Unknown

[MUSIC PLAYING] SPEAKER: Please, welcome to the stage Larry Ellison. [MUSIC PLAYING] LARRY ELLISON: Hi, everybody. OK, let's see. I took some cold medicine about an hour ago. So if I do sit down on stage, it's not because I tore my Achilles. OK, I'm going to miss Aaron Rodgers this year. Anyway, I'm going to talk about-- I'm not going to talk about the NFL. Generative AI showed up about a year ago now, and it changes everything. It's certainly changing everything at Oracle. OK, about a year ago, OpenAI demonstrated the ChatGPT. And much to the shock of

the people who developed ChatGPT, the baby talks. Seriously, the developers did not expect that. They did not expect. And I'll explain why. OpenAI was a company founded about 10 years ago by Elon Musk. And what they did after, again, about a decade's worth of work is they built this enormous artificial neural network made up of billions of artificial neurons or what AI professionals call a billion parameter model. And no one ever built a model quite this large before. And after they built the neural network, they trained it on enormous amounts of data. It read all

of the Wikipedia. It read the entire public internet. No one had ever trained a model on that much data before. But the artificial neural network itself was not different from the previous one or the one before that. It was the same architecture. The only difference with GPT 3.5 is the neural network was bigger. The training data was more vast. So just because of this change in scale, suddenly, when prompted with questions, ChatGPT supplied answers. It engaged in conversations. It was a big surprise for the people at ChatGPT, for AI professionals, for the entire world. And

it captured our imagination. Unlike most events in technology, most cool new tech does not-- does not get the attention of heads of state and heads of government and every day people in other professions. It just doesn't. This is like Sputnik. Now that shows I'm a little bit older than some of you in the audience. Sputnik, when the Russians first launched the first artificial satellite around the world, it captured our imagination. It was news. Everybody knew about it. Everybody knows about generative AI. It raised the concerns of AI professionals who were concerned about the risks this

new technology poses. It raised concerns about governments who've talked about regulating it. The writers strike not far from here in Hollywood is partially based on the writers being concerned about their jobs. Who's going to write the next great script in Hollywood? Is it going to be some guy you know or some computer you don't? Everyone wants to know what comes next. Well, I'm going to start by telling you what I think comes next, which is a worldwide race to build what comes next, to build better AI, still better AI, which will deliver a better future

and will have significant associated risks like all new technologies from the first technology, fire, which can be misused and spears, which can be misused to later technology like nuclear power, which can be misused. All new technologies can be misused. However, by and large, technology and advances in technology have made our lives better, has made us as human beings more prosperous and more comfortable. Generative AI, is it the most important new computer technology ever? Probably. One thing's for certain we're about to find out because countless billions of dollars are being invested in generative AI and large

language models. ChatGPT 3.5 is no longer the current best thing out of OpenAI. It's ChatGPT 4.0. Just this past-- in the past 12 months, Cohere, ChatGPT, other companies have developed new large language models, which generate not only language but they generate images, they write music, they write poetry, they generate computer code. That's getting a little close to home. And coming soon and very soon in the next 12 months, complete self-driving cars from Tesla. There's a new drug that's entering phase I clinical trials. The drug was developed in China. It's an antiviral drug. It's not a

vaccine. It's a therapeutic that treats COVID-19, all variants of COVID-19, every single one. But it doesn't stop there. It also cures cousins of COVID, SARS and MARS. It's a small molecule designed by a computer that binds to a protein on that virus and prevents it from replicating. I told you AI is going to make our lives better in a lot of ways. There are associated risks. At Oracle, we're working on a voice digital assistant for clinical purposes. So I remember once walking into Stanford to go look at my X-rays. And they couldn't find them. And

they brought in one, two, three specialist-- IT specialists to help the doc find my X-rays. Now, you just say find Larry Ellison's latest X-rays. Oh, by the way, show me where the bone is broken. There are more big surprises coming, lots of them. And most of the news is going to be good. All right, so how is this working out for Oracle? So far, we're pretty happy. It turns out that Oracle's Generation 2 Cloud is very different from other clouds in a number of ways. But one of the most interesting ways is the network we

use to interconnect our computers in the cloud is very different than what the other cloud vendors provide. We provide a remote data memory access network. That means one computer in the network can actually access the memory of another computer without kind of tapping that computer on their shoulder and getting it to interrupt itself. So it has the ability to move a lot of data from one computer to another extremely fast, many times faster than conventional networks. And the fact that our standard network and all of our clouds uses RDMA network means that when we build

a computer for training large language models, that would be a computer made of NVIDIA GPUs, when we interconnect those GPUs, that computer runs much faster in our cloud than it does in other clouds. Well, as I've said before, in the cloud, time is money. If we run twice as fast, it costs half as much. If we run three times faster, it would cost a third as much. We are much faster and many times less expensive than the other clouds for training AI models. That's why NVIDIA is doing AI training in the Oracle Cloud. Cohere is

doing AI training in the Oracle Cloud. Elon Musk's new xAI is doing training in the Oracle Cloud, plus dozens of additional technology leaders and startups are coming to the Oracle Cloud because it's faster, much faster and much cheaper, much more economical to build your AI in the Oracle Cloud. So, so far, so good for Oracle. The NVIDIA superclusters we're building, well, one of them will be the largest scientific computer ever built in the history of the Earth. I remember when I used to keep track of where the world's fastest computer was. Was it in the

