? about these will a on in of be And be you kind retooling. X put on these was given of all new like And a to more X and you What lot context Highland Are just bit the that running like like that the little to of vehicles same I follow-up, going they're can has think existing time should required today. that? that help a how please. the And context. like

said we've that on were think I front. we

follow-up? So your what's

It's interests talk over you're of to heart you, do Tesla Maybe which involvement your is many is a years? leading next about your can point in expect your the Elon, with and you more terms companies right at at one personal where for, lessen important X just now. you any that

majority much take a And it the time work day is I'm like constitutes it me Tesla very much my the afternoon rare make is and off. going will of quite it I a week, future. work it be so -- as in prosperous. Sunday Well, it's prosperous every to sure to pretty of and is for

to Adam from Stanley. Let's Morgan go Jonas

Elon. Okay. Great. Hey,

sales in a room expect X%? above what's words, on potentially achieved Or degree confidence leave other XXXX. statement of rate XXXX But team's that So lower you notably lower your team that in growth than does your volume for and on year-on-year? growth

year. No, than I think we'll year have this last higher sales

on Elon, future products. follow-up, My Okay.

closer would more but your $XX,XXX one one castings, to I range, XXX-mile say piece, that nailed next-gen aggressive had you cheaper Giga vehicles, aside to don't pack If to getting price unboxed, putting best nail your piece, you your assuming robotaxi, take point, execution structural want a on execution, those cheaper Chinese better you, How best of offer long that long copy that? could from it and to it to you vehicle features years take competitors copy a now? unique couple would how Chinese competitors

I do, I drop sales, our they don't better in what at done competitors our can our we've have except you sales know than mean theirs. our China in relatively was than drop competitors the if less look versus drop because in

So we're doing well. But I think thought robotics [ Kathy be or ] should would company. say it of best, like really, we AI an as

company, If value as -- framework. you auto just it's just Tesla to an the fundamentally, like you just have wrong

If question, you answer right then ask is the impossible. the wrong

the to mean believe hours to if probably going are. -- investor have goes will but a X.X I we hours a Like day think is in company. somebody autonomy, So they should a like of we but Tesla XX%, XX be an is and that And doesn't not then week, it solve use you same. costs the I that car from

think autonomy, I to look give by day a thing Supervised, FSD actually is you believe in coming. key day. this at the that's It's you you getting better especially if this that should if preview didn't remember, right,

Yes.

visibility XX.X, and than to FSD then XX.X XX.X the significantly tried that, better not possible. you those what's and on. It's not have we be understand don't like is you've things, better really said, I If and going into going even

a have can I car But car a a company like And we as just at just look just said, autonomous. can't be more than have would because here, because we happening. company car company the why a car. cars that's it's And Yes.

-- in Yes. the addition community Tesla, rapidly. to This is is like all improving AI increasing just

putting auto mean automobile. I we're in Yes. actual the

about well, you're really That's like, a like, future we of point. point. actually, of sort making. So I'm carriages tell whole us need sort go well, the horse. whole doesn't horse it like the That's

Alex Piper Okay. from Thank question from next The you. Sandler. Potter comes

Yes, be. that thesis desire future AI, AI, couldn't In I about control to I you've The all obtain hinges your of company. things. And training, net spoken context, the Elon, that completely the of voting understand agree so more. full of neural why self-driving these XX% on would completely

not you'll which the come any that. obtain could core the if that not, I'm part that So you've be by of asking potentially I'm asking mechanism Because level can of about at if with control. up then risk. voting ensure necessarily thesis you

So have on topic. you might any that additional commentary

to indoors for there's disappeared autonomy, would I for future Well, I I momentum matter do it you we'll That or think aliens with can't I we have of products, that a a solve buy or little to solve Tesla the shareholders yes, cannot vehicles respect even beyond maybe I'd don't lot if autonomy least. on we one. win humanoid for but opportunity if Optimus no talking an how will know but with respect slower, I'll compensation, tomorrow, autonomy at more I solve would and that vehicles. of can't reticent then whole future is what we're a -- respect Tesla, meaningful with that And If Optimus. that. -- back even sort I some positive to in Tesla super-sentient If that a have robot opportunity. the And is if with if there's Yes, it yes, range enough things to company of can uncomfortable be fact. would They there's to you deployed. an escape, be level cash flow, we see. follow shares. that terminator-level a can we could generates risk, -- reratify influence over of ratify say have obviously guess not there's the buy

head result these what of discussion activities more types ways qualitative are the right. in the also you potentially for are the very All Maybe interested a of actually it that parting one helpful implications of you're as sacrificing all count OpEx quantifying impact That's presumably these final reductions, there. context. question reductions. thank I'd and on. I'll pass for be What folks? with

