SES - Earnings call transcript Q2 2024

Due to licensing restrictions, you must log in to view earnings call transcripts.

Operator

Good afternoon. Thank you for attending today's SES AI Second Quarter 2024 Business and Financial Results. My name is Cole, and I'll be the moderator for today's call. [Operator Instructions] I'd now like to pass it over to Kyle Pilkington. Please go ahead.

Kyle Pilkington

to call welcome quarter our and results. XXXX everyone, our covering conference second Hello,

are Founder, Qichao Executive today me Joining Hu, Jing Chief Officer; Chairman Nealis, Financial Officer. and Chief and

today, shareholder after as our as results. financial our a issued We well market close business which letter update provides

You'll find our of link website ses.ai. with release to in letter section shareholder a Investor call conference Relations press today's webcast a our in the at

expectations or started, or as and this statements today. may is statements in implied results a different statements performance the cause of may assumptions that meaning today involve information our are certain materially of predictions These these get uncertainties, reminder contain statements. and discussion which applicable and future forward-looking we forward-looking Such actual to expressed on or those our legislation. be the within Before risks, from securities based

risks over The current our our business to SEC in review pass our detailed our differ afternoon, that, as that we from those release to, and in materially but limited will cause and quarter. it filings. for as With uncertainties Qichao. This the not earnings I'll results latest our include, to results expectations could well are

Qichao Hu

models, margins. the transformative for decade. you from next AI afternoon, this and industries language grappling our set about earnings second represents seeking quarter I of us Good thank a is to development This want seismic across LLMs. innovation with joining AI those to any large generative shrinking those and on shift call. pivotal the talk to technology disrupt S-curve industry by

and years players just year years completely the ahead become battery The market. now market The must completely ago. different The to fact ago is of is battery different battery something X X deliver EV dominate companies next-generation light global relevant. incumbent from even or today's that

addition that mobility mobility to market, UAM and we entering including their in compete we Previously, drones or work. EV announced cannot We air urban on are our terms. existing the air

For quality, very overcome why batteries, with to approach has development. we commercial at next-gen the scale: and slow. next-gen That's safety, & too human-based of compete been introduction material must technologies batteries battery long. takes traditional always incumbent future simply X hurdles The

world's the in are lithium We metal. leader

the automotive joint agreements with enter A-sample world's first were automakers. We to B-sample development and global

very We have in developed and capabilities exciting materials manufacturing.

manufacturing integrated strategically support. have our We into encompassing and development, operations, technology design, AI aftermarket

mindset, generate on by business when embarking into materializes the We've AI metal, both realized the are reshapes thematic it share the embedding to have initiatives. that adopting and Since model of lithium value external preliminary diligently to internal to excited of our we fundamentally aim this with platform-building value. a outcomes we approach and AI achieve worked

paradigm Today, we're shift. introducing a

solutions represents new forefront will but ultimately AI accelerate this technologies. agnostic will be next-gen AI our the of of Our the Lithium metal all any commercialization better to approach, battery technology.

yield lithium of hour amp the manufacture the of XXX This completion year. this capacity hour to largest our the with to quarter, are Korea. Last their with development hit our sector. build center South we the of in amp metal will B-sample announced start XX share globally will one EV target Let's I'm within large line on track Hyundai joint to line electrification Ui-Wang, and glad lines we a quarter cells. that automotive to fourth lithium of in B-sample partnership metal

to to reach and continue our XXXX C-sample a XXXX. with in SOP of production automotive goal OEMs with EV We work start in

drones, UAM demand. to For we strong and continue see

in EV South previous we Shanghai A-sample UAM lines are converting Korea our lines. to and UAM, For

in and test, September. producing expect August; cells start SAT, in Korea test, We UAM in the to line FAT, field acceptance acceptance September complete site

to UAM lines XX UAM in test amp and in testing amp hour aviation September SAT complete start lithium great lithium-metal making and rigorous October. and make for line on progress modules producing We cells based medium to these modules. expect the hour FAT safety metal Both cells the XX Shanghai We're both will certification.

and We later testing entered expect enter this to OEMs have with a agreements year. a already leading few more UAM few cell

we from estimated small both The especially billion according are to industrial to estimated demand market for in drones, for X.Xx growing and For Intelligence. customers, size AR/VR about the be seeing Concentric to XXXX, according drone drones. $XX defense $XX XXXX, in goggles was swarm billion SkyQuest, market

