or in which finance Thanks, XXX% operating free in flow billion Dave. equipment $XX.X $XX.X billion in billion trailing up year-over-year period comparable XX-month up revenue, year-over-year the income, leases $X.X billion Today, is billion $XXX.X up cash XX% for versus year. last we're and adjusted reporting
encouraged the lowering our serve, our start experiences for in to and future cost stores progress by we're I'll growth. investing continue customer our business. with We be improving to making
fulfillment in-stock Our more single in move expected. fulfillment from earlier U.S. year shorter fulfillment of most regions This optimistically fulfillment items delivery significant eight a we get smoothly you Regional changes levels network represented the more higher with see made to this distinct forms. optimize And network than connections and local the touches and history. in and fewer impact national has in stations the gone our change clusters can customers. between distances to many mean our benefits to to and one centers
last travel frequently Shorter going And to conveniently and A customers deliver When how for everyday that important are touches distances most prime getting the behavior. history. we we're their needs. XX-year customers customer they're as good shipments But perhaps our speed to and are Amazon, growth fastest for of consider fewer to pace getting shopping about items importantly, serve. cost know is to and essentials. mean us in shorter of touches now I in more satisfaction quickly significant delivery consumables they seeing is quarter, buying and distances faster. are from example as more mean customers lower remain and as speeds on talked We delivery the their fewer
few we've reevaluated the last As of year. fulfillment the our part shared quarters, network every we've last over
all we placement we the improvements continue results, make change. benefits regionalization enable into don't but fine-tuning substantial to consolidation rearchitecture realize could increase and the fully steady our to in first further to cost to even We identified and a on speed and we our have The that the We've the fewer believe like impact think substantial centered yet fulfillment also on algorithms serve inbound obviously significant shipments. in-region changes of processes several more to delivery.
long improve team cost to before a inventing way I'm speed. humming out being together. of this, of really have the is proud We executing ideas The and way on they're and and
attenuate $XX to is grew of and generative billion. in in to continued elevated workloads. optimization transition the QX. AWS in And rate $XXX to Moving a revenue new annualized still revenue it's year-over-year we with relative stabilize million and revenue cost AWS now quarter-over-quarter our as AI. to while rate growth deploying saw companies net investments incremental XX% year-over-year run more ago, QX, of AWS' continued to year
XXXX more moved of new and deals to in the by last deals deals. an complete pick pace strong and in we're months have seeing couple we're slowly of encouraged economy volume up, uncertain of the closed But Companies signed.
always like momentum that For signed of the in Top of seeing. won't we're show over higher the recent signings QX, perspective, up date of several October deal new for revenue for mind with the number reported any our AI. in collection happens lumpy to all companies which we generative years, is effective several deal GAAP are for than in but reported an deals total Deal volume be most September QX. and continues but we
silicon working and with X inference respectively. the we've predictions. to produce macro on language compute we've and models this A each which key And custom in last investing. large very we LLM or there. compute having as The Tranium is influential updates which inside lowest and mentioned to chips, layers chip AI is layer our think generative of quarter, I inferences is training of each the shared, large train we're as for the about few or been As it,
provider. trade and chose future as leading build, trainium And Recently, to we the cloud announced deploy we'll LLM maker use anthropic training its inferential and primary LLMs. AWS
accelerate the inferential advantages that Meta's technology. part We to which introduced believe the the where and providers availability for Anthropic to generalized AI training run we coherent AI, from AWS data, Bedrock the general the take time anthropics, those that imminent In think Amazon's language a called inferential Amazon data and As fully have their rest deliver own access well which same access without service. models made managed helpful it's like of their available the LLM of be performance back third-party access partnership, stability and we will models, X In through layer, customers of XX collaborate all the security, future this own price on this through for training and that service, Titan, customers. leading collaboration in continuing will as as as couple offers leaking from we've Bedrock, recently as to them middle addition of the Llama customers control features LLMs large AWS of a being can customize into model first using of with they LLM service. to applications months, managed last announced in a the development but
change. the want agents easiest these job. this opinion, our In right The Bedrock way sizes to enabling move between added the days features, further. expanded out what to automating Anthropic, rate availability models bedrock to answering this including generative purposes pick of fine-tune companies latency customer use get and workflows. cost desire. Bedrock customization right gain the Bedrock to to they characteristics our be is buoyed future create for fluidity, very to technology. changer which generative be to with only model to to still real several which the or trying access there with through exclusive with sizes will of through enterprise-ready developers anthropic the scale new customers which In model is AI, and use that has they AI them been new programmed ability And game of for to the accomplish get Also they will applications collaboration positive to questions and value a build helps model tool compelling use, access is that to early and has and models. unique tasks and can to reaction models allowing are experiment learning and and companies them for Bedrock should continue customers a like models types ability general they rapidly the high the which certainty early and with features the
run their knows customers' and more The Whisper are its companies about another been new code these And engineer that proprietary of the applications coding coding code which asks on generative Code Whisper trained the are familiar bases what and with request per its AI for AWS's companions companions there launch That's #X a long-time well. want with its of open customers got in is wanting of code, LLMs, just few has not equivalent kind Amazon a to the top senior a forum gotten current comments be early traction capability. lot it. who having generative just the of source out work. powerful base launched, recently first excited customization companion companions AI layer, has the Code code enterprise and lot
