Unknown

build applications faster contining to run with our large Mission critical build applications faster contining to run with our large Mission critical systems that support a lot of the world's payments to provide a topnotch quality of service towards our clients the relationship with Oracle has been hugely impactful Oracle is is very hard to beat in terms of price and performance they were uh quite flexible with us to set up an infrastructure that worked for us we were just pretty much mesmerized by CPUs memory storage seamless connectivity rapid time to develop rapid time to Market training jobs

that took 5 days to to run in the past now only took a couple of hours it's just enabling us to do more and faster we've been able to introduce things like machine learning and AI it is literally gamechanging the amount of good and the change that we can produce by giving that tool to our technology opens up totally new product experiences if you really want to focus on the important business problems Oracle is the way for you to go Oracle really gave us exactly the product that we [Music] needed please welcome to the stage

clay m [Music] work thank you all for coming very good to see everyone here today now look let's be honest with our eles um for those of us in person we're on the west coast you can't seem to breathe outside uh and if you have to choose between listening to me in about an hour and going to get lunch while I'd like to tell you that I'm confident in my own attractiveness we know you're going to choose lunch so instead what we're going to do is we're going to keep this tight we're going to have

a lot of fun uh and what I'm going to need from you is that I have several really interesting people coming out here today when they come out I need you to show your enthusiasm for them and in return I promise to get you out of here on time so okay thank you okay that's it's remember we're on a time crunch okay uh we're here today to talk about Cloud infrastructure transformation and the way that I want to start about that is by talking about a different industry that's gone through a lot of transformation of

its own take a moment to look at this video when I look at it I see a very interesting combination of both modern and Antiquated the ship looks very modern this is actually only recorded in 1950s it's not very long ago a few years before we landed on the moon what is it about this video that seems so familiar yet looks so wrong it's the way that we're loading the cargo we live in a world today post the Intermodal container this was an idea from a single person that has transformed an entire industry in some

ways it seems simple it's an opaque locking standardized steel box uh you can connect them together kind of like an erector set but when you take a moment to think about everything that's been enabled by the Intermodal container the fact that over the last 70 years we've had a 240 times Improvement in cargo capacity from less than 100 containers on a ship to more than 24,000 containers today it also enables you to scale down the fact that you can take one of those containers off of a ship or off of a rail car and put

it onto the back of a truck and deliver it anywhere in the world is incredibly amazing at the same time we've also achieved more than a 97% reduction in the cost to transport cardo previously it was about $6 per ton to transport stuff now it's less than 16 cents today security has been simplified in a variety of ways first it kind of seems obvious you can't see inside this container previously you just had cargo laying around it was in bags or barrels or crates much easier to see what what was going on in there you

can't get inside these containers because they have doors and you can lock the doors and the thing that may not be obvious but is also equally important it used to take up to 3 weeks to load and unload these uh ships well suddenly now in the span of a few hours you can do the same work so the time duration at which the cargo is actually at risk is significantly reduced so we talk about these nice things but how does this end up benefiting Society well um we live in a world where ubiquity has become

commonplace think about coffee I'm not a coffee drinker but uh there's a very good friend of mine that tells me that that she's so into coffee that when she drinks her tea she gets a shot of espresso inside of it that's that's how uh ubiquitous this this fluid and substance has become it's it's only possible by the fact that we have this incredibly marvelous modern transportation system the same thing applies to choice if you go back not that long ago it wasn't common to be able to have an infinite array of choices that you could

choose from today there are options that are much better suited to your individual needs you no longer need to make uh make a poor decision from a few suboptimal choices to achieve these effects it requires more than just the IDE of the Intermodal container it requires more than you actually going out and creating these containers so you for you to be able to see these significant benefits to achieve this manufacturing and warehousing had to be reimagined suddenly you need new Transportation Vehicles you needed cargo Airplanes new types of trucks to cart it around rail networks

and Rail Yards had to be adapted and our road Network had to be upgraded to accommodate larger vehicles that weighed a lot more the way in which people work is also significantly changed previously you spent most of your time loading and unloading things from ships or rail cars or trucks very manually and today you instead have highly skilled operators that are operating incredibly Capital intensive equipment uh to be able to do the same job so all of this transformation required significant investments in energy and capital as well as across time it didn't happen over a