X make X already have hour, and cell We metal our small to amp lithium small amp converted to modules. lines hour cells

our talk about let's Now AI solutions.

have for Science. X: AI AI We AI for for Safety and Manufacturing,

and to optimizing traditional quality approach First, define process least quality human and engineers Manufacturing. AI is experience at specifications where and takes manufacturing cell typically the design for X improving The human years. through optimize

an Battery good science, than more the a best. of often especially manufacturing is the and very ones art between

learning AI human While for It approach line cell lithium-ion define Manufacturing specifications it commercialization machine mature this in which past works based this data well the manufacturing believe and We faster technologies, can of human-based engineers. to quality accurate by uses more fine-tune time technologies. down than accelerate has much and on XXx. for today collected, next-gen process battery worked slows and large-scale is

series complex checkpoints cells B-sample, EV quality amount We produce time ultrasound steps month per than with X,XXX individual manufacturing anywhere both and than more There cells in of such largest x-ray, and an of lines line data process Our data, are lithium vision. UAM thousands per enormous drone more the of produce growing. relationships. the including are and per metal group X,XXX and images data CT, cell and growing, and There as world.

our B-sample, AI pretrained been on Manufacturing excited XX,XXX of all to metal lithim cells. announce from very of model for liter working for Our than AI EV the small to to already manufacturing metal drones. We're installation on UAM, more has

group optimization will and quality, will it the analysis. of manufacturing. drone expect accelerate accurate quality us and C-sample individual manufacturing steps We EV provide larger-scale relationship and UAM analysis and This for very further of detailed preparing step

the Manufacturing approach industry. development, for partner semiconductor companies plan the Manufacturing AI in-house latest to to tech for we big with addition AI also from In incorporate and

We AI enormous allows This for SOP continue a in goal value work partners. to capability OEM to automotive bring EV OEMs with XXXX C-sample and auto our in to Manufacturing to manufacturing XXXX. our with reach battery large and us

for conditions on example, AI name set vehicle Traditional monitoring to by of safety health, humans, physics-based based state state determined few. for charge, boundary battery current; of health These a include, temperature; are and Safety. of voltage; prediction Second, a models. time, SOH; SOC; capacity; would

best profiles, solely we with understood more pretrained predict actual of XX,XXX Safety, relying conditions, powerful life enough data on boundary and under at of well for our human-developed more cells than accurate than engineers. hours flight LLM rather AI In While more drones and XXX humans, the various they is cycling remaining our mission the anomalies have than AI are boundary a metal to lithium using useful battery modules. far than detecting conditions are by including not actually lithium metal incidents. human even cell

and addressing XXX% the which can that far AI these issue AI LLM features we early lithium vehicle are accurately. densities, warning and developed work detect models. safety These energy send signals the we all safety. training, metal next-gen core the Interestingly, guarantee near help identifies features the anomalies with that Safety batteries With believe is data working remarkably, more can in field, of of expandability higher more improving for battery and on

predict of AI Safety than incidents. with for OEM has XXX% model able partners, our been our more working XX In to

predicted model earlier to Our they sent warning XX the and signals. cycles XX occurred than incidents

to continue able our the to around predict We human-based only the verify actual of the also In accuracy. cycle models until incident. were comparison, test incidents the prediction XX%

been has Science. battery AI in single Third, technologies. step Human next-gen for battery research development materials of the and commercialization slowest on

For XX When to that are example, less than could the studying billion, studied should that's studied. lithium-ion global years the XX XXth industry unique entire unique XXX be and X,XXX molecules. be there spent molecules

We years AI On new enormous takes Unlike exist. and Manufacturing introduce that AI the Science it to Science do year. vehicles. actual can material. XX for database not in collect currently human requires does X for property believe for that lines and average, data molecular from Safety AI a that scientists an battery

massive map database labs properties initiative small participated Xth have have and Molecular in and companies molecules. molecules. property initiative, the Recently, big this subsidized to and computing national a universities, power. Several this started whose Universe, we we the to requires crowdsource tech XX of free to Synthesizing resource goal is already called about new computing mapped

of Electrolyte metal in the we Electrolyte built more lithium this to from to not and That's sufficient synthesis new and mapping been molecular a to are best reach drones, With expect GPUs, large screening the just Universe electrolyte of Once commercially and lithium-ion models not accuracy. ability formulation batteries to UAM map, any that we to enough since large synthesis this training and includes applications. model for chemists This tools, high-throughput purification organic for This complete automotive material Molecular which electronics, for testing. consumer molecular available. generative cell new we some map April now world, have our discovery have completely year. small will EV operational molecules these can we AI why and foundry, has power Most employs problem. accelerate and for but other the molecules we battery AI also

battery in has capability. complete such the No industry one a

as of transportation. represent streams These expect do electric we future how revenue to sooner-than-expected all this? be So monetize X what solutions and AI well we exciting as the

ensure near vehicle in AI XXX% In and Safety, quality data Manufacturing truly the safety be safety for to manufacturing must integrated. and field,