So said, certain some as efforts, like redundancies. count happens are duplication we've we there companies of There's across head the which reductions time, certain these over done board. in areas. And grow

get are, go look to need of these you and pockets rid at back So all it. where

remarks, went resilient this how So wherein This the stronger grows, we're any that through. the right company just future way end we set we're pruning. it future hey, and I about next to at my in of much because with deal it, said we'll which Like the get get be basically to the through needs growth? tree this like, through exercise have more And there. is for really the is to it which the going opening exercise we is phase bright. of do think and period And much we pruning

Yes, giving we're prosperity significant to of up of. had is now. I'm a not from aware that period anything XXXX that long We've

a growth and one company some with way. to a zygote just tail. it inefficiency. reorganize corrections really you that XX% reorganize and arms it, the need is growing you cell sort a and for start of year, of off and you kind when is and organizationally accumulates have But if and next of like start so company blastocyst to the X% human the legs along XX%, made time of we to phase wrong And per We've briefly,

the tail. But you shed

if company for it you is to different shed you then the basically, of have tail, A will be you're creature kind reorganize growth, you And organism. fail. like it of hopefully. growing. phases the And You a baby, don't

like your same -- cells. XX organizational feel walking versus colony roughly basically X the you're bacteria. XX cell versus have of that like doesn't trillion about feels number on Humans in depending can't like are a billion it's it and million body You X trillion, you're cells But XX XXX mass, It X one if Xx versus person. trillion. around structure versus

that for you've achieve got the anyway, fail company growth. So or it growth will new to to of a phase reorganize

Mark from to Goldman Sachs. Delaney Let's go

your The and it. in OEMs believing licensing business company previously potential licensing not how some you you had in the on been FSD really early discussions characterized phase much Can elaborate opportunity the today in milestones And with is be successful has at terms mentioned anything in technology agreement Tesla of achieve the progressed? licensing view? in a needs to reaching there order to product

just it we to right And Well, the is that -- is. approach needs approach. I obvious to be think think I our need it just

it around, inference the No if cameras is no have self-driving. our obvious can that LiDARs, you nothing. XX.X, radars, car think a and you I standard solution ultrasonic, now just with drive relatively computer with no achieve low-cost

work vehicle heavy integration for No manufacturers.

Yes.

becomes a remember, if cameras So the have wants inference fact, computer be don't our and just in car. nobody in king car. your of of car, hill... I same sort I -- case It's it obvious was Nokia having a it them Once software. remember you the of smart use this and really when that licensing a

[indiscernible]

They're -- a niche going basically still of limited they crushing. a brick understand did that with phones to functionality. and to be the phone the going with Android people way. And be And products. iPhone came all Yes, any sort then still out that are There's but not be a smartphone was product. that not

Or hand phones.

hand Yes, you the time last now saw was even a phone? [indiscernible]. When

I have no idea.

a In hotel.

hotel. Sometimes in a

the hotels Yes, have them.

licensing becomes becomes So obvious, I it. people you or will all don't that -- would optional. cars car to that buy sell will not smart will think not Once need understand be the or cars not nobody

Becomes a survival. method of

is. your it it License absolutely, will or nobody car. buy Yes,

I to thing lot want I the of time takes had in which put life before it cycle. take will add than OEMs, they a while. might point is years they deal mean product with talking have in these their earlier that, also but licensing They're which one just about a We conversations we've some product. their a out in I'll of other that it

is is. where them big this So difference between us the and

Yes, years. a would I mean, result it a car being now in really X deal in signed probably

early. be would That

basically. lightening like That's Yes.

eager OEM. That's an

Yes.

be do So sign surprised if we I wouldn't deal. a

maybe I inference is and it probably would chance think yes, car, even X computer. cameras need though a we more do deal have year, it's all you one. a But years before good be integrated than our with this a we sign

So about massive design just change. a talking

clarify, to which not have just again, we And they work which the Yes. it's do do, time. the which to take work it's the will have to

how understand has stay price offers prices with here basis from My the company you affordable credits lower were very the some also said especially free in the an still demand forward. on as reductions vehicles with have cars flow and helpful. from annual existing meaningful have reductions approach products? And pricing company making to the Tesla of the and Tesla's to here price can become, IRA for to driving follow-up was for access meaningfully product leasing going Do the current that Yes, cash that Previously, had incremental the see sense positive better that place. incremental set?