few years it took decades container ports had to be made deeper and larger to support these container ships you also required significantly more staging space for loading and unloading these containers of the 4,700 ports around the world today 20% of those ports only unload and load containers there are more than 45,000 airports around the world today about 20 times as many as we had in 1940 cargo airplanes allow you to have low latency access from almost any location to any other location around the world the rail network has also expanded significantly with today we us

having more than 1.3 million kilometers of rail race rail is still critically important because it offers the lowest cost land transport available and then the roads they operate as that last mile of our Global delivery Network this is all made possible by the fact that we have semi-rs that can take containers from a ship or a plane or a rail car and deliver that container exactly where it needs to be and remember throughout this journey that same container with the same security and the same efficiency moves seamlessly from one type of transport to another so

we're a few minutes in uh and I haven't used any of the AI bingo words you were expecting I haven't actually talked about computers at all uh for those of you that were around then it's also not 2013 and I'm not here to try to analogize between Linux and Docker containers and how you should deploy your software the reason that I'm here talking to you about the transportation industry is because if you can understand what it looks like for an industry to be truly transformed we can recognize how early on we are in our transformation

of the cloud industry today the cloud has an amazing Vision but if you think about it most Cloud providers have a relatively few number of regions it's not in many locations um the cloud cannot efficiently scale down security is becoming more and more difficult and these large uh public Cloud regions these Mega regions are becoming bigger attack vectors around the world so for any of you that have ever worked behind the scenes when you take a look behind the way clouds work building and operating the cloud actually has very few of the benefits of being

able to run on the cloud scaling down is impossible and scaling out is extremely difficult the cloud is not ubiquitous and customers don't actually have a lot of choice the good thing is that we're still very early in our Cloud transformation and we just need to continue this transformation to achieve these same benefits with Oracle Cloud infrastructure we are working every day to ensure that our cloud is available where you need it and you have choice and how you procure it and who you consume it through that's the purpose of the rest of our conversation

today is to explain to you how we're continuing to push forward with that transformation in the cloud so our first responsibility is to have public Cloud regions everywhere today we have the largest number of live and planned regions of any cloud provider with 59 uh 59 uh planned and live and we're continuing to grow more and more regions every year so to kind of illustrate why this is so important about having public Cloud regions where they need to be I want to talk about another industry and bring up a very important uh customer and partner

of ours that industry that's going through significant change is the animation industry if you think about how animation is done from 100 years ago to now technology has changed every aspect of how we work and to be able to use the cloud uh fundamentally anim animators need the cloud to be every everywhere wherever the artists are so up next I'd like to talk to one of our good friends at Sky dance animation here we go these are extraordinary individuals you have our attention you part of something bigger now that is what being part of a

crew is all about good right so what happens now we get to work please welcome to the stage Gregory [Music] brenty hey hey hey than good to see you you you look better than I was [Music] expecting all right so just so you all understand here what Greg does is way cooler than what I do in Cloud infrastructure which is why he's here it's uh through osmosis I think I became at least twice as cool as I was before um thank you for coming Greg oh thank you yeah so sky dance is all about pushing

the boundaries of what's possible in animation uh can you share a little bit with us about what you're doing to level up the quality of your Productions and what this means for your audiences and the industry as a whole yeah for us we need to make technology invisible to our artists they're they're skilled in their craft how to animate how to tell stories how to make amazing films and so for us the most important thing is to take technology out from from their day-to-day environment out of out of their working uh requirements we do this

with uh a lot of hard work from our technologist to make it invisible um but uh connecting our three Studios we have one in uh in Madrid one in Stanford Connecticut one in Los Angeles we have to connect those teams together allow them to to collaborate allow them to share content give each other feedback and move that creative process along um every day and as much as possible we need to keep the technology out of where we are um the big thing for us in compute and and where we're taking advantage of oci is the

back part of the film we have to render it we have to take these uh working models these proxies these animated characters and turn them into the beautiful pictures that you see uh on the on the Silver Screen when we're all done uh to do that we use a lot of high performance compute uh and for us pushing that creative storytelling uh to as late as possible in the process those decisions the ability to burst up in our compute to be able to deliver those final films that's the big advantage of oci and how we

use that Tech to keep the story moving uh we just finished uh our second film called Spellbound that one was rendered in oci so we use the compute resources to get this film done and I actually have a new preview of the film that I can show if if anybody wants to see that yeah let's let's take a look all right oh they're here be cool [Music] [Music] huh excellent service Lawrence I'm giving you five stars you must be princess aliens that's me your having some trouble with your parents yeah they went into the dark