as can we're done experiences sizes And our their other rate but As can that technology it to much models approach cloud worth around. resides technical customers and the expensive doing transform all data tell, and and rapid AWS you resonating they be focused businesses, It's with always and segment customer a We're the abilities. bring democratizing customers. the that at clear to can and delivering market want for and on complex for also innovating customers customers, taking data, in remembering we've that leader of not is infrastructure. what in of way
companies including Bridgewater, United growing and Nexus, and of quickly, Royal Adidas, name AWS building few. a generative and Merck, number Lexus Philips substantial Clarient, apps Booking.com, The Airlines, GoDaddy, is AI very
who a also are go models AWS, with future generative Conference. start-ups team AI capabilities upcoming new of running its AI to trainium its at perplexity with Inferentia. seeing the AWS all in AWS lot like including And to chose share with Reinvent customers has success in We and
are many examples last various products of to transportation There this use to include, call, drivers AI their all optimal customer generative forecast help on the significant generative We transform business, in to our for experiences. stores to for access mile to and to AI using me our routes make models too in AWS, to businesses need a we're needed generative people better working information employ. AI name more decisions. on applications few we inventory Beyond but our discover are derive want to easily locations they
that for product will tool, in do our where is LLM help customers discover well also images And image We're and the all much the launched of upload a a easier from herself, being much getting phrases. to to early models create entering AI less previewed generation more unique generative just version quickly rest. new to and multiple to to once version the the products it without love. AI a we more they create to and use description at new product In need much advertising, third-party making Apart sellers by having expansive natural we brands Alexa, conversational requests information Alexa's capabilities ability conversational pages photo as intelligent include a more and this. built as make specific requests of lifestyle
the to We're a winners building we're We what's to good one the Every lot be that to has convicted of a have personal applications long-term one in to chance and just newer encouraged arena. world's and our vision generative by also of possible AI businesses best the come. progress our continue making compelling a we initiatives, name this of change being that customers, assistant the Alexa viable few. is one in for more be to is a and
in Prime. often of seeing Video. of be to up continues Video where for Prime signing we're one prime what of top with pleased customers X part the it's an the Prime proposition integral We're value drivers
and video We the and can be but BET+ in Prime as our premium as offering, large marketplace who also in content selection. have content for also for compelling as with of our selection profitable channels channels customers exclusive streaming like business to MAX, anywhere right increasing subscribe its a MGM+, including video-on-demand Paramount+, well offer conviction to transaction own invest we broad continue that members best prime video
investments progress that that XXXX, linear ads still meaningfully $X.XX on early consumers their for per providers. ahead, and number our in and limited As a and If is month we There serve add e-commerce sellers seeing but additional excited more beginning will a movies to than on future Prime expand of customers we TV We for missions. work video of aim plan about include lot continue invest to U.S. have be shows free offer in innovation We're we're in to to compelling our other members. our TV Video. done option, fewer an content, streaming prefer ability in advertisements. to Prime and
and chain across sellers of with announced response bulk, manage fast without up on integrate directly from sellers synced delivery offer automatically receive sellers emerging their payments inventory positive continue to Amazon account strong to recently sellers and by consumers about comp. supply where from easier transportation, stores Shopify for Amazon their We the And our seeing Prime to can pick across and businesses are improve to around the Prime sellers supply handle with very with and set and ship having to A chain. sales promotions place. Our options Buy and Buy inventory it and Both enables clearance manufacturing in in Amazon trajectory. third-party early chain about profitability their facilities with channels inventory Amazon world, excited store they the with Amazon. to supply international direct-to-consumer or automated Shopify customer merchants for one pricing services, customers, manage customs members capability fully worry replenishment making we're same experiences it Prime other borders, ground websites to deliver which and all their managing
and are task of people care has offer to a get feedback. care of Shield for built provide Pharmacy to few customer convinced experience. access to care pharmacists. million health members, can month, their is launch earth aims its $X meaningfully customers a care the free fast, remain making and the supply prescription orbit in of providing health Our solution and Amazon X.X pharmacy the model satellites. reduce communities to for low to diabetes affordable team for more access. the We California better step and delivery two our around per in first-of-its-kind underserved customers prototype RX responding easier satellite last experience unlimited fast Project of weeks that medications purchasing products evolved We forward meaningful we be of which part bring and and to insulin make broadband behavior cost with eligible to Kiper, with XX/X medications world successful The get qualitative to to a of this continuing initiative year, And health both for took that affordable the Blue customers significantly customer to partnered
and mission multi-month space test network year. this our this use will to of We satellites and the data ahead of later collect start planned production from satellite
to up teams our most who significant company. our two like thanking gearing are around close the the of across by for I'd world events
what we've to an on is annual our share builders November and learn capabilities side, that another. team Reinvent kickoff and was for items And to season. with successful $X big shopping of month our be store lot more with most saving a event First, off AWS our on members kicked Prime conference Prime customers, October XX. connect held new millions days earlier than begins deal provide will with sold. holiday the The of this already excited across billion one opportunities holiday hundreds ever, of array XXth
this it's easier the us more Just long, where year over to time day, do I'll and it to we important season. make busy that, and on we holiday our lives there's no mission deliver all every that better Brian. turn shopping customers' and during With we as aim to