UAM comes Here's are time cells that be lithium the safety safety. ensure where manufacturing EV, SES XXX% and metal and near our for data first will integrated to in, AI drones

accurate data, of next-gen become. more data with model safety consolidate batteries both and some peers We're more diverse and The working to models and our the training. larger also manufacturing lithium-metal our in for the lithium-ion

We life structured prediction these prevent value incident costly the that as remaining extend help premium can period. for battery expect valued warranty be more accurate OEMs a prediction the to proposition recalls The entire lifetime. can useful and is pricing for

not Science, for We SES their resource capability. OEMs help AI good solve develop intelligence Many challenges. In the materials. electrolyte the and companies development electrolyte strongest have to and them in-source do AI has battery can battery

We to metal lithium by seeking human scientists. by Efficiency the record Electrolyte Coulombic will beat start set

applications then performance to expand lithium-ion to grid and such automotive other expand to applications. will non-volatility consumer electronics from We and low temperature fast and as many storage charge

analogy an model find the AI enjoy and The a in fee type This much licensing profit that in-source the intelligence industries development be royalty. of higher based business may pricing pharmaceutical structure Science recurring for on can margins. for

lithium apply metal been to have discovery. material and this expect to We material applying discovery lithium-ion to

everything. we're AI. All-in So is AI going changing on

Our of they But drones. lithium can Science models lithium commercialization, time broader for Safety AI to and the Manufacturing, revenue be UAM metal also AI to ion AI the applications. and accelerating EV, are for and profitability for for applying

of developing ground up technology proud particularly numerous I'm navigated a from Having deemed impossible. industry cycles, that many the

portfolio efforts. with validates world-class collaboration further a Our of diverse customers our

However, I I've never integration been aspect than with I on our this of more AI believe global change a am excited now us business drive operations. will to enable firmly into our transformative about every scale. of the

and in science fortunate truly this am period AI. to I transportation, in exciting be living

we In delivering addition top AI the UAM and talent, to EV the year priorities AI focusing industry. attracting efficiency, vision top to drone leading and our progress on solutions, battery cells for the capital of continuing for partners transformation the our have lithium outlined metal on our X remain make to

like your continued interest over to for now I'd to And for SES it Jing financials. you AI. Thank turn in

Jing Nealis

the operating will second our were operating Qichao. and our quarter Thank year I expenses cover second XXXX. full XXXX our $XX.X In Today, for the quarter, discuss results financial capital GAAP million. budgets you, and

in ended and million in $XXX.X expenditures quarter capital used the second were liquidity. We million. was with million, Cash $XX.X operations $X.X

cash management As we to prudent updated with guidance. XXXX our expenditures, we full our continue and of year be very

cash of to comprised now range of We million to to in cash from to This total $XXX $XXX down $XXX expect of usage million compared $XX the from operations usage range with $XXX million, $XX be million previously. previously. to $XXX million million million $XX is million

$XX million $XX million And to compared with million to million $XX $XX in previously. capital of the expenditures range

XXXX. into our strong liquidity provide well the expect sheet We for to balance company

Going to potential drones our UAM share OEM capital our C-sample upside beyond, we expect partners, expenditures with and capacity forward in solutions could to buildup AI commercialization. earlier and provide

operator to for With that, I'll questions. the up hand back the to open call

Operator

Jed [Operator first is question Instructions] with William Our from Blair. Dorsheime

Mark Shooter

these really I'd incremental been to hear All-in approach. for You in you've Shooter on the on I time. AI positive what Jed. data some on for to this know AI really Qichao, But like Mark seen to what push strategy push applications for you've was this incrementally so working shift. here have background

Qichao Hu

really Safety started we working all AI in AI back for the AIs. areas, end X A-sample, really So of on of B-sample. X really, these beginning Manufacturing XXXX in for then and towards

and of get with actually checkpoints we end A-sample line and more manufacturing, per X,XXX towards about B-sample really cells helpful. then about hit we as per cell month begin per So we quality data, once in now found and X,XXX the you is