Yes.

positive think I free we cash flow can meaningfully. be

customers. Vaibhav his I said in we back cuts give price efforts, our it think like trying to cost it down opening offsetting remarks, like to basically were we're the the

given any have the if at price like excellent. true be a a you day, end at -- product for sell product great if company, will a the of any great the a I great price, great of That's sales Yes. you at area. mean

that over that price. moreover, accessible product need time, we're keep at great making is that So to do to great -- And it's people. price sure a a we

So to solve the affordability both you question. for and it's money not have value -- the fundamental

standpoint. dollars question they is sometimes fundamental people a hundred a somebody paycheck. of majority to affordability of don't overlooked. is a paycheck are fundamental several But If thousand year, think car affordability earning The living from the vast

difference cost month important a is improving the the makes refinancing price... is or and actually lease $XX it if So the way making keep to one the per keep to It other. affordability for

More accessible.

that the to the for keep over value Make exactly. accessible, better, and Yes, money time. more improving price

cash also But ] if want the buy. make people [ to flow

going at great a standards be price increasing. a great great it's what a product for Yes, at And product price. the constitutes great to keep

be -- like there's you static. can't just So

car and price, cost better, keep production, what making the we're the the to You but that's doing. have improving of improving

X people fact, car. think Lars Not the low explaining mean looks videos you added and value $XXX different. for and feels it is different, how as only out put it. don't that like and I said which and to lease my I can I in has fully like in have actually remarks it effort month. even also, -- much the car fantastic of gone, is we've engineering a it updated lot But as a the different, understand And so Yes. team opening Model much feels like looks revised

Yes. Without gas.

Yes.

comes from George from next The Canaccord. question

First, could timing China? additional including about comment of the launching clarifying maybe in help geographies, of us recent your please some you understand FSD

a new some we that bunch where should be cars I are we in. currently of yes, -- of selling don't are markets markets, mean acceleration like cars We'll there sell that. see we

And FSD new markets?

Yes.

neural a it end-to-end net-based is autonomy almost well So human, actually modification think like works about any in that pretty without the just market.

So we any plan that, China. on the -- a releasing in -- we as with for which autonomy market includes get of regulatory the that it approval regulators, we supervised can think where system approval

a you it's can in -- obviously, in and rent country foreign that drive live yes, a like drive and human, car you just country, go can a So pretty well, you if better. you'll

other car well drive make these country-specific the almost in drive we'll countries But training. quite better so And can it with everywhere.

Like are is driving basics traffic car, of same the The there's lights... everywhere. a basically car a

understands shouldn't It no it it where drives. that hit things, matter

that some Exactly. solid a in I rules need There U.S. you fined line do China, lane think. road cross to shouldn't it's over you And heavily change. get are you China, a you if to do recommendation, follow. that. a

something. the reductions. mostly actions, more like entire type it's change It's We but some have smaller not or to or do

have George, do you a follow-up?

Yes.

not release the you can why has And that results and or us curious you supply confidence I'm some deliveries that you impacted XXXX? mentioned my the as throughout that So whether . is unit maybe the you help and to feel in follow-up with first that? constraints have in quantify do quarter to growth

remarks little big macroeconomic one, a did We environment. Seasonality was of Red in from continued the had things opening attacks at our different bit this you. we We cover to QX are had lot pressure a think had which a I happening. the attacks. Sea factory.

X. Model We are ramping

things gives recur. mean feels these confidence it that, things that We we a in expect hey, to constrained are I these of are a that those activities All us almost ramping Cybertruck. like happening. culmination don't all period. And

be We think QX lot a Yes. will better.

Yes.

just thing sometimes, Or It's after one another. crazy. it's

It's delivered Yes, but are a it if and got logistics Y extremely one Model you cannot be -- got there, demand for you've on have they ships, situation. that just a you've if And the quite cars people. it's complex don't in sitting it's Model X obviously exactly. market, to excess

have So or overcomplicate week past process, so the the also which did in just the I'd simplified. say -- we greatly sales we've

far you can became buy it Tesla, buy whereas too So a should a under complex just minute. to be in a car it

you So to getting under we're what minute interface that from can a complex. quite in was Tesla buy a back

Okay. you. Thank

Colin Let's from Rusch go to Oppenheimer.

approach in happening now? talk really beyond your unlocking what's as Tesla world, leader distributed physical a is the of pursuit can the comments vehicle right in around Given the AI that inference, what in for you about

to something? want you say Do

be like probably mentioned, XX used a the week. Elon even when hours Like full to robotaxi, going Yes. car around it's it's a