United States? Was it in Japan? Where was this computer? Was it built by Cray in the old days or by Intel or who built it? Well, the answer-- the new answer will be NVIDIA and Oracle. And it's a bunch of NVIDIA H100 16,000 of them connected together with our RDMA network. Now, there are two things that make a computer fast when you're dealing with AI. One is how fast are GPUs? How fast do you process the data? The other is how fast can you move the data to the computer to process it? And we do

that much, much faster. And if you're not moving the data fast enough, the processors are just waiting. And our big advantage is we move data better than anybody else because we started using that RDMA network for our database, which specialized in moving-- in moving data around. It so happens that's the same problem you have with GPU clusters, moving data around. It's a problem we solve for database and reused in building AI models. OK, Oracle's been using AI for a very, very long time, many years. But this is different. Generative AI is a revolution. It is

a breakthrough. It's transformational. It's fundamentally changing things at Oracle. I'm going to see if this works. I'm going to bring a chair over. Forgive me. This is great. OK, I'm happy. I like this. Now I don't have to get down on the ground like Aaron Rodgers. All right, so I was thinking about that. My family's watching on the internet. They're going to, you know, like what? I'm going to get calls after this. OK, so we have been using AI for a long time. But this makes AI central to almost everything we're doing. And it fundamentally

changes how we build applications, how we run applications and changes everything. For example, we're not going to be writing new applications anymore in Java. Not new ones. I mean, we're continuing projects that started in Java while ago, and they will go on for a very long time. We're not converting from Java to anything else. We're continuing to use Java. But if we're starting a brand new project, we're generating that code. We're not writing it-- we're not handwriting it anymore. We're generating that code in this thing called APEX. And this application generator is a no code

system. And we have been-- because of the new technology, we have improved APEX dramatically over the last couple of years to the point now where virtually every one of our new applications is going to be generated by APEX. The Autonomous Database, which has got a small percentage of Oracle databases in the world are our new Autonomous Database. That's going to change. It's going to become a much larger percentage. The Oracle database, there's been a long debate about the Oracle database or relational database. Is that the best way to go or you really should use an

object database like Mongo and use conventional-- gets inputs and not worry about this schema thing you have to build? Well, we've solved that problem. We've ended that debate. And I'll explain in a minute. We've added vector capability to our standard database because most people want to take these AI models, what are called foundational models, the ones built by Cohere and OpenAI, xAI and others, they want to take these standard foundational AI models, then they want to specialize them for law or for medicine or for some other field. And they want to use their own training

data to specialize this. And the way you do that, the best way to do that is to put that supplemental training data in an Oracle vector database. And that's what we built because we can move that data into the training machines faster than anybody. And the next thing is our data intelligence-- our analytics platform, what used to be called our analytics platform is now a data intelligence platform because it is looking at the data using AI, and then it itself is the repository for so much data. You're using that data intelligence platform as a source

of data to specialize large language models. I'm going to go into-- I'm going to deal with each one of these. OK, APEX from Java programming to we're generating new applications, Cerner New Millennium generated with APEX, Fusion Marketing generated with APEX, Banking applications-- back up the slide, please. I didn't mean to hit the button. My fault, my bad. But please back it up on. OK, the Cerner New Millennium, Fusion marketing, banking, retail, hospitality, virtually every new project is being built-- is being generated with APEX. That means the teams are dramatically smaller. The development process is dramatic--

is fundamentally different. We prototype features and iterate, so we develop things much more quickly. When we generate an application, guess what? We don't generate security bugs. Every one of our apps that we write has to go through a security audit. Well, not the APEX app because we didn't write it. The AI system wrote it. We generated the app. It doesn't generate security bugs. And it generates an application that is, quote, "stateless" without going into all of that. If the computer it's running on goes down, you can immediately failover to another computer. So it's fault tolerant.

They're much more reliable. So as long as you're willing to spend fewer people and less time building your application, it will be more secure and more reliable and have a better UI and all of those things. This is a very big deal. We've been programming in Java for a long time. The Autonomous Database, again, it's very cool. I mean, it's 100% self-driving database. You don't need a DBA. It installs itself, configures itself, updates itself, patches itself, tunes itself. It does all of that stuff automatically. Still, it was not used by a majority of our customers.

