GSIT - Earnings call 3 2024 | GSI Technology, Inc.

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

Operator

Ladies and gentlemen, thank you for standing by. Welcome to GSI Technology's Third Quarter Fiscal 2024 Results Conference Call. [Operator Instructions]

Before we begin today's call, the company has requested that I read the following safe harbor statement. The matters discussed in this conference call may include forward-looking statements regarding future events and the future performance of GSI Technology that involve risks and uncertainties that could cause actual results to differ materially from those anticipated. are Securities the and uncertainties filed Form and the with Exchange risks described company's in XX-K These Commission.

at today, Additionally, you I've XX, recorded that advise of XXXX, January this conference Technology. been call is asked the request to GSI being

call and Shu, the to today would President Sales. Didier Chief Lasserre, President company's Schirle, Douglas now Lee-Lean go him ahead, Please turn Shu. I over With Vice Mr. Officer. are is the like Chief the Chairman, and Hosting Executive Financial conference to Officer; of sir.

Lee-Lean Shu

to to joining product Good family. eventful updates several thank us X third advance you afternoon, that track an we development, and us. the for I from Gemini milestones the am pleased keep major share APU quarter. achieved with Starting will key on

the to we confirm tolerant test a The radiation testing November, in-space for [catalystic] on results processor. APU in successfully met the be compute has harbor Gemini-I First, applications. radiation that Gemini-I complete

several companies meet strong We are encouraged we satellite their that on actively radiation APUs, interest. by are and with engaged total

and We of customers the compact accessing offer that starting internal should edge in AI anticipate This has Gemini-II entirely sampling enabling target cost. give XXx Moving models half advantage a deployment a processing used. loadable chip in-place XXXX. performance APU. to spins the in Gemini-II and chip, Gemini-II of us the This cost with much applications. In We many to applications will calendar third allow to quarter. external fit of than expect [indiscernible] like dramatically XXXX. of cannot Gemini-I broader half internal Gemini-II memory on begin generation and can lower Xx performance megabytes memory second of XX from table without second better Zone at memory ADAS. range improve be the May-February, the initial in select The complete in we late our the evaluate calendar in Gemini-II of initial data tremendous

Lastly, programs, third million. quarter, shipped radiation contract SBIR we X to new we hardened $X.X and a amount II in second [issuance] for of the Phase the received in X customers [indiscernible]

ahead, stage with design innovation hyperscalers of for We early their suite late had architecture center. several best our Gemini data in discussion in address options lead chip. about we Looking developing a emerging to are the the the APU markets

have lower our In by data X users significant initiated with with the help, technology center in can consumption, We who on discussion costs. hyperscalers reducing continue APU addition early-stage helping interest also to teams to and power to preliminary Gen expressed AI influence benefit concepts. work provide hyperscalers major end various internal

our to financial performance. Turning

mix million our product guidance gross The quarter. revenue improvement a margin XX% make A reflects gross at the third margin midpoint of quarter $X.X of in of with sequential our guidance. was this

and both Gemini-II Our reach. Gemini-II. operating for as a including mark as sequentially, greatly preproduction market $X.X growth onetime an expenses, increased We this in which future expense for investment APU of year-over-year extend will million view

development. In partner potential with launch suite Gemini-II assist strategic we support parallel, and with Gemini discussing to continue

chip extension partnership to management ambitious prudent we investment our require capital mitigate Gemini significant road to fuel anticipate suite today, and product on financial new lease. most our funding the map As focused remain our markets. and We help will into

the and paper location efficiency University spotlighting discuss APU of details published the of comments. his its our in unique the accelerate research will [indiscernible] map. paper for GSI implication Didier Cornell Recently, Gemini-I a to

our driving prudent focused our market quarter on our and Gemini to In closing, creation. APUs increase growth. tremendous long-term financial expense roadmap, with to laser We was on an innovative it our and progress support eventful flexibility bringing management actions third to value remain

go call the business ahead, our Please discuss Didier, further. performance will hand who over to I'll Didier. Now

Didier Lasserre

on so inference context Thank we especially confident potential you, our the workloads. Lee-Lean. market in for are why I for architecture, provide additional want some APU to