forest and when they came back they were a little different how different well oh wow you ain't kidding her parents are monsters wow I thought you are just being dramatic why would I do that because you're a teenager mom wow monsters ready the Battalion stop this is your king and queen that's the king and queen for the safety of the Kingdom they will be sent to the far reaches of the world if they don't change back soon they'll be monsters forever we have to get them out of here Mom Dad we're going to break the

spell come on let's [Music] go we're going to make it together put down your weapons princess I said drop them what guess we got to be a bit more specific bad King Bat Queen hey chew away sire dad put him [Music] [Music] down Greg I uh I'll be honest with you I find my myself like I see that and you tell me that all of that as well as the entire film was rendered on oci and I sometimes kind of go I don't think that's true I it's like but I remember when we didn't have

any of this stuff so I it's it's really encouraging to see the the the great stuff you've been able to achieve with this technology um so how is the cloud fundamentally changing your strategy for how you can so you can be more creative so believe it or not when these films get started they're actually pretty uh crappy and the stories that we start with are pretty rough and the process unlike liveaction uh film making is we iterate on these on on the script we iterate on the story We iterate on the sequences we piece this

together and then we take a look at it we we uh we screen it probably 10 times through the life of the show these films take about 3 to five years total to complete with the last two years being a big rank ramp up in production each time we have a screening we get notes we get we get feedback about where we need to improve and uh uh the ability to do that the ability to respond to those creative changes that's the key to being able to make a great film um what you saw here

this one's really good Spellbound is going to be a really great film a little while ago Spellbound was a little different than that so we have to be able to respond with the technology to be able to scale up to burst into the cloud to put 10 times more technology resources than we planned for against these films to make it successful I never want technology to be the reason we can't do something we can't address a creative note or a change in say even a character or character design all right Greg last question in your

mind what does the digital animation of the studio uh you know of the future look like for us it's a studio entirely in the cloud uh we're in the middle of a transition right now between uh our collocated data centers our on-prem data centers and a mixture of oci for cloud resources uh our goal by the end of next year is to be entirely in the cloud we need that Global reach we need the to shrink our studio into uh something that takes three different time zones and three different locations and makes it feel like

we're all in the same room that studio in the cloud is our big Target that makes us the the largest animation studio that's done something like this if if goes well by the end of next year we'll be the first to do this um it gives us that flexibility when schedules change or when when shows need a little bit more love before they're ready to go in the can so yeah that's the future for us studio in the cloud hey I we're very excited to go on that Journey with you thank you Greg for coming

and sharing this with us today we really appreciate it awesome thank you [Music] you just heard how oci supports customers with low latency and highs scale requirements and how bringing the public Cloud to them can really uh make a difference uh now what I want to do is take a moment about how uh talk about how oci is transforming the cloud industry by offering different scal down options to give you the flexibility that you need introducing dedicated region 25 the cloud is small as three racks we've departed from the era of custom coachbuilt Solutions this

is the era of standardized modularity where the cloud comes off an assembly line this means you get a fully managed solution with all the services and flexibility to scale it's easier to order with a few simple inputs and since it's just coming off the shelf it's easier to get into your hands easier to locate because the racks don't have to be side by side easier to install with multiple power options we can have it delivered to you in just a few weeks standardization also means it's easier to network with pre-wired cables and an assistant to

verify everything's good to go easier to configure and easier to scale with your needs all the way up to a full-sized Cloud region we've re-engineered the cloud from the inside out bringing you flexibility and scalability like never before oci dedicated region 25 your cloud is [Music] here thank you well look uh so I will say though this is not actually running any of your workloads so you don't have to worry about uh uh our our choices here today impacting anything that's a that uh any actual Live customer workloads look this is something that I've been

working on for a very long time the fact that we can fit an entire Cloud region into this space is something that uh if you haven't been in the industry if you don't it it may not seem that awesome this is an incredible achievement and I'm very proud of everybody on our team and Oracle who's worked on [Applause] this dedicated region is really the evolution of a simple idea that we should be able to scale down as well as able to scale up uh sometimes people have a lack of understanding about why that's so important

so what I wanted to take some time today is to explain both why it's technically hard and what we've been working on to make it possible as well as the value of once you've done that hard work how does this actually enable new use cases that you didn't think about before so first let's talk a little bit about what we did to optim imiz our servers the performance and capacity so first thing it's not just an optimization problem for space it's also for power we had to focus on increasing the amount of cores that we