Even inside of Oracle, it wasn't used by Fusion, wasn't used by NetSuite. That's all changing. All of our new applications are starting with the Oracle Autonomous Database. And Fusion and NetSuite are converting to Autonomous Database. Everything we have that's current, everything that we built a while ago, everything that's strategic and on the cloud will be Autonomous Database. It's because it's so much more secure. If you don't have human labor, yeah, you save money, but all the security errors, almost all the security problems are caused by human error. And if there is no human labor, there

is no human error. The only way you can build a secure system, a truly secure system is to eliminate human error. Self-driving cars will not crash or they won't crash-- they'll crash 1% as often, a tenth of a percent as often than human driven cars. Automation is much more reliable, much more predictable, much safer than doing it by hand. The Autonomous Database is also fully elastic in the cloud. So let's say you're running basic accounting system, and once a month or once every two weeks, you run your payroll. And you're a big company. You're a

big retailer, and you have a few hundred thousand employees, and the payroll takes a while to do all the calculation, all the withholding tax vacation days, all of this other stuff to do all the calculations. And to get that done quickly, you'd like to allocate even more processors. Well, the Autonomous Database is completely elastic. If you need more processors, you immediately allocate them, use them, run 10 times faster. When you're done, you give them back to the pool. So it's much faster. You get the processors you need when you need them. When you don't need

them, you give them back. So you only pay for what you use. This is not how other cloud databases work. You say I want eight processors or 16 or 32 or 64, and you pay for them whether they're working or not. In our case, it gets what you need, you go faster, returns it if your database is doing nothing. There are no processors allocated. And it costs nothing. So it's much higher performance, much lower in cost and because of no labor much more secure. Object data relational database-- the big-- relational database is much more powerful.

The issue that programmers have with relational databases is they have to predefine what's called the schema. Again, I'm not going to go into a lot of details. But you have to predefine the schema before you start programming. Programmers say I don't want to do that. I just want to start coding. I'll define my-- I'll define my user interface, which are JSON documents or what are called JSON objects and JSON notation. And there are databases that support JSON. There's JSON data-- we have one too. We have one too. But what we do now is we generate

the schema from the JSON documents. The programmer doesn't have to worry about the schema anymore. We automatically generate the schema. You just want to worry about that your user interface and your documents, start programming, we'll generate the schema, and you'll end up with a relational database foundation. Even though you did nothing, the system will generate it automatically. And you have the power of SQL for as a query language. And you have the simplicity of a JSON just dealing with documents, just dealing with your user interface program that you write. Easier to write programs, but the

programs are durable. You have a powerful query language. You get the best of both worlds. And they absolutely coexist because we generate the schema. Lots of things are changing. Most customers are not going to be building-- are not going to be building the same thing that xAI is building or OpenAI builds or Cohere builds. They're not going to be building one of these foundational model, large language models. They're not going to be trying to solve the self-driving problem and process vast amounts of image data and build a special neural network just to handle the image

data. They're not going to be doing that. Again, Cohere does. OpenAI does. xAI does. But most people are going to take what Cohere builds, and they're going to-- just like sending your kid to college and then after college you send them to law school. So they're going to start with the college educated kid from Cohere and then send them to law school or send them to medical school to train them specifically on medicine. And that's what Oracle-- we make easy in the Oracle Cloud. You can come to the Oracle Cloud and start, let's say, with

a Cohere foundational model and then take supplementary data. Let's say we take the Cerner EHR data anonymized, of course, we anonymize it, and we use that EHR, that medical data to specialize the Cohere AI model on medicine. So we train it on medicine. And that's exactly what happened. A company in Israel called Imagene actually did that. They took a bunch of cancer biopsy slides and fed it into the AI-- into the AI model, and they're now able-- the computer is now able to diagnose cancer in a matter of minutes. I don't know if Imagene is

here at CloudWorld, but it's a remarkable team from Israel that's done absolutely brilliant work. But this will be a more normal thing to do than-- a lot of people will be building these specialized models. Not many companies are going to build foundational models. And that makes complete sense. We're doing the same thing, by the way. We're not taking biopsy slides and trying to train a model for cancer detection, but we are taking a lot of electronic health data and training it to give doctors orders, whether it's discharge notes or-- but again, the idea is to

give the doctor a draft of a discharge note or a draft of an order. The doctor reviews it, edits it, and then submits it. So the model is making the doctor's job easier. It is not taking over the doctor's job. The idea is to-- this is a very, very powerful tool that makes the physician's job easier that whether the physician is diagnosing something using Imagene or placing an order for a patient in a hospital using Cerner New Millennium. OK, one of the things we do is we take the image-- the biopsy data from Imagene. They

keep it private. It goes into our cloud, but one of the interesting things is we can't see it. No one else can see it. It's private to Imagene. So a lot of people are very concerned about sharing their training data. Let's say we have an investment banker that's a customer, and they have a bunch of financial trading data that they don't want to share with the world, but they'd like to use it to train a model. Well, in the Oracle Cloud, your training data-- you can train and specialize the private model with your private data