X key, cycle Bit the real by processors I'm real APU bit can -- is cycle, variable to our from that time million bits undefined processing can to our efficiency. adapt flexibility unmatched X of bit to million First, sorry, X time with be in maximize toggled

are This strips models. configurability deal for different X can dynamic shows an as for bit time bit different process inference at bit-wise long a a that research since efficient precisions more

delivers Second, higher by consumption. architecture removing lower fetch power This innovative our model the Von the Neumann performance breaks APU data function. with design

As capabilities data for consumption needs users. reducing centers costs directly power lowering Gen and center applications inference AI critical end these Lee-Lean of address and data mentioned, emerging the by

our our architecture and efficiency gains speed has Importantly, ultimately and I compute in logic difference and chips memory, the compute APU -- in-memory to emphasize near-memory as actually memory. This they architecture. compute physically true sorry, to I'm competing that that are achieve claim scale. we in-memory will fundamental integrated want APU APU true represents Unlike our enable the transformative in anticipate we

memory us architecture sustained and competitive compute true advantage. gives Our a

development, monetize University. recently ecosystem APU focusing hands strategy cases we military, deployment helping use research partners support go-to-market of our and from The and paper will the One of academia. in benefits, accelerate To is and the Gemini-I example expand on capabilities. Cornell hyperscalers real-world promote us as are libraries this in getting published a key and showcase the

our unique APU We APU's to Gemini-I demonstrates Paper for sorry, benefits -- Cornell performance our I'm pleased that are genomic announce the applications.

potential simple APU's needs, medical for budgets faster a advantage our revealed and in-memory performance for multiply lower filtering to packed cost-effective scaling, comparisons. accelerate XX massively data data Xx can that be This deployments into architecture, this to researchers workloads other for our With technologies DNA data-intensive showcases up GPU similar high-density compared can study servers compared search, requiring low-precision Leveraging Cornell showed analysis, also to to sequencing more. strong including APU parallel real-world to applications CPU. rapid core solutions. hyperscalers security with power The matching the

within our We sets. on opportunities to performance remain industries. multiple significant compute market patterns efficiently reinforce across sorry, massive sectors and data on I'm focused game-changing rely -- in-memory delivering enforce results across that These similarities customers the finding

initial devices will this silicon target and we Gemini-II after. evaluation As anticipate summer spin After we shortly February. of in first initiate second a benchmarking debugging, Lee-Lean receiving and mentioned,

million $X.X funding SBIR this will support development. in Our

a As X.X direct to includes Phase SBIR our create second announced contract reminder, Air algorithms Laboratory. Research II recently U.S. Force to million this the specialized for

The target and its GPS memory using U.S. develop object showcase from applications rescue, as space also include APUX in-craft search and data and will SSIM integrated circuit. to indication, and change situations. algorithms the compute benefits moving applications detection GSI such of force target absent performance the detection,

engagements us gives opportunity. market In in confidence summary, partnerships the and versatility of architecture our customer ecosystem hands-on

coming map a can that capitalize have APU road deliver multiple innovations in to years. will we product We believe continuous industries across we robust -- the that adoption

to net or year $X.X or customer were the XX% third net ago, revenues for now third to XX.X% switch of in the me revenues of XXXX, In and breakdown in and compared the $XXX,XXX quarter. quarter of XX.X% period of product million Let to $X.X Nokia sales million revenues prior fiscal or a same the quarter.

sales of fiscal the quarter. prior in period XX.X% defense of to shipments ago XX.X% comparable and quarter. compared and of in third in XX.X% of third the shipments were the SigmaQuad to Military prior the quarter year XX.X% XX.X% a quarter sales shipments shipments XXXX were quarter XX.X% third compared of in

shipped to note, one product prototype third quarter, sales in SRAM a different deployed of radiation-hardened over be satellite customers. we in the X These X programs. $XXX,XXX on last separate On will

