This is part 3 of a 3-part series; click here for part 1 and click here for part 2.

In our previous posts we went over two of the three sections in our framework for describing how you can build the future with data and AI. Those sections were “Build faster with AI” by using GenAI services and “Bring AI to your data” with the powerful capabilities built into our database services. Now we’ll talk about how you can “Deploy and scale anywhere,” and by this we mean using distributed, multi, and hybrid cloud options as well as the freedom to use open source and your preferred tools.

Distributed Cloud options for everyone

Let’s start with distributed cloud options. Oracle is the only hyperscaler capable of delivering more than 150 cloud services across any environment, anywhere. You can literally run any workload in any location:

The OCI you experience across all these environments is the same: same architecture, services, experience, and value. And our Roving Edge Infrastructure puts cloud at the edge of your networks – often where data is generated, such as with Internet of Things (IoT) applications.

Man sitting and looking toward colleague

GPUs for every workload

We talk a lot about large scale and very high performance core infrastructure. But we can do much more, across a range of scale and performance needs, because not everything needs to be gargantuan. OCI has many GPU options for workloads ranging from small-scale to mid-level and large-scale AI on both virtual machines and bare metal.

Now to harness all that GPU and CPU compute power, most customers are turning to Kubernetes to orchestrate their AI resources. As an open standard, Kubernetes has the benefit of not only being extremely flexible with loads of automation, but also portable across environments, whether cloud-to-cloud or on-prem-to-cloud, in any direction.

And while this post only just skims the surface of OCI Supercluster, ultralow latency RDMA networking, and high performance storage options, it’s important for you to know that OCI Kubernetes Engine, or OKE, can help you orchestrate and automate it all.

OCI Kubernetes Engine

Indeed, OKE is optimized for AI, whether we’re talking about Training, which requires high computational power, typically in batches, or Inferencing, which typically runs more frequently yet less intensely than training and may or may not require the full horsepower of GPUs, depending on the use case.

With inferencing, OKE can recognize ongoing changes in demand and dynamically allocate more or fewer resources to efficiently meet the need while minimizing cost. And of course, OKE has centralized management with built-in security and resiliency, enabling you to schedule and prioritize jobs across a resource pool while ensuring high performance for your mission-critical workloads.
We really do believe that OKE is the best managed Kubernetes platform for AI workloads. Whether it is model development, or training, or inferencing, OCI AI services run on OKE, and we’ve just released a slew of important updates including:

  • Improved observability for GPU nodes
  • Support for Ubuntu OS, the default standard for AI
  • Container Governance with CloudGuard
  • OpenID Connect (OIDC) Discovery support
  • And OCI Full Stack Disaster Recovery (FSDR) service support for OKE, which we just announced

Futuristic control center with large status monitors

Open source support

In keeping with the themes of open source, flexibility, multicloud, and choice, I’ll reiterate that with OCI you are welcome to use any of your preferred technologies, frameworks, tools, and languages. We’ve given you the freedom of choice to use what you want, whether that be tried-and-true, or cutting-edge.

Our DevOps continuous integration / continuous delivery (CI/CD) service will happily interconnect with your existing Git repositories and tools, and you are also welcome to use OCI’s own cloud native services. Whatever you choose, you are never locked in, and you can use what you want. Bottom line: Our goal is to make things so simple and efficient that you’ll be happy to do your work using Oracle’s comprehensive application development environment—but it’s always your choice.

Java hits 24, 30, 41, and 63

Speaking of languages, I want to give a shout-out to Java, one of the world’s most popular languages. It’s no exaggeration to say that Oracle and the world RUNS on Java. And we’re proud to celebrate 30 years of Java innovation, which also happens to coincide with the recent release of Java 24. There are over 63 billion active Java virtual machines (JVMs) worldwide running critical applications that businesses depend on, with two thirds of those (41 billion) cloud-based JVMs today and dozens of billions more forecasted by 2027. In short, it’s a great time to be a Java developer.

Oracle APEX for low code development

Now if we go to the other end of the application development spectrum, it’s low code. Oracle APEX is one of the most popular and powerful low code application platforms available. With APEX you can build and deploy 20 times faster than with manual coding and it’s easy to quickly add AI to your apps – things like adding a natural language processing interface, simplifying AI Vector Search setup, creating text from a user prompt, and more.