that remains private after training. And we've done that for software companies and for financial services companies, medical companies. And in fact, last year, I said the reason we're optimistic about being able to make real progress in automating the world of medicine, which is enormous, is that we're not trying to do it alone. And it's important that we have partners like Ronin working on part of the problem and partners like Imagene working on other parts of the problem. And we provide some of the fundamental tools that make their job easier. Another thing, I'm not sure there's

anything more important that we're working on right now than our new Oracle Cloud Data Intelligence Platform, especially-- and that's a combination of Oracle analytics plus generative AI. And the specific example I'd like to cite is Oracle's new public health data intelligence platform. Now, this used to be-- back in the day, this started as a project called Cerner HealtheIntent. But we-- fundamentally, we've rewritten Cerner HealtheIntent. We've rebuilt it since we bought Cerner. And it's become the Oracle Public Health Data Intelligence platform. And we've unified-- we've Unified National Population Scale Health Data. This is designed to take

all of the EHR data for a country, for all the patients in a country and put it together, all the diagnostic laboratory data, everything, putting that into a single Oracle Autonomous Database for the entire population of a country. When you do that, when you take all of this health data and put it in one place, you get enormous benefits. The first benefit is when you go to train AI models, you have 1,000 times more data than you used to have. You have all the data. Often, for a drug, the best data you had you obtained

during the clinical trial. But once the clinical trial is over, they don't gather nearly as much data after the trial is over. This is like a clinical trial that goes on forever, a clinical trial for everything that goes on forever. We keep collecting all of this data. And then we use that data to train models. And that enables-- the answers we get from those models enable detailed personalized medicine. Now, one of the most common prescriptions in the world is a drug called a statin, which lowers cholesterol. And there are a number of statins. And I

happened to be on a statin called Crestor. And I talked to my doctor, who's a very good doctor, a very fancy doctor and a molecular biologist and said, well, how did you decide Crestor versus this drug or that drug or that drug? He said, well, you know, you're a guy. You're European descent. We really-- the truth is we don't have a lot of data. It was kind of an educated guess, if you will. We thought given your family history and what we know about you, we thought that Crestor was a pretty good guess. Well, wouldn't

you-- wouldn't I like to know, given my genomics, given everything you know about me if you had all of my health records, you could tell me exactly what the best statin for me was and the best statin for everybody, but not just statins, a variety of different drugs. You just made better choices based on who I was and what you knew about me. You can do that. You can do that if you have this national population scale data on public health and you collect all the laboratory information and all the diagnostic information. You have this

wealth of data that will help doctors make much better decisions of what therapeutics to give you, and that will deliver better outcomes at a much lower cost. By the way, better outcomes-- anything keeps you out of the hospital is lower cost. So strangely enough, everyone likes better outcomes, the patient and the payer. That works out for everybody. Mistakes, mistakes are very costly in terms of human suffering and sometimes human life and money, just bad. So with this new Oracle public health data intelligence platform, we have this fabulous training data, and we can do personalized treatment,

get better outcomes for millions of patients. I'll say it again. I'm not sure there's anything we're working on here at Oracle that's more important than this. T. K. Anand is having a session on this, and I invite all of you who are interested in health care. Please don't miss that session. OK, so other things that we're doing. One of the most important things if you're training AI models is having the information, having the data to train the AI models. And you'll be shocked that some data still isn't digitized, some medical data isn't digitized. The quantity

of medical image data, whether they're MRIs or these fancy two tube Siemens scanners, even sonograms, sonograms aren't saved. Biopsies sit on glass slides in drawers. They're not digitized, by and large. I mean, a small number are digitized, but the vast majority of them are not. We don't capture this data and digitize it and vectorize it for training data. And we should, and we're in the process of doing that. The Cerner care-- what started as Cerner care where, again, has now become a huge IoT project for here at Oracle where the patient monitoring that goes on

in the hospital to see, you know, and they keep track of your blood pressure, obviously, and your temperature and how much oxygen is in your blood, all of that because if you drop below a certain threshold, an alarm goes off in a nurse's station, and someone comes in to see you. But that's not recorded. That's not permanently recorded. We'd like to start permanently recording all of that. So we'd like to make sure that the systems in the hospital, the patient monitoring systems records all of this data so we can use it to train systems. Diagnostics,

these are the captures. The images are so large. Sometimes they're not captured at all. If you go in and you're pregnant and you get a sonogram, that sonogram is not stored digitally. There is usually a technician or a physician that reads-- looks at the sonogram and actually measures the skull of the fetus and the length of the spine and looks for problems like the umbilical wrapped around. That's all done looking at this sonogram image, a human being looking at the image. Having a computer assisting with that would be a big help for the physician or

the technician. Recording that would be a big help if you miss something and you go back and find-- go back and find it again. You shouldn't be measuring stuff on a screen. The computer should provide all of that. Plus a lot of these home devices and wearables give us the opportunity to collect even more data. But all of that needs to be integrated and used-- and used as training data. And that's a project we're working on. Again, we have partners in that effort, and we will be announcing them. We capture other data. It's really interesting.