I'd over like to call ahead, the hand please. Doug, go Doug. to

Douglas Schirle

on net revenues million $X.XX net net per revs $X.XX of $X.X of of of fiscal diluted of diluted you, quarter Didier. the million of million per of losses the for XXXX. or quarter or diluted loss share of $X.X share compared third $X.X net the million of share Thank million to on revenues We and $X.XX or net third $X.X quarter reported $X.X $X.X in XXXX a XXXX a million fiscal second per fiscal loss for net

Gross in product margin period the primarily second in at The preceding compared changes in the in to periods. XX.X% year the XX.X% changes margin mix to XX.X% and gross X due quarter. were volume sold was in prior

the expenses quarter. the development the prior in compared to the were $X.X expenses million at year and $X.X the million million and third million operating period and $X.X prior to XXXX fiscal quarter prior quarter fiscal XXXX million in million quarter. Research $X were from third Total in $X.X in of compared of $X.X

and in administrative were XX, December $X.X the general expenses prior the $X.X in million ended million and quarter million previous $X Selling, compared quarter. quarter to in the year XXXX,

in $X.X was million Third quarter. compared quarter an million prior and operating million of $X.X loss operating year $X.X loss prior in period the the XXXX to fiscal

and ago. $XX,XXX, interest a a compared and the included Third XXXX income same and interest income of provision other a quarter $XX,XXX provision net year other $XX,XXX loss tax of fiscal net period to of for of tax $XXX,XXX and net

the was XXXX, $XX.X a year short-term investments in tax as interest second capital $XX,XXX. quarter. $XXX,XXX compensation $XXX,XXX debt, and prior compared included March March in loss cash at equity had million December cash XX, million March and $XX.X provision in to XXXX, XX, income XX, stock-based no was compared $XX.X was quarter, at year Working $XX,XXX $XXX,XXX XXXX. fiscal XX, in the December of we the quarter of compared preceding ended stockholders' as million comparable expense equivalents Total versus pretax In December net to the XX, a XXXX. equivalents million quarter ago cash, third cash and net million of to XXXX. With other $XX.X $XX.X and At XX, million of as of $XX.X and

of to to $X.X XXXX, fiscal range gross of quarter anticipate approximately in fourth the of million net million, $X.X XX%. the revenues For with we XX% margin

call like this at to we'd point, the Q&A. to open Operator,

Operator

first Brett Scott. Instructions] with question Janney Montgomery from Reiss [Operator is Our

Brett Reiss

like me you is infrequent Gentlemen, in more give preproduction I'm talk an if what just you clarity an million a on old. engineer, years can for explain some you charge mask? I'm X not that. just $X.X me to can Could

Douglas Schirle

Yes.

wafers. mess time product, the we the a we to fab have in every have to have to manufacture set a So prepared run

charges for it for Now prepaids we mess capitalize production over XX-month that a product, set will typically, when period. we into and amortize incur

that X a the process prepared before, R&D when is run we've that we expensed to new expense. set never to that mass what have on a technology exception However, charge

at last process used. a this out we've quarter It's taped in Gemini-II we on well, quarter, technology process -- So a XX-nanometer TSMC. never

that charge that never $X.X we before, million since Now used expense. we've R&D to process

Brett Reiss

right. All

X is It's this reoccurring? So going to not shot. be

Douglas Schirle

It recur tapes no. product we've that no, process we have in before. the used out technology never will on when the another that future No,

it's so most year, So X every It happen X years. infrequent. doesn't or every it's at

Brett Reiss

Okay.

was keeps cash, dwindling margin my down. of my that. safety a about I'm little The in which concerned investment Okay. here

in we through And how we that? with on the Some bringing do the joint of these burn other remaining long cash? timetable before partners. What's initiatives venture have

Douglas Schirle

Didier mentioned others. To it's to next and going opportunities, Well, product, various currently talking hyperscalers require we're with our looking significant at to come out and investments.

So we're of other or looking funding. for sources partners

quite we near In that, a building building money worth looking addition be in to we'll that own, we of and bit a at the potentially have selling that that's future.

Brett Reiss

you that? do What for? sell Right. the case base is think a what appraise you And could building scenario, on it

Douglas Schirle

We $XX get to think million. we probably the in range million of somewhere can $XX

Brett Reiss

Right.

