Oracle wants to be enterprises' 'one-stop-shop' for AI

Oracle has cast off the shackles of a long-running “perception of lock-in” according to a senior company executive as the cloud giant targets customer flexibility and choice as the key to success with AI.

Speaking to ITPro in the wake of the company’s annual conference Oracle AI World, Jason Rees, SVP for technology and cloud engineering, said the firm aims to offer a “full stack” ecosystem for enterprises ramping up AI adoption.

Key to this approach is a concerted focus on the vital, interlinked layers that are needed to successfully build, deploy, and roll-out AI tools and agents. From the underlying infrastructure to database management and marketplaces packed full of custom AI agents, Oracle is betting that it can be the one-stop-shop for enterprise AI.

The company’s established reputation for reliability with its Oracle Cloud Infrastructure (OCI) forms the foundation of its value proposition with enterprise customers, Rees notes.

Deals with OpenAI and TikTok, for example, are a testament to this, according to Rees, underlining OCI’s compute capacity for organizations seeking scale.

“On stage [at AI World] this year we had TikTok,” he explains. “Now these aren’t necessarily for AI workloads, but just from scale you’ve got OpenAI who talked about the partnership they have with Oracle.”

“The fact is it went from a ‘we need some compute capacity’ to actually ‘you are the one-stop-shop as a partner that will allow us to scale the business,” Rees added.

“The key thing is the underlying infrastructure of OCI is the right platform to be able to allow our customers to scale and do training and inferencing at an infrastructure level with AI.”

Full stack service upgrades

Oracle AI World saw a raft of updates from the cloud computing giant, including deals with AMD, a new AI agent marketplace, the launch of its AI Database 26ai data management platform, and the AI Data Platform for building applications.

Rees tells ITPro the breadth of announcements showcases the company’s approach, with database features in particular seeking to address long-running customer challenges with breaking down data silos. This is a problem many have faced during rapid AI adoption projects.

“I think one of the challenges we often hear with our customers is they've got silos of data everywhere,” he explained. “So you've got different silos of data in our cloud, in another cloud, on premise, whatever those things are.”

“We understand customers are going to have it in our database, they’re going to have it in data lakes, on object storage, et cetera.”

With 26ai, this is about “bringing the AI to the data”, Rees added. Through the data management platform, customers are able to essentially break down these disparate environments, allowing them to fully capitalize on private data sources which feed enterprise AI tools.

Combined with the company’s AI Data Platform, it’s here that IT leaders can fully unlock the value of the technology and begin delivering on promises made to the C-suite.

“Our customers know that to get the real value out of AI, they need to get access to the private data, and they want to do that in the most secure way possible. So the ability within something that we announced, the AI data platform, one of the core tenants is the ability to have a catalog of data,” he explained.

“So whether or not that data exists in our systems or exists in Snowflake or Databricks, the fact is you can reference against that data and use that data as an input into asking more reasoned questions.

“We're looking to bring together all the elements of data in a secure, scalable way, in order that our customers can really get the value out of AI.”

Choice and flexibility

Oracle’s new AI Agent Marketplace, also unveiled at the conference last month, is another example of the company building a more open, flexible ecosystem for both customers and partners, Rees tells ITPro.

The marketplace offers AI agent templates from approved Oracle partners and systems integrators. Independent software providers such as Box and Stripe are also available through the marketplace.

Building on the launch of its AI Agent Studio last year, Rees said this new marketplace has two key motivations. It provides enterprises with flexible options based on their individual needs, while also giving partners an opportunity to engage with a wider array of customers.

“We're not the only ones who are innovating, and if our partners can also innovate, why would we not allow that to be,” he said.

“It's all about choice. It's all about choice. We want to get to a point where a customer can get the value from what we're building, the value of what our partners are building.”

It all started with multi-cloud

This current strategy at Oracle is the culmination of a long-running process at the hyperscaler, starting with delivering interoperability with counterparts in the cloud computing space.

Oracle has focused heavily on multi-cloud capabilities in recent years, securing integrations with key industry stakeholders including Amazon Web Services (AWS) and Microsoft through its Oracle Database@Azure and @AWS services, as well as with Google Cloud.

“The reason we did multi-cloud originally was giving our customers choice, and not having to refactor Oracle databases in different people’s clouds,” Rees explains.

“You had all these workloads, custom applications, commercial applications that were underpinned by the Oracle database. In the beginning we were not able to offer a database service that we can offer now,” he adds.

“When we started with Microsoft, with Azure, we began with just an interconnect, right, a high-speed connection between our clouds that were located on the same campus. That was what customers were happy to take – a split architecture, low latency, and they were happy to do that.”

In forging closer ties with providers like Azure, AWS, and Google Cloud, Rees tells ITPro this allows Oracle to embed its database “directly in the fabric” of other hyperscalers, thereby offering a greater degree of choice.

“This is part of our strategy, making it easy – removing engineering headaches from our customers, our clients, and just making sure that they can get the best use of data and of course, AI going forward.”

Naturally, the company is still keen on encouraging customers to stay within its own OCI ecosystem, but the flexibility afforded by these integrations now makes sense given the state of the industry landscape – and AI focus on the part of enterprises.

“It does feel to me that customers are realizing that multi-cloud gives them optionality,” he told ITPro.

“It really is that ‘what's the right cloud for the right workload’ type thing, and it really does allow customers to say, well ‘I've got no legacy issues, I’m saving on egress costs with network traffic, but I'm getting the best out of the tool set that I'm choosing to use’,” Rees added.

“I also think it's just the right thing to do.”