Inside of a hospital, one of the biggest problems is their inventory. You're having a certain kind of surgery on Monday, and the surgical nurse, he may on Friday run around in a variety of places and collect all of the things that you need, tranexamic acid to stop bleeding, the doctor's favored gloves for operating, the doctor's favorite scalpels, people have personal preferences, all of this, take it all out of inventory and hide it in a ceiling tile outside the operating room so when Monday comes, they'll have everything that they need. They withdraw it from inventory that

way. They don't know what their current inventory is. They don't know where things are. They have crash carts. If they need tranexamic acid very quickly, they can go to a crash cart and get it. Tranexamic acid is something that just stops bleeding very fast. And everybody knows the first thing you do is stop the bleeding. So one of the things we're building as part of our Fusion products, actually, is an IoT system that keeps track of all of this stuff. It has RFID tags on all of the inventory, and we know exactly not only how

much a hospital has but where everything is located if you're looking for something. So that's all being automated by supply chain. Cerner had a robotics group that is automating laboratories. Automating laboratories is great, but the laboratory-- but the laboratory devices should also send the results and store them directly in the electronic health record. We need to do that-- we need to do both of those. We're working on that. And then robots need to build some of these medical devices if you want to make them inexpensive and highly reliable. And I think here we have a

really interesting video of the Cerner-- the Cerner robotic system that automates the laboratory. I guess do I press another button to get that video to play out? Yeah, we'll find out. I mean, it's very interesting. This is the Cerner system that we got when we bought Cerner. And it actually didn't have a lot of people working on it. It had been a de-emphasized area in Cerner. It was very important for us-- that was a quick video. It was very important for us because we thought we need to automate the labs, we need to automate the

labs, and we need to collect the data directly out of the labs and put it in the EHR and actually send it-- the lab results to the doctor, to the patient electronically on the patient's smartphone, for example, you go to the lab and have a PCR test for COVID. And yeah, there should be an automated way that you can notify a patient what those test results are, an automated way you can notify your physician what the test results were, an automated way to put that in the electronic health records. Next video, I'm going to-- I'm

going to show you is another robot project at Oracle. This was an ongoing Oracle project during COVID where we're building a microfluidics blood lab for a COVID immunity test. And this is kind of a next generation-- I hope they play the video. They're building these microfluidics disks that test you for your immunity levels for COVID. I mean, right now, I mean, I think I'm scheduled for yet another booster tomorrow with the latest variant of COVID because there's really no convenient way, easy way for me to find out what my COVID immunity is. I just know

how many boosters I've had. I didn't know what my immunity level really is. So there's this company that's working on a blood test that will measure your immunity level, and you do the blood test in these little disposable laboratory in a disk, and we're building the robots that build the disks. So we're building robots that build blood lab robots, if you will. And that robot is attached to the internet, and it will tell you and your doctor at the same time what the results are. So it is, again, collecting data and making the data available

to patients, to physicians and the anonymous version of that data available to AI models for training. There's enormous amount-- again, enormous amounts of data that's going-- that we need to collect. There's enormous amount of training data that we use-- we need to use to build these models that will provide information and help to people in the health care industry. But right now, this is going to shock you, genomics data, genomics data is discarded. And Oxford Nanopore was using-- trying to store their data, their genomics data in the cloud, and it became so expensive they decided

just to throw it away, just throw it away. Imagene, which wants to-- would love to look at 100 times more biopsy-- more images of biopsies than they currently have, but those images of biopsies are on slides. They're not even digitized. And one of the good things about the Oracle Cloud being less expensive is the Oracle Cloud now provides an economical solution to this problem. We can digitize these slides. We can save the genomics data. We're not going to be throwing it away. Instead, we're going to be feeding it to big AI models where we get

a huge, huge payback. Again, payback in terms of better outcomes and lower costs in medicine. OK, I focus-- I've used my example for most of AI medicine, but there are other industries that can be transformed using these 21st century technologies and generating applications. What does it mean when I say we're going to generate applications rather than write them in Java? I said teams are going to be smaller, and they're going to write applications faster. Does that mean that we're going to have massive layoffs here at Oracle? No, no. We're too ambitious for that. What we

are is going to try to do a lot more. We're going to generate. Because we can generate programs, we can tackle things that are much bigger in scale than we've been able to tackle in the past. We have better tools for building applications. We can build better applications faster. So that means we're going to be building using the same technologies to build more applications and more industries. And I thought I'd give you a couple of interesting examples. So APEX front and center, an application generator, don't hand code things. Autonomous Database, population scale data runs without