Douglas Schirle

And tied paid because to fully there's it's that no for. debt

Brett Reiss

of a pay payments right. investment? the milestone capital the with you take? look be outlines What do it Right, broad would to they equity would venture of an an Could structure Would that like? partners' joint of in other would me structure investment think a you follow, just GSIT give upfront payment what

Didier Lasserre

mean either be talking that's could funding I Gemini-III. Lee-Lean Next program, is the So of specifically a those. for about when And what partner for it was

that, So a customer be for a partner. more would likely Gemini-III, of of that that, NRE But But funding we for partner looking for Gemini-II type most fully funded from was internally. funding. aside an are

aside we would investments there So well. that. open But associated be to in from the with are also company milestones as equity that,

Operator

Bernstein next Capital. Our Jeff Silverberg, Bernstein instructions] [operator from with question is

Jeffrey Bernstein

So the question on placing those -- valuation. and parts on just a rad-hard for SRAM congratulations

those were from might be satellite those over about gain you you programs. programs time If win X how to to what period able much and revenue

Didier Lasserre

Sure.

quantity prototype the of again, and a is one was So demonstration this purposes. for -- programs just

was on them $XXX,XXX. was that the ASPs about over are One it quick shipped, you totaled of And one do calculation a parts just what we can XXX,XXX those. of other and XX the

And just again, are quantities. so prototype the, those

quantities that looking next you are by year. have These launch So the to the obviously, don't something. can programs multiply they're within We yet.

But certainly, it so second they soonest can be half north prototype time be might is a of And at before XXX,XXX, imagine the production. would XXXX you what if the be. would release some it production --

Jeffrey Bernstein

And be so GEO these would satellites. like

individual larger going satellite aren't be to These satellites? those So networks. are

Didier Lasserre

actually is So are is GEO, these X X LEO. x, ones

Jeffrey Bernstein

Okay. on to actually ride get that? from still of going these And evaluation providence parts, get sort else? are somewhere Or get we Will you automatic one with space a of and are these to you into that looking get

Didier Lasserre

No.

as spoken other -- already the have we out the in last certainly, we've in years. we've devices of So couple have past, shipped we prototype

in And so it we programs the or the could be past said these any X. of

these like is their accelerated table. One looks of in time fairly

So chance quarter of get X just there's we could a hit ones quickly. fairly last this the up

Jeffrey Bernstein

-- the from story in And sell lowest on outlook what's revenue router just Nokia then and you for that Okay. on them that Nokia forever, into? the

Didier Lasserre

to inventory, manufacturers and because had were so they they bit X inventory. both contract through their a manufacturers parts separate were So contract send of we these some burning little --

in Nokia forecast -- the think for the quarters. coming those we -- As in far rate last seeing are run as couple been we've from I rolling past talked we've around still get the XX-month a and of the

north a of $X.X bit a quarter, the this quarter. $X did So kind of what million somewhere past we just in little range million,

Jeffrey Bernstein

at also mentioned partner, a think I partner, those both And for potentially financial for Okay. in of of a a Needham, development to possibly hybrid a Gemini-III. partner, you addition potentially looking

around sounds some talk an also about I potentially stream? you semiconductor about upfront of you that a memory licensing be just Can probably Would an royalty like give kind license something for for that? just what would in IP. edge then of processing. be It IP think case and sort Or color doing that. mentioned for

Lee-Lean Shu

more focused the Yes. Gemini-III. IP on are we Well, --

be development the product for could it or the IP be customer. could it So

So over have as good I we mean, conference progress basically, we the call, made think I -- in quarter. we pretty mentioned the

it. So still we're yes, working on

Operator

hand to Lee-Lean further There comments. call are to like any time. no questions for back closing at Shu the this I'd

Lee-Lean Shu

results. our full forward Thank speaking report the you year with quarter we and all us. XXXX you We for again joining when to look fourth fiscal

Operator

conference. This concludes today's

may time. lines you your disconnect this You at for your Thank participation.

GSI Technology (GSIT) 25 Jan 242024 Q3 Earnings call transcript