Like Oracle Integration, APEX has a huge assortment of pre-built adaptors to help you quickly prototype, test, and deploy. And let me be clear: just because it is low code does not mean that it is basic; there are literally hundreds of thousands of enterprise APEX apps out there, and developers within Oracle use APEX to quickly create sophisticated new application features every day. If you haven’t tried APEX, I recommend it!

"AI" with connected icons superimposed over people

AI Solutions Hub

OK, so now let’s talk about how you can take all these amazing capabilities and put them to work in your own applications. The very first thing you should do is go to the AI Solutions Hub. Simply type oracle.ai into your browser, hit enter, and you’re there. The AI Solutions Hub is where you’ll find a curated list of ready-made solutions – think of them as OCI-verified blueprints.

We’ve distilled each one into a brief description and quick video along with step-by-step instructions, sample code on GitHub, service documentation, and in some cases, longer-duration hands-on-lab-type videos to follow.

We created the AI Solutions Hub precisely to help you get going quickly and remove any barriers to your progress. And if you don’t already have an OCI account, simply sign up for the Oracle Cloud Free Tier to get $300 USD in credits to use over 30 days to determine if these amazing capabilities are for you.

Yet more Oracle AI customer success

Here are two customers who are having strong success with multicloud and OCI AI solutions.

TIM Brasil is a large telecommunication company which had been operating a hybrid multicloud environment with Oracle Database on-premises and dozens of applications spanning AWS, Azure, and Google Cloud. As you can imagine, it was expensive and complicated to operate the networking, firewall, and integration needed to maintain this topology across so many entities. With the help of Accenture, TIM moved 100% of their IT to Azure and OCI, using Oracle Interconnect for Azure and Oracle Database 23ai to simplify the architecture while boosting performance and resiliency.

Now, TIM’s app and data integration supports any SaaS, application, or data source while the direct interconnection with Azure and Google Cloud means no data transfer fees. It’s an incredible success story and I encourage you to watch the video as their Chief Technology Information Officer describes the journey with our CEO Safra Catz at Oracle CloudWorld 2024.

Widelabs is an applied AI startup who chose OCI to train and fine-tune one of Brazil’s largest large language models (LLMs) in service of reminiscence therapy, a proven psychological treatment that helps patients suffering from cognitive decline through the recollection of lived experiences. Widelabs’ LLM generates biographical content for an Alzheimer’s patient—a written version of the patient’s life story—in just minutes based on a vast amount of text, audio, and images supplied by patients and their families.

It’s a very inspiring use case, because it is heart-wrenching to witness friends or family members slowly losing their memories, but Widelabs’ application can help with reminiscence therapy. It’s great to be able to support such an enriching endeavor.

Together we’ve come so far

OK, so that brings our journey to a close. Let’s recap what we covered. We started with how AI is the most transformative technology of our generation, how AI use cases are everywhere, and how professional developers like you are harnessing AI to increase productivity.

We talked about how Oracle has infused AI at every layer of the stack, and how our mission is to help you build, enhance, and scale applications with AI to extract the full potential of your business data. Then we discussed the three pillars for how exactly we help you, namely:

  • Build faster with AI by using GenAI services
  • Bring AI to your data with database AI and integration services
  • And deploy and scale anywhere with AI infrastructure, multicloud solutions, open source support, and more.

And throughout our journey we shared multiple stories of real customer success.

Make no mistake, Oracle has fully embraced AI. We’re all-in:

  • Vertical applications with embedded AI models that are trained for industry-specific challenges
  • AI Services packed with options ranging from GenAI to task-specific capabilities
  • Converged databases with Vector Search, Select AI, in-database ML
  • Cloud infrastructure with a unique AI layer based on our Supercluster technology, Kubernetes, and more.

Animated image of colleagues building a bridge

Now it’s your turn

Now it’s time for YOU to go build applications of the future with data and AI. We want you to discover, learn, and build. Start with the AI Solutions Hub at oracle.ai, and if you don’t already have an OCI tenancy, sign up for the free tier at oracle.com/cloud/free.

Learn about all the technologies we discussed at the Developer Center, oracle.com/developer.

Then take it a step further and learn and earn certifications at mylearn.oracle.com.

Lastly, start building right now with OCI Generative AI service, OCI Generative AI Agents, and Oracle Code Assist, which is in beta.

This is part 3 of a 3-part series; click here for part 1 and click here for part 2.