a DBA, without a driver, robots and sensors. So what are the kind of things we're looking at? Well, along with a partner, actually, Danny Hillis's company, Applied Inventions, we're looking at an industry even more primal than health care. We're looking at the agriculture industry, which started about 10,000 years ago in the floods on the Nile Delta when Egypt was the huge supplier of grain to the Roman Empire. Actually, there was no Roman Empire 10,000 years ago. That was a couple thousand years ago. But agriculture started 10,000 years ago on the Nile. So when you grow

indoors, it turns out you use a tiny fraction of the water than when you grow outdoors, 98% less water. And water is our most precious commodity. We are running out of fresh water. And as our climate changes, we need to adapt to those changes. We need to slow the rate of change, and we also need to adapt to the change. And droughts are now occurring more frequently in the Horn of Africa and Ethiopia and Somalia. And we-- one approach is to grow with less water, to grow indoors with less water. You also use much less

land when you grow indoors because the plants don't reach full size until right-- almost right before they're harvested. And if you have a greenhouse and you migrate the plants, you start them small and give them a little more room and a little more room and a little more room and then a lot of room right before you harvest them, harvest them then you go back to seedlings, you save about 90% of the land when you grow indoors. So we don't have to burn down the Amazon rainforest to create more arable land for farmers. We don't

have to steal animal habitat in Africa because we need more arable land for farmers. We have enough arable land to grow all the food we need, but we need a combination of growing indoors and outdoors. We need to grow near population centers. So we don't grow in Chile and then ship it-- ship that food to China. And, you know, just grow in just in Chile's outdoor growing season. The great thing about growing indoors is every day is a growing season. Every day is a harvest. You can grow near population centers. And if you don't transport

the food as far, you dramatically lower the greenhouse gas emissions produced. And by the way, agriculture I think is something like 35% of all greenhouse emissions, gas emissions. By the way, I'm not talking about indoor greenhouses you're growing in. I'm talking about the planet Earth is one big greenhouse. Agriculture is responsible for about 35% of CO2 and methane and other greenhouse gases. So grow closer to big cities. A harvest every day means you have less expensive, fresher, more nutritious food. And you have better job security because it's not seasonal. There's a growing season in Chile,

you harvest a bunch of strawberries in Chile, you go down there and then once the harvest is over, you're done. You migrate someplace else to harvest some other crop. This is in every-- it's more like a factory job. You go there every day. There's a harvest every day. There's more job security that way, but there's also a better job quality. If you're harvesting, even if you're harvesting by hand, it's not a robot harvest, you're harvesting by hand, you're doing it sitting down in a climate controlled area that's way better than leaning over with a knife

and cutting-- cutting an artichoke off a plant growing on the ground. So it's all together vastly more productive and better for everyone to grow indoors. There are challenges to do this. You've got to build a greenhouse that works efficiently any place in the world. That might mean you have to put in a local solar power station. We do that, or you got to build a complete growing system. It's got to be the greenhouse building, the irrigation systems, the sensors, the robots, you got to manufacture-- you got to design that all together. It's like a car.

Every single piece of the greenhouse is a unit. They're all the same. It's a standard greenhouse design. It can grow lots of different crops. You have a robot factory that builds them. You put them into standard shipping containers. You have a robot that digs the foundation of the greenhouse, and you snap it into place. You snap it into place very quickly. You assemble it very quickly. These modules very quickly. And it's always connected to the cloud. It's always gathering data. It's always being directed by the cloud. The nutrition is-- we detect when a plant doesn't

have enough water, doesn't have enough nitrogen, it's by a cloud system. And the irrigation system puts nitrogen in the water and takes care of that problem. It's a modern approach to an industry that's 10,000 years old. Sound familiar, we collect huge amounts of training data about sunlight in different areas, nutritional data. The plant images tell us when they're ready to transplant, when they're ready to harvest. And the plant genetics allows us to develop new varieties, more robust varieties that grow better in different climates and are more desirable in different cultures. And in the hydroponic greenhouse,

the heavy unpleasant jobs are done by robots. It moves entire rows of plants. People aren't allowed to go in the growing area. People go in the harvesting area. People don't go in the growing area because you don't want to contaminate the plants. You don't want to have infestations of bacteria, E. coli, insects, and like-- so you don't need any pesticides because it's really a clean room. You're growing in a clean room. OK, and it's not like a fantasy we're just imagining. The company has about almost a million square feet that's currently growing. This is the

latest version of the greenhouse. You see on the left-hand side of the picture, you can see the overhead robot lifting up several rows of plants to be transplant-- moved to be transplanted. It just will be moving into the harvesting area. The new generations of-- so there are greenhouses producing in Hawaii. There are greenhouses producing in Canada. The latest generation of greenhouse is being developed in Southern California. And you can visit it if you want to in the fall. So congratulations Danny Hillis and team. They're doing a fantastic job tackling a problem that's 10,000 years old.