can provide the memory capacity increasing our Network bandwidth as well as our storage capacity we ultimately we needed extremely efficient servers and so we did that and built it into a complete converged platform such that all of our capacity can run on the same unified Hardware platform uh regardless of which service it is right now we also did a lot of work to improve the security of our racks so normally in the cloud what we do is we have a secure area whether it's a secure cage or a secure data Hall and our racks go

in there what we've done now is we've shrunk that security perimeter down to the scale of each rack each rack locks we have live Telemetry that shows when anything is being done or not done to the rack and we've also controlled the access both front and back with how you can plug cables in ultimately what this means is that um everything that leaves the rack is secure uh and it ends up being that uh with the combination of this and what we're doing on the network side the fact that we have a new scale down

Network architecture the fact that the network architecture within the rack is fully redundant and that all the traffic that leaves the rack is fully encrypted when you combine those things together the Locking rack and the network changes it enables us to place these racks anywhere where you have space and power we don't have to put them together and then secure the perimeter they're already pre-cured within each rack so I've talked a lot about scaling down but customers are not happy if we can scale down the solution but then you also don't have a way to

expand it so the fact is is that start with three racks it can expand up to six but once you decide that you need more than that six rack capacity we have an easy option for you you just add on a medium expansion kit this allows that region to scale up up to 64 racks and then if for some reason you've gotten to a place where 64 racks is not enough capacity for you we have a larger expansion kit what that does is it scales the region to the same network design and overall architecture we

use for our public Cloud regions which means that you can now support thousands of thousands of racks in a single deployment the last thing I want to talk about is the significant Investments that we made in the software stack to support this new version of dedicated region first we've been working for years on our fully automated software installer you probably take for granted the fact that most of the programs that you use have an easy way to install and uninstall them honestly that's not the way that distributed systems typically work and it's definitely not how

Cloud regions have been built across uh you know any of our competitors or throughout time uh instead we developed an entirely new way to deploy and manage resources across these Cloud regions we also made the decision early on to make that all of our services are available in Everywhere by doing so it supports a unified security model and ensures that we don't have any dependency mismatches what that means is that customers receive every service they need everywhere all the time and the last thing that we do is we have a focus on ensuring that our

Individual Services can scale down and that's not something that we do once and forget it every year we're constantly working to improve such as the minimum footprint for a service requires fewer cores less memory and less dis space so big takeaways here dedicated region is a long-term investment for us to achieve a goal of putting an entire cloud in three racks we had to update all elements of the system the compute the network storage and every part of the software stack the ability to scale the footprint down allows us to deliver cloud in locations that

were not previously possible at Oracle we are continually committed to ensuring that we offer our full Cloud capabilities regardless of the size of the region and the fact that we have this flexible sizing actually means that it's really easy to use dedicated region so distributed cloud is continuing to accelerate and you I talked before about the number of our public Cloud regions but when we add up all of the uh dedicated regions and our alloy regions we're quickly approaching the same number of those that we do for our public Cloud a mature industry fundamentally gives

customers choice and where we're at in the cloud today we're not there we're still too mature aloy is our Cloud offering that allows others to become Cloud providers themselves and then to extend our Cloud capabilities with their technology fundamentally Choice gives customers better products over time to talk more about alloy and dedicated region and how it's transforming their business let's hear some of the latest Innovations from Fujitsu Fujitsu aims high as the sustainability transformation partner of choice for customers from diverse Industries in over 100 countries its 124,000 employees work to resolve the greatest challenges facing

humanity and its position as the number one it service provider in Japan and one of the most Innovative companies in the world ensures the best results Fujitsu and Oracle have been partners for over 40 years this strong partnership has supported customer [Music] transformation please welcome to the stage kazushi [Music] Kaa thank you for coming kosan thank you thank you for having you there a it's uh it's very good to see you yep so to start can you tell us about fujitsu's strategy and how alloy Fitz in it first of all thank you very much for

giving us at this great opportunity fit's purpose is to make the world more sustainable by building trust in society through Innovation to achieve this purpose F currently focus on two key things UB and Technology fans is our new business model which solves social issues across Industries globally technology is a at the heart of a sustainable uh tolerance information we focus on five key Technologies AI Computing network data security and converging Technologies for example we commercialized Japan's first quantum computer and provide the fastest super computer hugak which bring us unique Japanese large language models under these