And engineers with experience. then human because the don't -- you design, the have new when we cell start cell make design, so new no you basically, a

specs You to have what use. no idea quality

process year AI of first, started very and seeing year. this, we last say, then applying actually traditional would the So the started end and then the And this I we really slow. the just of would then we this beginning And too collect models, specs, recommend all data. model interesting quality and is towards

you optimal the the or the is to tell you for capital pressure you role the it's X sealing engineers a of all electric hot anode is between of rank on quality to human this bigger going lots actually millimeters. grams than one, between Actually, relationships going example, and your so GAP figuring is this hour issues amp AI X.X amount temperature. And instead tell experiments So optimal the steps. and model these has actually, all okay, what then then then per the out, during impact press for jelly trying

shocking, like way. a powerful was that but So really in

it's engineers. traditional powerful. helpful lithium-ion engineers model the quality a doesn't it good instead way sorry, the big to quality, industries, We manufacturing So engineers. this really of like of -- quality out still human the but supplement complement was the improving tool have of replace still from just And world,

language at much sentence. if the And safety actually, side, very curve, and a large a and on we charge it's And you like cycling charge training the all the so data started actually model discharge then with look discharge.

And we're then examples also in for OEMs then So against conditions, an those SOH and language alarm. where any And set and is the dozen SOC, -- -- model. and we B-sample. get also would you triggered, than physics-based have traditional, another then testing a drones. like the several where if you case are mission train boundary in profiles now as sometimes X, have a we models this then more large And of UAM have those

those mature But a process new like existing long to chemistries. of specs of time set you it of [ models is works that. process let's head -- physics-based takes works don't work, to Again, the manufacturing, don't but physics-based work. That I the ] of just develop [indiscernible], existing chemistry, a none existing none the only when introduce that mean for conditions metrics models, set quality the boundary manufacturing the cell

the too continue you take process, if will to long. it So use traditional

models And actually this on an model, AI today, about this one of we that cycle XX anything large then a cycles had which sent had it's actually we model, predict feature, But XXX-something. explain that. this cannot cell one other So on found example where large XXX, incident before. an model language language the before that warning actually cycle like then actually, none able physic-based to the was

So that's really powerful.

And a quality your both cell then safety then so work side, when new anymore. introduce manufacturing design, and experience doesn't -- you

Your metrics work existing of don't set anymore.

electrolyte will So actually new actually last that of our X testing. hired AI just for year, faster. and then help human expanded we're then was -- actually team find we our And actually since you of much in Science, AI end team. molecules, model able to them we and both scientist standardize model we so XX and team, AI And develop AI then

that out then the only scientists this so find performance since XXth, right? the Xth, as XX just works better. after If math to we're math came we XX can Xth, that to molecules the the to And years as confident in that the XX something with are XXXX. heavy past human far the And good XX the we pretty

year, beginning the why just the I to spend much signals okay, improving doing use So towards battery when that of just do Why introduce at want found convinced, X think you XX a this the AI quality, you new improving faster. these we're year end safety B-sample, and then and years years? last of spend things we that made scale, C-sample, if chemistry us X can the

Mark Shooter

great. software we XXx sound that's of It Well, appreciate and making AI Qichao. a your color But engineers, AI engineers make engineers. to there, XXx time sounds -- the control it lot your all I is hear quality that scientists engineers. applying you're like material

to So hear. that's great

Science AI OEM follow-up One thinking I occur. you an -- AI are out for they sorry, a the they what that the for it convince in are electrolyte for partners, currently the to about comes the interested Science, to -- comfortable the with not of they needs are space attractive how sorry, necessary I'm some safety OEM are was traditional a in particularly specifically they're partners. at vast Manufacturing there. at that a as looking of looking proof agree they because that of to How Science this Safety, new looking at is of Or not as chemistry that lithium-ion? and it EV point partners? mapping with. don't OEM have I'm the I bonus such

Qichao Hu

Yes.

