Thank Lee-Lean. you,
respective establishing highly point strategic partnerships I These strengthened vector the and developing the hyperscalers fast in by our individuals who extensive start their search have our earlier roles in addressing wealth skilled want knowledge pivotal fields. in team of we mentioned market. presence play by experience Lee-Lean, a addition to with will bring two with of a and professionals
government related Israel. applications, from some to some team, division including U.S., functions undertaking with of members, software U.S. the our team assume that new based transfer the and in our With will role scientist senior of projects on who require various projects developing will data lead we team the workload this One our will collaboration functions, and employees. offload
benchmarks play platforms. science the U.S. crucial assisting in role conducting compiler and customers with Our a team different across data will
Our search, of new business company. collaborate to the data will the scientists vector paving for with our deployment this successful for line way for plugin this optimize team our fast
wealth Our second engage form leading having of top experience the hyperscalers. strive as companies. industry new the background from resource afforded sector, FPGA will for be him invaluable extensive our with partnerships brings we worked and a semiconductor This which has to connections
the lead platform our APU the explore our development. We strategic APU to platform partners revenue to will sorry, resources he - building to technology develop lead I'm of building future our fund and licensing will for of service
were hyperscaler our for architecture for growth. holds On we added recently we mentioned the resources last we a major Gemini language of team. relationship large models. potential And great inference additional working with this to call, based on This
We a have architecture. future this progress great in study we APU. utilizing GPT exploring conducted feasibility study I say And delighted specifically is to Gemini on focuses a making are inference am that The prospect.
fast memory models technologies the fast a APU way GPT is store existing to achieve levels gigabyte and memory. very that GPT while to process external the large XXX to XXX the required intensive billion memory very from processors can performance it the found hierarchy application, enhanced compared requires parameters. memory working We And a utilizing when technology. the same significantly of the storage internal model, is memory all
high the a on performed large fast requires incorporating very and HBM, as to matrix the next processor memory internal memories. a bandwidth d core. by core be working to large It be X.X several This a multiplication are also internal memory processor which by and we'll processor accomplished memory And the substrate. can put support die
provides built GPUs achieve throughput, as very performance computation. only large inherently memory high not and in architecture node. similar Gemini process large but memory technology memory per that tops same can supports on watt also bandwidth the APU peak has state-of-the-art
and However, our the the large tops size GPU. sustain can tops nearly peak with same APU bandwidth, a average LX massive as unlike
performance than that six generate we than and more three XX plus with point the node. to technology token XXX calculated watt, Gemini slightly a a input size six times X-nanometer XX output per dye per be a dye tokens second more better GPU single This is in by module model. HBM the that state-of-the-art have tokens a delivered of XX context could And a achieve could of GPT composed of billion we
study levels, The was power memory in done peak further low in distinctive the The processing move advantage III conjunction out holds roadmap for capability. the with Gemini territory. laying a at in delivering generative consumption performance given to AI development into APU
like seen, applications As we ChatGPT, each have generation. are more generative with AI capable becoming
is more higher computation, parameters parameters model and energy them. a require leading force to combat reduce of power larger carbon capability trade-offs researchers much requirements. More carbon driving new compress behind large are usage used format's growth, footprint. the language the To to improvement between this that to the are are number These The ways memory by that the footprint investigating. researchers help exploring data
resolution trade-offs and set these solution. GSI Unfortunately, data need Technology's, a APU technology to limited flexibility formats. size be performed methods. navigate to To on flexibility single a level bit. explore can the with new this element small provides CPUs they flexible fixed lack are bid a of By GPUs to and allowing any computation fine as as a the performed data computation be at
transfer. we developed. each energy solutions to by number allow usable bits the optimizing reduce of And be will innovative data This for
of awareness our of strategic biggest work with capabilities partners, potential some challenges. the can we increase we AI's solve licensing to As
Regarding we notable two our SAR work with progress I of on Gemini targets. solution, made our have
these one presence commitment in of our in have FY Underscoring we expanding with sale our closing a a XXXX market, set customers. to this goal of
As in of I mentioned, we a support recently to additional search our SaaS we resources place anticipate building end fast added vector vector fiscal with revenue year. solution fast source beta the the With before resources customized customers. customers search for
or switch compared revenues and In prior the $X.X XX% million XX% Sales or period quarter ago, Kia $X.X the first XX.X% the were product net a new same to net in of me $X.X fiscal now customer of net revenues Let and revenues the or for year to in million of quarter. the to of first XXXX. breakdown quarter. million
prior quarter the compared first in to and quarter of prior fiscal XX.X% first sales XX.X% of XXXX the compared in first shipments a in XX.X% and shipments XX.X% the XX.X% ago SigmaQuad were shipments XX.X% quarter. in defense year of of quarter. shipments comparable the Military were to period quarter of sales
like to now I Doug. Doug? Please the hand ahead. go call over to