two pillars of urance and Technology our 124,000 employee are working towards a sustainable world oracular oid will be an ideal platform for ubas because o through function are useful for solving social issues while meeting customers demand of sovereignty including AI sovereignty so thank you for sharing that uh I know across fujit so you serve a huge diversity of Industries and use cases what is it about these workloads that makes them so demanding and so complicated good question so as you mentioned f is supporting many Mission critical systems globally including Tokyo Stock Exchange many of these

customers still struggling to move to hypers Scala due to uh they need to have achieve three things simultaneously to utilize the latest Crow technology to keep operational transparency to control their data for example a huge retail customers fail to move to hypers Scala and back to fit own Clow because they need to control the Clow F Cloud strategy is a hybrid it combined with a on premise environment and M Cloud capability such as Azure AWS and F Cloud however hyp scalers don't provide operational transparency and data control F's own cloud does provide these but limited

functions to solve these issues finally this year fit has announced fit Sovereign Cloud powered by oroy which will be launched next ail thank [Applause] you foring crowd provide more than 100 of oci services include generative AI while keeping operational transparency and controlling data because this crowd is run from a fit Data Center and fit to manage this Sovereign Clow this is also because oracular Roy is a oci with full functionality but a smaller independent architecture that is a isolated from a public cloud in addition to existing function the Oracle and phys discussed for six months

and 106 items specific for Japanese Hing requirement and decided to implement include uh this F Sovereign Clow will be supported by or employee who have a Japanese nationality and who live in Japan oracular o is the only solution that provideed this sovereignty and functionality or strong support makes this clowd achieve full compliance with the needs of a Japanese Sovereign crowd which covers government agency and 15 fundamental Industries such as the utilities transportation and financial companies designate from a economic security pro promotion Act of Japan this is the embodiment of a you vision of a distribion

cloud with the oroy that meet the needs of each region we really appreciate uh your support and the or collaboration I well the the partnership has been incredible so far I've really enjoyed working with you and your team that's true for everyone here at orle so thank you very much um what are you most excited about in terms of the future Fujitsu W what excited me most is utilizing this Fujitsu soving cl to achieve a sustainable Society globally with accelerating customers TR Journey not only in Japan as we are planning to deploy F sing Cloud

globally each country has unique Market demands therefore we need to have a barious size of a platform as you mentioned before utilizing oracular alloy finally as technology is a technology company which is aimed to contribute high performance Computing and AI for oci such as our newly developed two nanomer Arm based processor software defined liquid cooling systems both are coming from world fastest con Haku and our experience of 7,000 AI project these will contribute to Oracle thank you for your wonderful partnership uh no thank you kosan um it's been incredible working with you so far and

I'm very excited for us to launch this in April and see where it goes yes I mean look thank you [Music] so as we heard from kosan dedicated region and alloy are one way for customers to receive more choices clouds working together is another way to enable that same choice so I'm going to take a moment and talk about our history of multicloud so uh we've been on this multicloud journey at Oracle for more than five years now it started with our relationship with Microsoft where we created the Oracle and Azure interconnect um we then

over a year ago launched Oracle database at Azure where we took our highly differentiated uh database services and made them very easy to use within within Azure by integrating them in the console integrating them into the procurement experience Etc uh then last quarter we announced a great new partnership with Google where we launched uh many interconnect regions as well as the same Oracle database at Google Cloud uh which as of today uh is now live in four regions around the world and then you might have heard earlier this this week where we announced that we

have a new partnership with Amazon web services in which we're going to do Oracle database at AWS we continue to grow very quickly in our multicloud world with more than 50 multicloud regions live or planned when I talk to customers the demand is extremely high and they're very excited that cloud providers are finally working together to solve their problems so next up I'm excited to sit down with my friend Amit from Google to talk more about our new partnership Oracle and Google Cloud have partnered to enable Oracle Cloud infrastructure to be deployed directly in Google

cloud data centers allowing Oracle databases and applications to be combined with leading Google cloud services like vertex AI for a transformative multicloud [Music] experience please welcome to the stage armit Z [Music] hey clay good to see you it's good to see you thank of course so I'm it right we busy yes we now have four Cloud regions live um since we launched this partnership in June what have you been hearing from customers yeah first uh thank you for having me here and congratulations on successful another successful Cloud world uh I think the partnership which we