They're using-- Applied Invention's using the Oracle Cloud, the Autonomous Database, NetSuite to manage the business and collaborating with the Oracle engineering team on AI IoT robotics analytics and with Tesla on the satellite network. OK, another thing I think we need very much is better support for our first responders. And here, this is an Oracle project. So we built a network for our first responders, an audio-visual network that never fails, that always works. I don't know, I mean, the tragedy in Maui, one of the things that contributed to the tragedy was they lost the cellular network.

The cellular network failed. This will never fail because it's not just a terrestrial network, it's also a satellite network. Both of them operate simultaneously. One backs up the other. So it's a net-- for first responders, the network is not allowed to fail. Again, always on. It's a combination of 5G and satellite connection. It's for electronic or emergency medical technicians, firefighters, police officers. They have wearable-- they wear computers. The police, they actually wear computers. If a fireman falls down, if a policeman falls down, you know. The accelerometer on the computer does that. The firefighters have always

on cameras. The police have always on cameras. Those cameras are not recording on the camera. The camera is transmitting to the internet and to a command center who can see what the firefighter can see, who can see where the firefighter is, who sees what the police officer sees. So there's a command center, a real time command center. So the EMT is never alone. They're always in contact with the hospital emergency room when they got a patient in transit. The firefighter is never alone when they're fighting a brush fire. They're connected by satellite. And if the

wind changes, when the wind changes, the command center will notify them of that to avoid a tragedy. A police officer who's in a difficult-- a young police officer who's in a difficult situation, a rookie police officer always has supervision. There's always someone in the command center who sees exactly what that officer sees and can assist them and supervise them. So a lot of the tragedies-- sometimes under pressure, people make mistakes. EMTs make mistakes. Police officers make mistakes. Firefighters make mistakes. They're now always being supervised and assisted in doing their very difficult jobs with the people

in the command center collaborating with the people in the field. That's real. That was the first Oracle police car-- excuse me-- and we had to put in, if you will, a Tesla-like screen in the Tesla position where the application-- the navigation application checking drivers-- checking license plates, things like that, all is done, all voice-- all are voice-activated. Everything is voice-activated. If there's an incident the officer is driving to, they will see what the officer who's already on site sees from their camera. They'll see what's on the audio video network. Our next generation police car is

coming out very soon. It's my favorite police car. It's my favorite car, actually. It's Elon's favorite car. It's incredible. I know too much about it. Some of it's still to be disclosed. But among other things, it's very safe, very fast. It's got a stainless steel-- stainless steel body. And we don't have to add a screen to it because or we don't have to add cameras to it because we actually use their existing cameras and their existing screen to put our application-- our application up for people in the Tesla vehicles. You can use a non-Tesla vehicle,

and we can enhance it or you can use a-- you can use a Tesla vehicle. All this stuff is up and running. It's actually deployed in Stanislaus County I know for the police and I believe also for the fire, which one command center. Stanislaus County is a rural area in California near Yosemite Valley. And of course, it's an area-- California summers are getting drier also and an area vulnerable to brush fires. And we're very worried about, again, brush fires and firefighters who are in remote areas where there is no cell service. So we have to

communicate with firefighters by using satellites, and we have to inform firefighters in a safe way using, again, robotic drones. And the drones are fabulous during the dry season. They can be used for monitoring fire-- monitoring for forest fires. They can be used in urban areas during heavy traffic. They provide improved situational awareness, very important during fires to know what streets are open and what streets are closed and know immediately and be able to let people know immediately what's the best way to exit the area, what's the safe map to get out of the area. OK,

this is my-- you'll probably be grateful. This is my last topic. My last topic is MultiCloud. Last week, I was up in Redmond, Washington with Satya Nadella, the CEO of Microsoft. And we had a wonderful chat. And we agreed on one big idea. And that big idea was clouds should be open. They should not be walled gardens. Clouds should be interconnected. And customers already use multiple clouds. I mean, people don't go to one and only one cloud. In infrastructure, there are four hyperscalers, AWS, Microsoft, Azure, Google, and us. And there are a lot-- and there

are a lot of application clouds. But there are still clouds. But people use stuff from AWS, and they use stuff from Google, and they use stuff from Oracle, and they use stuff from Salesforce, and they use stuff from ServiceNow. And it's our job to let customers-- once customers choose what they want to use to make all of that stuff work seamlessly together to interconnect those clouds in such a way that if you want to have-- use the Salesforce sales automation application and then have a data warehouse in Oracle, for example, where you mash up that

data with other data, that that's a very simple, straightforward thing to do because the two clouds are connected. If you want to use OpenAIs, ChatGPT with an Oracle Autonomous Database, so you want to use something from Microsoft and you want to use Microsoft Teams, you want to use Microsoft Office, you want to use ChatGPT, any of the Microsoft services with ChatGPT with Oracle and Oracle, that should be easy to do. There shouldn't be a wall between Microsoft's cloud and the Oracle Cloud. There shouldn't be a charge if you want to move your data out of