R&D kinds talking with it's and C-sample, at a lithium of anymore, cars We're A-sample. to in new field thousands convince all with chemistry So talking necessary of of not for batteries the is them UAM. metal commercial putting tens a things. We're battery we're not and EVs One about it's approach about scale. on X users seriously doing B-sample of

now because field. not safety we lot experience in talking lot AI need but and we're lab in also data, safety real-world guarantee of to the So a point, this safety really the of at model a about

is necessity. thing one So

batteries. lot their automakers a want to own of these is make Second

-- hands battery And the power in the so the LGs CATLs, they so is the of the far, manufacturers, are of far, the large world.

their for having then access and and is to battery control control destiny, own automakers powerful. and the to quickly they having performance need to in vehicle battery So the manufacturing very data data control really battery

same to automakers compared the manufacturers. get to battery allows safety quickly and get in of up the level to to terms then large It and quality the proficiency speed manufacturing of

And for also both metal, are next-gen for any X lithium-ion. for important then OEMs. these the but lithium really So it's

Operator

Instructions] Water [Operator Severson with next Research. is from Shawn Our question Tower

Shawn Severson

clear trying understand And able expand what mean I the the lithium Qichao, does model wanted as the own -- to your you applications right, outside to better monetization this you've a far I'm make to to for you I been monetized the industry, for metal it's the just is OEMs, were battery, and as simply back go talking to the AI. how think I that pathway of have. information in does lithium-ion manufacturing. about How AI? uses of get what with

Qichao Hu

Yes.

as AI So more get both get lithium with models the and get And data Safety, the AI you our very becomes we diverse, becomes. more Manufacturing data we for train our other ion then once larger from in metal partners, model, then lithium than that sources model the this cells train it metal and lithium AI, as smarter the the XX,XXX from actually OEM well in-house we the size for we chemistry-agnostic. that do data,

stack to the specs apply that could this, design new lithium-ion for to can cell process, also apply don't new, pallet to So also be because the battery. commercialize quality then and to And can we know line model Manufacturing, happens so AI they what manufacturing cells. it We silicon apply it's this company example, a for are this also wants that. who to apply because cell next-gen we they to manufacturing also that

for cell to vehicle and life train proven And in predict monitor put lithium-ion data less the incident, then this remaining model if health then the also, cell the wants OEM silicon lithium-ion any that. and and to anode a also can also then the or an be the large used language on useful

Shawn Severson

of you you from pay you, So this solution in the the they effect, kind license AI. they or would, for for license

Qichao Hu

Yes. Yes.

to And vehicle be Safety, our it'll be So could then fine-tune of data year. part Manufacturing, them. could example, first premium over so they the or And month model then engagement model. it per for is line day model basically a for per that first it the be to we once fine-tuned, the the for provide also XX-year per fee phase period. free, And a AI in us for warranty our license and -- AI per for then

Shawn Severson

for a to science. Do to license would from then used you things with own come you are you will OEM the that this, a those will these new for own and and I the the that mean specifically you patent combination and IP by comes going solution expect or own new particularly be up chemistry, it? they things? Or they in the

Qichao Hu

Yes.

also AI own. we models might models, case, Molecular the faster. And models -- Universe, open we can develop -- open be smarter. then others that definitely so in actually when in will the Science model plan be database become so so it the model some that the especially the source to the we then So this for and But property can source can then model cases, make the open part trained source. we of molecule And the

efficiency can we of ion. that has, improve going molecules, on low-temperature once new lithium model fast a we're course, then lithium our coulombic then the for and of example, silicon molecule proprietary or license. will higher use Then IP that that metal those generate But be output, charge to

Shawn Severson

And battery the will mean and and solve AI fact? be there's by after I say, for reactive. that the a My predictive are that, understand there as manufacturers well. is what occurring. right? last is take a of that experiencing solve portion problems question -- can then Can is that this But you OEMs data that proactive Something let's problem you and is is happening, the I

Qichao Hu

Yes.

then the for trained then the is meaning spec. to quality data it's just get going So cells, rank and collect And actually actually we all then manufacture specs. And blindly to for recommend sure, cells, collect can manufacture data in it. AI manufacturing, example, without going any and you quality initial

impact quality higher For steps have certain will example, steps. than other on

for #XX, and a lower going And quality then that Yes. -- step you, pressure the that's you to the so need tell that's to press to example, improve hot

So for then definitely, fix. to get will the manufacturing Manufacturing, point a with AI you can in blind tell to AI where start where can and you you

for is so So then the monitor not health the Safety AI on then it. goal predict to vehicles, -- in yes, but the to and then really control and

we're what them. OEM. So whatever to we the that then the prediction make, send to And do going with OEMs back the to signal that's up

Operator

We have no questions. further

So team all call any participation. I'll pass concludes the That back Thank the call. your remarks. for closing to management today's for you

disconnect now may You your lines.

SES AI (SES) 29 Jul 242024 Q2 Earnings call transcript