Oracle made its foray into the generative AI market, forging partnerships to offer large language models (LLMs) to its clients and announcing substantial commitments of cloud computing resources from startups that were engaged in training complex AI systems.
 
Three months later, Oracle is unveiling comprehensive access to this potent AI technology across its cloud infrastructure, electronic health record (EHR) software, data analysis tools, ERP solutions, and customer service software. Notably, signed contracts with AI developers have doubled in value, reaching an impressive $4 billion.
 
Larry Ellison, Chairman and Chief Technology Officer of Oracle, remarked on the transformative impact of generative AI. He stated, “Generative AI showed up about a year ago now, and it changes everything. It’s certainly changing everything at Oracle.” He made these remarks during his keynote address at the company’s CloudWorld conference in Las Vegas on September 19.
 
At the CloudWorld conference, Oracle introduced a generative AI cloud service, implemented artificial intelligence enhancements in healthcare software, and unveiled new database capabilities tailored to store and swiftly retrieve unstructured and semi-structured data essential for AI systems. Additionally, Oracle announced enhancements to customer management software, empowering marketers to create campaigns and enabling service agents to summarize conversations. Furthermore, a generative AI-powered assistant for business data analysts made its debut.
 
Generative AI systems, which acquire insights from patterns found in internet and private data sets to facilitate the generation of new insights, have spurred extensive development and investment. This has created a surge in demand for the substantial computing power necessary to train and operate large language models. These models find application in diverse areas, including medical diagnostics, market analysis, conversation analysis, content creation, and coding. The computational demands of LLMs exceed the capabilities of conventional data centers, prompting developers to turn to public clouds that offer the requisite networking and specialized hardware.
 
To place generative AI tools within the reach of businesses, Oracle is expanding its use of NVIDIA’s powerful H100 AI processors in its cloud infrastructure. These advanced chips bring advantages in networking speed and cost-effectiveness to Oracle’s cloud computing service. The response from AI development companies has been robust, with contracts for over $4 billion in training capacity on Oracle Cloud Infrastructure (OCI) signed, representing a twofold increase since the end of Oracle’s fiscal fourth quarter on May 31.
 
Oracle distinguishes itself with a cloud computing infrastructure designed to swiftly move data to locations where processing is required, facilitating rapid data transfer among processors during AI model training, as highlighted by Ellison.
 
Oracle is also addressing businesses’ desire to enhance standard models by incorporating their own data to improve predictive capabilities in specific domains. “They want to use their own training data,” noted Ellison, adding, “The best way to do that is to put that supplemental training data into an Oracle vector database.”
 
Businesses harnessing generative AI have the potential to reap substantial benefits. Goldman Sachs estimates that generative AI could boost global GDP by nearly $7 trillion, equivalent to 7% of the global GDP, over the next decade. This boost is driven by increased productivity among office workers, accelerated drug discovery, and expedited software development. McKinsey & Co. predicts that generative AI and related technologies could contribute up to $4.4 trillion to global GDP as automation gradually replaces half of the work currently performed in areas like sales, marketing, coding, and research and development between 2030 and 2060.
 
Oracle is introducing AI-powered tools on multiple fronts. In June, the company partnered with Cohere, a large language model developer based in Toronto, Canada, to offer AI capabilities to OCI customers based on Cohere’s technology. The first offerings stemming from this collaboration are now being introduced.
 
AI Vector Search: Oracle Database 23c gains the capability to store semantic information as vectors, enabling large language models to swiftly locate related terms. It also enables retrieval-augmented generation (RAG), combining LLMs with proprietary business data to provide knowledge-specific responses. AI Vector Search enhances developer productivity by incorporating generative AI capabilities into Oracle Database development tools like APEX, facilitating application and query development without coding.
 
OCI Generative AI: This service enables businesses to integrate LLMs developed by Cohere into their applications through an API and execute these models on Oracle’s fast and cost-effective OCI Supercluster service. While currently available for testing, it is expected to work in tandem with AI Vector Search, enabling customers to create RAG applications that blend pretrained models with their own data. This service also underlies new features in Oracle Cloud Applications, NetSuite ERP, and healthcare applications.
 
Increased AI Computing Power: Oracle plans to expand the availability of NVIDIA’s latest H100 chips within its cloud infrastructure. These GPUs offer significantly enhanced inference and training performance, outstripping the previous-generation A100 chips by a factor of 30 and 4, respectively. OCI Supercluster enables customers to connect tens of thousands of H100 chips over a high-performance, low-latency network, initially available in Oracle’s data centers in Chicago, Illinois, and London, England. OCI also intends to offer NVIDIA’s L40S GPU for training smaller models or AI inferencing.
 
Oracle Clinical Digital Assistant: Designed to complement Oracle’s electronic health records, this AI-powered tool streamlines paperwork by providing doctors with a voice- and screen-powered user interface. It automatically generates notes during patient interactions and suggests actions such as ordering medications, lab tests, and follow-up appointments. Physicians can also verbally access MRI results and other patient information from EHRs without the need for manual data entry. This software is expected to become available within the next year.
 
Oracle Advertising and Customer Experience: Tailored for marketers and service agents, this platform empowers users to create AI-generated customer responses, write and search help articles, and summarize essential information from customer interactions. Marketing teams can design campaigns that adhere to their brand guidelines through a guided AI-driven process.
 
Oracle Analytics Cloud: This data visualization and collaboration platform now supports interactions with data using plain language, with AI generating responses. The software also offers automatic extraction of key information from text and image files.
 
These developments reflect Oracle’s commitment to bringing the transformative potential of generative AI to its customers, enhancing productivity, and enabling innovation across various domains.