In a rapidly evolving technological landscape, Oracle is at the forefront of integrating artificial intelligence within its cloud ecosystem. In an exclusive conversation with Express Computer, Leo Leung – Group Vice President, Oracle Tech and OCI, outlines how Oracle’s “full stack” AI strategy empowers organisations across various sectors—from tech startups to established enterprises. By offering tailored solutions that enhance operational efficiency, Oracle is reshaping business dynamics globally. With significant investments in AI infrastructure and strategic partnerships, the company is committed to enabling clients to harness the full potential of AI, ensuring secure and effective data utilisation in their operations. Leung along with his team is responsible for Oracle Cloud Infrastructure, which includes core infrastructure products like AI infrastructure, distributed cloud offerings, and multi-cloud solutions.
Question: How is Oracle helping organisations leverage artificial intelligence within their cloud ecosystem to enhance operational efficiency and drive business growth?
Answer: We refer to our AI strategy as a “full stack,” as it offers various levels of services that are tailored to meet the requirements of customer needs and specifications. Let me explain it in detail: for instance, if there are customers who are skilled in technology and focused on building AI-driven products, we provide them with high-performance AI infrastructure, utilising both our own technologies and partnerships, such as with NVIDIA. Companies like Suno, which creates AI-generated music, exemplify this category. There are such companies that are literally building AI-based products.
Further, there are technology companies which are integrating AI into their existing products and services. Zoom is one such example which has recently added an “AI Companion” feature that summarises webinars and assigns tasks. Zoom leverages our technologies to enhance their offerings. Zoom did not have AI-capability for some time.
Besides this, we have many business users who prioritise functionality over the technology. They benefit from our AI capabilities embedded in applications, be it HR software and other industry-specific solutions like construction project management software, etc which use AI to streamline processes.
So, we can summarise that our strategy incorporates various skill levels among customers, ensuring that our AI technologies are tailored for organisations rather than consumers. And this is what distinguishes us from some competitors which focusses on consumer markets.
Question: What are some typical challenges these companies which are integrating AI within their operations face? Additionally, what motivates organisations globally to fully utilise artificial intelligence? Could you also share examples of how Oracle has transformed the business dynamics for some of these organisations?
Answer: We have witnessed significant interest in generative AI technologies across the world. However, I am of the opinion that many companies are still in the early stages of adopting these solutions. While they may be utilising certain capabilities, like those in human resources, there’s still a long way to go. For instance, many customers, including our own organisation, are using generative AI to streamline processes which help us create job descriptions or enhance the review process by suggesting suitable jobs based on user input. Our software-as-a-service offerings can recommend other jobs within the same company as users refine their searches, demonstrating good adoption of these capabilities.
Now when it comes to comment on “Transformation”, the definition varies to a great level. Let me tell you about a large organisation like ours. With 160,000 employees, if we can save even an hour or two each, it can significantly boost our overall efficiency. Please note that the driving force behind this interest lies in our investments to integrate more AI capabilities within our database and facilitate secure integration with other AI technologies. Many customers already have sensitive and critical data stored in Oracle databases, raising concerns about how to utilise this data to inform AI models without compromising security. To address this, we’ve implemented various measures to secure data within our databases, allowing for features like retrieval-augmented generation. This approach has helped us to enhance AI-generated answers with context from private data while maintaining confidentiality. This capability opens up a range of new use cases for enterprises, moving beyond consumer-focused applications that typically rely on publicly available data.
Question: With Oracle’s recent expansions in regions like Malaysia and partnerships with organisations such as NVIDIA and AWS, what other factors differentiate Oracle from its competitors, especially outside of India?
Answer: Our ability to integrate generative AI within our database and connect securely to various large language models (LLMs) is a key differentiator. This technology is widely adopted, and we believe it enables companies to leverage AI securely. Additionally, our infrastructure is uniquely positioned. The significant investments we’ve made and that you mentioned are aimed at building not just data centres but also AI infrastructure capabilities worldwide.
Companies at the forefront of technology are training large language models and multimodal models, and they require substantial computing power. Recently, we announced the development of a single AI cluster comprising over 131,000 GPUs, functioning as a single machine. This surpasses the capabilities of any supercomputer in the top 500 list and is unmatched in the cloud. This level of computing power is essential for our customers engaged in model development. We believe that’s a major differentiator.
Our OCI supercluster, which utilises a networking technology called RoCE v2, enables this massive scale. Each machine is a bare-metal, cloud-based unit, allowing customers to harness every bit of computing power and storage available. This capability is a significant advantage for our large customers, enabling them to pursue a multitude of AI use cases effectively. No one has previously discussed an AI supercomputer of this magnitude operating as a unified system. Its uniqueness lies in the capability to network such a large number of machines, thanks to our technology known as the OCI supercluster.
We can proudly say that our database, AI infrastructure, and the applications that utilise these underlying capabilities set us apart and support a wide range of AI applications for our customers.
Question: How is Oracle using AI to enhance security within organisations?