XX a like hours guess, my That's week.

or left in and but hours and cleaning be going to It maintenance for be Yes. certainly some could that less, more charging then there's world.

like buying we send it. and close of for Like lot other to a these for of use in want can are large the And You them that. have bunch want right the a to workloads. workloads. those a expensive be do batch be lot they We [ wise and we those ] we and let leaving now neural like as this workloads them possible XX and chunk networks of of compute take XXX% lot use not pretty seeing, where already that as it basically a through Even now, through and compute are much documents run computer a use like companies this time example, have of be might to cars, these them idle. machinery to to make of don't paid

I Amazon when bookstore. it part started Services didn't most analogous to would it's Amazon the as out a where valuable think Web expect be AWS people that of

extreme compute then they nobody's that for levels of radar. they compute would found the spike the had needs year. the idle and year they periods because had for on rest to But the excess was brief compute of that So

of... for that rest of excess what year. they is the pull compute do kind then So That's the

it. Monetize

It a idle the say, of it. tens the of and that then useful. vehicles computers got of have no-brainer out of okay, like seems millions are monetize most them to kind Yes, there we've millions do well time we might something where if

exactly. Yes,

-- to, to think some and and vehicle useable a then I get compute million you own which I we kilowatt then anyone I you've maybe X will, of more than And think time, gigawatts point, then which at got the more useful -- mean XXX if the you you than any than like your might of Hardware by level, XXX company. order or could company, X have more on that compute, any probably of be get really

Yes, ] takes scientific solving car it terms this the of put this intelligence anyway. drive horse just to problems or else. or other you of uses, answer lot probably not because when driving to it's something car, intelligence [ a the And in

We've to already learned about these deploying workloads nodes.

cell And under it Tesla's phones, totally unlike control. our and is laptops

asking products as opposed own easier very be it's tedious. would So phones permission their nodes plus different to on cell the for users road see to

on Well, phone. you're the battery the just draining

Yes, exactly. The battery is also limited.

like you I yes, technically, get use just like suppose distributed but battery. phone at can't the amount you you So compute, -- full can't the run power would have Apple the it can't because and the of drain most

if kilowatt kilowatt which hours it XX is crazy the it's power even you for run car, hours use of whereas you plug not, could you've of XX for XX deal. and XX a kilowatt not a or compute hour big power. or phone, So a kilowatt-level your pack, compared you're computer, inference Whether got to still if

got with we there like built management. cold thermal car. Where exactly for data centers. in Yes, it's in already It's the

access distributed cooling. for. -- That power paid and distributed to generation was It's Yes. already power distributed

costs cooling, power and Yes. underestimate lot distributed of money. I people a mean that that

a big... Truly Yes.

but the a by chunk profit also the maybe. of, of CapEx small world is shared entire out and get sort a And Yes. then get it, they

Yes.

follow-up bit mundane. my just little And is more a

to capacity incremental target might XXXX how close that Looking can you when on the talk were accelerate you to start technology? and you ramp, at yields expansions about

the that. super in important good at near it's for don't I term. But making least progress on think We're

it As the by will -- continue end Lars will of suppliers And then it be exceed competitiveness the to said, improve. we year. this think of we'll

that it's to important think also I like ramp right the to mean ramp. note I is now relevant the Cybertruck Yes.

to And place we're them. a build just XXXXs to going so put randomly we have not like unless

technologies and suppliers we done XXXX, prudent have going found them. and great They're a make with partners, their all also our with putting cells. work of great in sure chemistry a so But about investments advancements And lot of cell to we're into you and in they've lot we're they're And that. also vendors. the development

happen Yes. part Because XXXX, a suppliers. Tesla difficult every a what our put with internal very it while for was mean there, orders. big against would hedge carmaker of the doing a because battery big I in cells was massive

of hour went crazy price And crazy numbers, so per lithium ion to the batteries kilowatt levels.

Bonkers.