did a few months ago when we announced it have been great to work with your engineering teams between our side and your Engineering Group and the progress we made I think customers are really appreciating it right so the customers been speaking to one they really care about multicloud they'd have investments in Oracle uh database Oracle applications they want to figure out where they want to run some of those things or they want to modernize many of them many of those Technologies and they want to partner with other Cloud providers like us uh so we have

a lot of joint common customers we've been working with very closely like uber uh we at uh the world with the CME and others uh they are really keen on like making sure a lot of these Technologies can be available in multicloud environment they can innovate using lot of gen capabilities they can do analytics workloads on top of the data which is coming from Oracle database Oracle applications and build the Next Generation applications as well so a lot of interesting conversations a lot of excitement there yeah we're we're hearing very similar things it's exciting to

see all the possibilities so between the interconnect as well as Oracle database at Google Cloud what do you see as the most exciting possibilities for customers what does it really enable them to do yeah I think few things right one uh given that now you have we have oracles database running inside Google Cloud uh data centers as well as we've been able to now have a common management plane right so the seamlessness where we used to have friction where you had to manage things differently now one common tool makes it very easy same commercial structure

so if you paid for Oracle database before Oracle applications before they can bring that and make it run with or Google Cloud Technologies is also so very good for them similarly if they have Commit structure with Google they can use those commits with the various things they might want to use from Oracle as well so I think the possibilities the endless in terms of what you can do commercially and technically uh applications new applications you can build uh we have Vortex AI as as a foundational piece for customers to build gen applications so if they

want to now take the data from database from Oracle which is pretty common for many of the Enterprise systems and they want to be able to modernize or do some data analysis on top of that or be able to add a lot of agents-based interface they can do that very easily now uh going forward uh the management plane becomes simple the the things are much more seamless going forward as well so that's a great addition for them as well yeah great to hear so look to kind of summarize at the end we've now we've got

you know I think it's 12 or so interconnect regions we've got four uh U Oracle database Google Cloud regions live and we have many more coming soon what should customers do to go about getting started so I think one of things that many more regions are coming we should be probably tripling by end of this year so I think the work which we're doing together is definitely uh great one thing which we' seen also now uh when we talk to customers specifically around the DAT data analytics right so big query for example is heavily used

by many Google Cloud customers and given that they have analytics platform on top that with looker and others they could not easily take the data out of article database and make that part of the analytics platform so that barrier going away are able to now kind of take this this information do a lot more insights and really have a seamless experience is something which we going to see continuously now yeah look uh the partnership so far has been incredible I completely agree with you the engineering teams are working very well together I'm impressed by the

speed at which we're able to roll these things out and the quality of the product that we're we've been using I I've been using it myself it's actually quite impressive thank you for coming here today thank you for the partnership now thank you pleas [Music] it's it's really cool to hear all the different choices that customers are receiving whether it be from Fujitsu whether it be from Google Cloud um I want to talk about another industry that has been undergoing significant transformation uh the way in which we all get around uh is very different than

the way it was uh not that long ago the technology to enable this transformation in personal Transportation requires a lot of scale as the problems that we're trying to solve are fundamentally Global in nature so to talk more about that I'm actually going to give my stage over to a very good friend of mine who's going to share some more about what's going on with them at Uber we all know Uber today for our rides food easier connectivity and movement but the company was originally founded in 2010 to solve a much simpler problem providing access

to Transportation at the touch of a button now 14 years and more than 50 billion trips later Uber is a platform that continues solving more problems around getting us closer to where we want to be changing how people food and things move through cities and opening the world to new possibilities and still at the touch of a [Music] button please welcome to the stage Camron zagari [Applause] [Music] hello uh last year on this stage our CEO darai shared that profitable growth is a top priority for us technology is a huge Factor empowering that uh growth

with seamless experiences for our writers eaters and drivers today our tech empowers for 18 million predictions per second and 1 million trips per hour that's a massive scale so as you all know the basic stories stem from balancing scale and efficiency this is about powering a sustainable growth while controlling cost and not letting cost becom a barrier to growth this is the second year of our partnership with oci and uh basically it's very fair to say that out of every 10 trp that Uber serves for our customers many of you here a few of them

actually are powered by Oracle and O cloud and that's targeted to be doubled next year this partnership has proved that everything is possible especially when we talk about efficient scalable growth uh this is not just what's important for the partnership but also we never let technology become a barrier to seamless experiences uh oci provides a lot of great features for us as as uh such as Ander Arm based systems moving data across regions and these are key factors to making those seamless experiences happen the can do Spirit tops it all basically every time we want