the AWS cloud and put it into a database in the Oracle Cloud. It is your data. I think-- so those clouds should be interconnected, and they should seamlessly interoperate. And we used to call that open systems. And we think-- Satya and I think the world's going back in that direction. Customers are insisting on it because they use multiple clouds. We started experimenting with this back in 2019. We internet connected our clouds. We had a lot of successful customers, and then we decided we need to do even-- make it even-- but it wasn't that easy to

do. We had high speed interconnect, but you had to be fairly sophisticated to connect Microsoft services to Oracle services. We decided to make it completely transparent. So you can go straight to the Azure portal and configure an Oracle service like the Autonomous Database and connect it to a Microsoft service and make it really fast. So what we've decided to do was actually build Oracle clouds right inside of Azure data centers and full Oracle Cloud inside of Azure data centers that you can go to the Azure portal and configure everything you need from Oracle or from

Microsoft, and it just works-- and it works super fast. And there are no charges. There are no data egress fees. It's your data. It is your data, whether it's in an Oracle Cloud, in a Microsoft Cloud, in an AWS cloud, in a Salesforce cloud, it is your data. They shouldn't charge you for moving it. OK, and again, it's the full Oracle Cloud. It's everything-- all Oracle Clouds are the same. All Oracle Clouds have the exact same hardware, the exact same software. The only-- all Oracle data centers. The only thing they vary by is by scale.

Some are big, some are small, but they all have all of the services, all of the services, every one of the clouds. Because of that, it's very-- because they're standardized, when you standardize something, you can automate it because they're all the same. So they use the same automation. So our cloud is very different. I mentioned the high speed network. The other thing is the fact that they're all the same, that they're standardized allows us to have a high degree of automation in our cloud, which makes it much less expensive to operate and much more secure.

I think I said right at the beginning of the presentation, almost all security problems are caused by human error. When you automate the cloud, you get-- we have Autonomous Linux. We have Autonomous Database. When you automate the cloud, you eliminate human labor, which makes it less expensive, and you eliminate human error, which makes it much more secure and reliable. You've just got to be willing to spend less. So we put the whole Oracle Cloud inside of the Microsoft cloud. They're co-located, so there's no latency. The network's super fast, very convenient and easy to use. Oracle

has more-- this surprises a lot of people where Oracle has more cloud regions than AWS, than Azure, than Google. Right now, we have 64 regions, major regions, but in addition to these major regions, in addition to these major regions, we have this very interesting thing called the DRCC, a Dedicated Region Cloud@Customer. So we can put the Oracle Cloud, all of it inside of a customer data center. And so the customer can just choose what data goes in and who the users are. They can have complete control of who goes into that data center. It's the

full Oracle Cloud. Oracle owns the hardware. You pay for what you use. Oracle manages the hardware but doesn't-- but doesn't look at the-- we don't look at the data. It doesn't manage the data. It doesn't decide who goes in there. That's your cloud. That's a dedicated cloud. You decide who goes in there. Interesting company, Nomura Research, has a couple of these DRCCs, and they're buying two more. They're going from two to four of these. And what they do is they actually-- they're a big consulting firm, an IT consulting firm, and they sell their services plus

Oracle Cloud services to the financial industry in Japan. And they actually run the Tokyo Stock Exchange in the Oracle-- their Oracle Cloud. It's the only case of a major Stock Exchange being run in a cloud any place on earth. I mean, those applications haven't migrated to the cloud yet except the Oracle Cloud, which is highly, highly reliable. It has to be reliable. Stock exchanges can't stop or people get very upset. And they are also a big APEX user. I mean, they are APEX fans. I'm not going to say fanatics. I was going to say fanatics,

but they've dramatically cut their development costs by using APEX and generating applications rather than hand coding them. We think this is an important new development, the fact that the clouds are now small enough. The full Oracle Cloud is small enough. We could put one right in Azure. We can put them-- in fact, we're building I think a dozen of them, a dozen data centers in Azure data centers with Microsoft. We're building them at specific customer sites. A couple of big phone companies are getting them to run their network. There are a lot of examples of

this. And we think right now, we have more data centers than any other hyperscaler, but still, we're on our way to 100. We think looking at the price of the DRCC, we're going to be the hyperscaler that measures its data centers, not on the way to 100 but really on the way to 1,000. And this thing connects to all the other clouds. So it will connect to the Salesforce cloud. It will connect to the Microsoft Cloud. It will connect-- it will be your interconnect to all of the other clouds. Plus it will be your own

dedicated cloud region, fully automated, fully modern, RDMA network, everything at a very-- at a cost that may fit your budget. Thank you very much. [APPLAUSE]