Yes, just bonkers.

that whole cell it. want we've for okay, gigantic way were crazy then hedge due like like, Earth. to address deal have the the to kilowatt demand. in we order we an bunch hour have We did It's with cost cell take really much hedge a double orders here numbers suppliers cost from If not So program per got production, in to to have we carmaker kilowatt increase That's of hell on that it. the some think against by demand our hours on placed problems to what just about the we shocks every internal to anticipated. too per of with

Ben comes Okay. the question Kallo Baird. from from Thank And you. last

about guess, Well, to you really thinking strongly It drive really I to the that again, do this. do -- to recommend I can't not stock the should impossible just want is, anyone Tesla XX.X. who we'd say if like company it's understand you --

All right.

is from Ben Patil since question. So Research. Wolfe not try unmuting, Shreyas let's Final

for COGS current and vehicle the auto that mentioned you Y. year, decline XX% would next-gen last Day versus Investor the during the Elon, Just, per by X unit

curious around opportunity $XX,XXX other would that those talking be largely some the cost you're now the see COGS. or something powertrain reduction an But of around to the of some the about manufacturing I'm unboxed was products -- drivers material that of X/X transferable if you About of think coming was that that implying new process. cost savings, from I introducing?

but we of was possible. it as are in method that's seating, in mean, that's reduction the fast short, unboxed risks all because interior it units, we're revolutionary, in automation, yes, whether production and improvements I as to powertrains, and not, of and the components Yes, drive sure. manufacturing like transferable, products thermal great systems, subsystems battery some integration is lines get certain it certainly what mean, doing, manufacturing LV all comes their developed, but and trying controllers, with I

And not and -- so and we the cars the make to to yes, put to throw advantage we're going utilize that and it and and best it cars away take the we're and it just engineering it work, make. cough we utilize trying future of

battery you as other that I like rising some thought would you seems know the a topic of mentioned today, against costs seems just Okay. cost hedge have it a advantage it from you Great. like cells, it, OEMs. have even given the of plans against vehicle other of opportunity other more talking automakers. And on XXXX then an it of But really XXXX cells And about revenue there's to those you're I'm about now, those and I'm maybe you and your any automakers additional stream, to generate that convert manufacturing sell an thoughts really wondering, if maybe rationalizing just if curious that.

is What automakers orders be other batteries happening the from for -- to something, missing Great. seems that the have dramatically. I'm unless declined

is for a sales our capacity. past. from excess of It So prices we're competitive lot more much dramatically more in seeing than the clear that our suppliers competitive have suppliers

Yes.

is this to [indiscernible]. what -- Elon In addition

suppliers. what chain of supply us cell Elon did to what upstream about that's from us addition perspective help a XXXX, In said chain understand XXXX was for supply the our

XXXX, back for can partners, that to struck So overall a Tesla. to those also materials we our lot of reducing had the supply we the deals cost

the by supply that. in So we're basically chain doing upstream inserting ourselves

So reducing excess these pricing, have. addition the that's that the to been capacity also in suppliers overall in beneficial

wane, is mean I to No, Yes. this obviously. going and wax

boom Internet production in Russia. I the additional like Republican And Reduction be future. then and funny are complications true true be like There's Inflation something. it with from But cycle Yes, the today incentives boom and production a and Agency of or is a Like forth in kind somewhat there bust so IRA. So Irish like where a supply here. name. going back know, The what Army, to Act, exceeds Research government like, is will the going bust production be and yes, exceeds supply to it's and DRAM there's not battery cell then of don't

Retirement thing. Independent

vitamin It's situation, But IRA. is forced it complicates exactly. Yes, like which [indiscernible] incentive -- wins. the it IRA structure.

So the demand the IRA produced lasts? that U.S. cells But how U.S. is I -- is are than for long don't there's know. that in outside the stronger then

it's cells and why against of that important all [indiscernible] hedged is have many Which we cells both this.

Yes.

Okay. But that investment wanted Relations. month today. I the the I'll let have of like all I at same from Thank time, moving Investor And announcement. and That's met a be very to to the you time on and about would with the up that announced ago, much. make Vaibhav community we world Elon short I know a

hanging couple be another around for of or months I'll so.

my watching free hell and some from And quality out the just production of for beginning my inside years. the super the a But and of sprint, this the break at X with have especially thankful at family. going say time. to people the I'm to over wanted feel greatest reach So literally I'm professional X-year become worked of taking to the I'll dozens any it's this that in I after and these outside I've forget the spending that to what company to started been I good today. be time never privilege from memories room the life. people room see And of years

else is I've team's think anything Tesla strength I the my in career. seen teamwork and unlike at

and opportunity regularly you with very me. feedback for in thank much back that XXXX. debating this and seeking Thank I Elon, investor it for you got

best reached was mean Yes. your was had thought you Well, I because reason I I seen. the the out I Tesla to analysis that of

Thank you.

with Yes, been is over to thank to where It's years. helping pleasure Tesla it you get for working a today X you.

and so ones. the side, in. years met the over thousands much. walked even listening the yes, all for be next tough very around the of and to but the looking you you you shareholders Thank thank that the And Thank X we've on factories, I'll call yes, in all months, And interactions, loved much. forward other the