Answer: There are multiple fronts, Oracle addresses the security challenges, and to deploy our cloud stack anywhere is one of Oracle’s such unique capabilities. While many large providers have focused on building bigger data centres in only a few locations, we offer a solution called “dedicated regions.” These regions mirror our public cloud stack and can be tailored for specific purposes, such as serving governmental needs or individual companies.
For instance, Fujitsu recently purchased multiple dedicated regions in their own data centres to cater to their specific customer base, while Vodafone has similar arrangements for their enterprise purposes. This allows us to create isolated cloud environments for individual customers while maintaining the robust security features of our public cloud.
In addition to dedicated regions, our database technologies are extremely secure. We utilise advanced encryption, automatic patching, and ongoing threat protection to ensure our customers’ data is safeguarded. Many security breaches occur due to human error, such as outdated software or misconfigurations. By automating security processes in our Oracle Database and cloud services, we alleviate these risks. So, what I mean to say is that we can create a cloud environment that is dedicated to a single customer while still providing all the features of our general-purpose public cloud. This setup offers physical separation and isolation yet maintains the robust security capabilities available in our public cloud. This is a distinctive offering.
Moreover, we’ve introduced a new technology called Zero Trust Packet Routing, which separates network configuration from security configuration. This approach focuses on protecting data itself, allowing organisations to establish policies that govern how data moves. Even if an attacker gains access to a credentialed user’s account, our policies can prevent unauthorised data transfers, enhancing security significantly. We’re actively developing this technology within OCI, with the aim of setting a new standard for security across the industry.
We’re tackling some of the toughest cyber security challenges out there. We’ve just introduced our initial implementation of OCI Zero Trust Packet Routing (ZPR) technology, pronounced as “zipper and our goal is to refine it further and establish it as a standard for other technology providers to adopt. That’s our approach to security. OCI ZPR is a user-friendly solution that empowers customers to define security intent by creating security policies in natural language and vocabulary, making it easier to restrict access to sensitive data through specific access pathways.
Question: What has been your experience with Indian clients? To what extent has India leveraged AI, and how much further do we have to go?
Answer: Adoption is progressing well, though I haven’t seen many pure AI companies developing models from India. However, we’ve observed a wide range of companies embracing our technology, including large enterprises, public sector organisations, and small startups. On the enterprise side, it’s no surprise that we have customers like Cognizant, along with an interesting case like IFFCO, a fertiliser manufacturer. I’m particularly intrigued by IFFCO because they’re involved in various innovative initiatives.
One of OCI’s strengths is not only in developing new applications but also in managing a company’s existing IT landscape. Most large companies I’ve spoken with still rely heavily on legacy technology—typically, 70-80% of their systems have been in place for years. IFFCO’s story stands out because we’ve successfully migrated over 100 of their existing applications to our infrastructure, allowing them to move away from data center management. Additionally, they’re now building new applications with us, including mobile solutions for managing drones that fertilise or spray crops in their customers’ fields. These transformations are particularly compelling, as it’s often easier for companies to start fresh than to adapt and modernise their existing systems.
Another good example is iHub, which is building modern healthcare management software. They leverage our autonomous database, which offers ease of use, scalability, and strong security, critical for handling patient healthcare records. Many companies may not realise that Oracle provides several open-source products, including PostgreSQL as a service, which has gained popularity in the Indian market.
We aim to provide not only our own high-quality technologies but also to cater to customers’ preferences and familiarity by offering a variety of services. Many customers may not realise the extent of this flexibility.
In India, there’s a vibrant open-source community that is quite knowledgeable about these technologies, making it essential to mix different solutions for that market. We’re continually improving in this area. For instance, Cognizant is leveraging AI significantly, both in their back-office functions like finance and HR, and in areas such as code generation. We also engage in some code generation ourselves. More companies are increasingly integrating AI into various aspects of their applications and workflows, aiming to streamline and automate processes, even if it’s not completely automated yet.
Question: How does Oracle measure the return on investment for clients implementing AI-driven solutions in the cloud?
Answer: The investment required for many of our technologies is minimal. For our Fusion Cloud applications, AI capabilities are included as part of the subscription, so there are no additional costs. The same applies to many platform services. This is particularly important as generative AI is still in its early stages. For other AI capabilities, such as AI vision and document understanding, we see broader adoption and strong ROI.
For clients with established AI initiatives, we offer competitive pricing for our infrastructure and services. Our generative AI service features competitive pricing per token, designed specifically for business use cases. Unlike general-purpose models, which are typically larger and more complex, our models are streamlined for specific business applications, making them more efficient and cost-effective.
We also prioritise making our capabilities available across different cloud providers. Whether a client chooses AWS, Microsoft Azure, Google Cloud, or Oracle, we want to ensure that our unique capabilities are accessible. Recently, we launched our database services on Google Cloud, and we’re also working towards similar offerings with AWS. This approach allows customers to leverage Oracle’s strengths regardless of their chosen cloud platform, ultimately expanding their options for AI services.