to do something that seems very impossible our friends at Oracle clay definitely himself makes it possible for us and we are very grateful for that it's been a great journey so far and we are looking forward to do more later thank you so much [Music] [Applause] [Music] appreciate AI is influencing every aspect of Our Lives training these AI models requires the largest and highest performance GPU clusters creating and operating the largest GPU clusters is extremely difficult it requires research and development in GPU Hardware networking storage and data center power and cooling oci superclusters deliver the

highest performance largest GPU clusters in the cloud to hundreds of customers today our newest superclusters support more than 131,000 gpus all working together to train and deliver the latest in AI capabilities oci superclusters combine advancements and liquid cooling power generation RDMA networks and file storage services to support the most demanding AI workloads [Music] while I was back there I was able to get a time check we're still on track for those of you that were worried uh I promise we might even in early uh so don't get too excited um we've had immense needs for

compute capacity and storage capacity for a long time uh think about different workloads like search or social media or high performance Computing uh but there are different requirements for these AI training workloads and the different requirements the name of supercluster it's It's the clustering uh clusters require everything to work together on the problem at the same time that requires things to be physically close to each other it also requires extremely low latency and high Bend with access to all components of that workload those requirements create Downstream needs that then flow through every aspect of the

infrastructure you have to rethink your networking your power generation your Cooling how you do Hardware design as well as your overall availability plans so I'm going to spend some time now to walk you through some of the changes that we're having to do at the infrastructure level to support the world's largest GPU clusters so take a minute to look at the numbers on this slide um I find it hard to grasp and I spend a lot of my time thinking about networking so uh for those of you who don't know what a cadrillion is a

quadrillion is a thousand trillion so inside one of these networks there are 300 quadrillion packets per second that can be sent and with an acceptable drop rate of zero one of these networks is not just a little bit harder than a small Network to create it's exponentially more difficult we would only do something like this for a technology that is valuable as AI every single piece of this cluster has to be working together simultaneously or else the workload fails to function gpus are driving a significant shift and how data centers are cooled until recently a

typical rack would be on the order of 10 to 20 kilow uh it's air cooled fans blow hot air out of the server it goes into the data center that all gets blown uh outside gets cooled comes back in is cold air now you have gpus where each individual GPU is a th000 Watts or more suddenly you can't cool that with air you have to have liquid that goes direct to the chip suddenly you have dense racks where the rack itself might have so many gpus that the power envelope is no longer 10 to 20

kilow you're talking 70 100 or 120 kilowatts that means now that you not only have to get more cooling liquid into the data center you have to get it all the way to the chips inside each rack what a data center looks like like is also changing think about this you need far less space and a lot more power data centers for AI are smaller they require both air and liquid cooling and they have a higher power density than even the most state- ofth art data centers we were building for non- aai workloads recently customers

choose oci for their AI clusters because of our Relentless focus on performance availability sec maturity and best-in-class economics so now let's take a moment to look at some of these things that these clusters can produce when you get it all right grock 2 is xai's Frontier language model with chat coding and reasoning capabilities grock 2 is multimodal and reasons over human language images and generate codes sorry generates code in this demo we are using grock 2 to image understanding to explain details of the first image then text generation is used to generate a packing list

and finally we use image understanding and text generation to summarize what is shown on our boarding pass to Mars all of this capability was trained and created on oci infrastructure transformation takes time often transformation takes much longer than we like but thankfully it's happening even when you aren't watching back to the beginning when I talked about the transportation Logistics industry uh the changes that were required across uh all the different aspects whether it be from containers to shipping to Railways to all of that took decades to accomplish it's still very early in the transformation of

cloud infrastructure many of the benefits that we can achieve are still yet to be realized however through constant Innovation from the smallest dedicated region to the largest super clusters at oci we're bringing Cloud ubiquity and choice to customers all around the world we are still very early I you heard you know kosan talk about uh the new Sovereign Fujitsu Cloud um you can hear Greg from Sky Dan talk about the transformation that we're making across the animation industry by having a large number of regions very close to things like artists I want everyone here to

walk away with an understanding that the cloud has a lot of benefits and we've got to fundamentally transform how the cloud is built operated scaled delivered so that all customers can take advantage of that thank you very much [Music] we hope you've enjoyed this morning's program please join us here this afternoon at 2: p.m. for the industries [Music] [Music] keynote get [Music]