Oracle Q2 Earnings Preview: 5 Things To Know

Oracle Cloud Infrastructure long-term growth, vendor partnerships and channel partners seeing durable Fusion demand could come up on the next Oracle quarterly earnings call.

Generative artificial intelligence. Growth prospects for Oracle Cloud Infrastructure. And Oracle’s appetite for more and deeper partnerships.

These are some of the topics expected to come up Monday during Austin, Texas-based cloud and database product and services vendor Oracle’s report on its second fiscal quarter earnings. The quarter ended Nov. 30.

Oracle Q2 Earnings

On the vendor’s previous quarterly earnings call, Oracle reported $12.5 billion in revenue, up 8 percent year over year ignoring currency exchange.

Oracle co-founder and CTO Larry Ellison told listeners on the call that more use of GenAI—a technology that has incurred great investment and interest from vendors and solution providers alike—“is a boon to our business.”

“The demands on data are getting stronger and more important,” he said at the time.

Here’s more of what to expect Monday on the Oracle earnings call.

Long-Term OCI Growth Questioned

In a report from earlier this month, Morgan Stanley said that its checks with large Oracle channel partners and resellers had the firm awaiting “more clarity on the durability” of OCI growth, including AI-related use cases, beyond the first half of Oracle’s fiscal year.

Part of the reason for this was feedback from Oracle channel partners leading Morgan Stanley to believe “bookings are below capacity on a go-forward basis” despite partners being “unchanged on core OCI growth,” according to the report.

The firm reported that “several partners” said “they saw fewer but larger deals coming out of this year's CloudWorld conference,” possibly due “to the back-office refresh cycle which began at the end of the pandemic beginning to roll over.”

Morgan Stanley said in the report that it will monitor this refresh cycle into the second half of Oracle’s fiscal year, but the firm is “confident that the extended backlog should insulate Oracle's back-office apps growth in the near-term regardless,” according to the report.

Along with the partner feedback, Morgan Stanley said that it suspects AI bookings “have largely slowed outside of the” contract with Elon Musk’s X.ai.

Oracle “recognized $2 billion in AI bookings by the end of F4Q, which was up to $4 billion by its F1Q earnings call but including the X.ai contract,” according to Morgan Stanley. “This would imply that the F1Q intra-quarter bookings were likely closer to $500 million, which is a steep downtick from the F4Q bookings number, and much of the incremental bookings likely more multi-year.”

Morgan Stanley also believes that the prior quarter’s slowing capital expenditures growth “remains a potential negative read on forward OCI/GPU IaaS demand,” according to the report.

An October report from KeyBanc seemed more optimistic on the long-term growth of OCI, with the firm saying that “meaningful revenue from the $4B in AI OCI backlog” should appear in Oracle’s 2025 fiscal year because GPU-based OCI supercluster deployments can take up to 12 months.

KeyBanc noted that Oracle could see “extremely lumpy revenue growth as large contracts do or do not get fully deployed in a given quarter.”

Short-Term OCI Growth Clearer

In the near term, Morgan Stanley feels “more constructive” on OCI’s setup partly because of capacity issues potentially addressed by the September general availability (GA) of bare metal instances powered by Nvidia H100 GPUs.

Oracle has said that by now its OCI superclusters can scale to 16,384 H100s in addition to the previous 32,768 A100s, according to the Morgan Stanley report.

Morgan Stanley estimates that OCI GPU IaaS revenue could gain a sequential $50 million or so during the quarter due to “a more material ramp of H100 GPUs in F1Q/F2Q.”

Along with the capacity issues potentially addressed, Morgan Stanley saw signs “that customers are already consuming against this added capacity.” X.ai said in November “that it finished two months of training for a beta of its LLM model 'Grok' using 'tens of thousands of GPUs',” according to the Morgan Stanley report.

“This comes after the company signed a $1.5 billion multi-year contract on OCI at the beginning of F2Q, which likely makes it one of the largest AI customers given total AI bookings amounted to $4 billion inclusive of this contract.”

SaaS Growth Looks Good

Morgan Stanley said that its channel checks showed “continued durability in Fusion applications growth, as partners continue to cite healthy backlogs and strong win rates consistent with prior quarters,” according to the report.

The firm heard of “particular strength in the Financial Services and Healthcare verticals.”

Strategic back-office applications revenue growth has easier year-over-year comparisons, according to the report. Morgan Stanley predicts a 25 percent growth in Fusion, accelerating from the 20 percent growth reported the prior quarter.

Oracle’s database offerings could contribute a potential growth acceleration in SaaS due to “a renewed customer focus on refreshing the data estate and its recent 23c release,” according to Morgan Stanley. However, the database products face a tough comparison year over year.

Morgan Stanley’s channel checks suggested “higher customer spend on larger data strategy initiatives typically for one of two reasons, either because some of these larger projects were put off amid lower budget visibility in the C1H or because organizations are re-prioritizing these projects in order to get the data estate primed for AI initiatives coming in late-2024 and beyond,” according to the report.

“We see this largely affirmed by strong reports from data management peers and believe that the Oracle Database 23c release in Sept. 2023 probably aids the company's ability to capitalize on this trend in a more meaningful way,” according to the report.

An October report from KeyBanc said that the firm was “reinforced in our view of Oracle's opportunity to capture meaningful share of both the IaaS/PaaS cloud market and of the evolving AI/GenAI infrastructure and applications stack” in part due to Oracle’s “breadth of database capabilities in Document/JSON, graph, multi cloud (Oracle Database@Azure), and soon vector.”

KeyBanc said that a potential differentiation for Oracle is “combining context data through vector capabilities with business data (i.e., structured/relational) to augment (including Retrieval Augmented Generation or ‘RAG’) LLMs while benefiting from traditional Oracle DB features like Exadata storage, RAC cluster, and security.”

The vendor’s “vertically integrated infrastructure and database functionality” could also benefit “a strategy around embedding generative AI across both horizontal and vertical apps,” according to KeyBanc.

On NetSuite, Morgan Stanley’s report predicted 21 percent growth, flat from the prior quarter. This reflects “slightly more caution given more uncertainty in the SMB and mid-market headed into the C2H.”

Growing Partnerships

KeyBanc’s October report on its “reinforced in our view of Oracle's opportunity” also came in part from Oracle’s partnerships with Nvidia and Cohere.

Oracle and Nvidia announced this year DGX Cloud and Nvidia AI Enterprise availability on Oracle Cloud Marketplace. Cohere serves as the large language model (LLM) layer of Oracle's GenAI services, according to the KeyBanc report.

Oracle has also partnered with Nvidia and AMD on GPU access, according to the report. GenAI deployments can involve up to 30,000 GPUs.

Overall, a more partner-friendly Oracle is good news for OCI, according to the Bernstein report.

“Historically, Oracle has been reluctant to enter into partnerships and wanted to be all things to all customers, but we have been seeing Oracle recognizing the value of their partners (especially the Microsoft partnership) and the need to leverage partners to drive adoption of their cloud offerings,” according to the report.

The report continued: “Partnerships open doors to win deals that Oracle might not have been able to get on its own, thereby expanding its cloud reach and unlocking the value of the Oracle database.”

A Potentially Lucrative Microsoft Deal

A November report from Bernstein said that Oracle’s recent deal with Microsoft to use OCI AI infrastructure with Azure's AI infrastructure to power Bing conversational search could imply an additional $171 million to $341 million for the OCI business, an increase of between 4 percent to 7 percent of revenue.

But the firm said “we believe that the direct incremental revenue is far smaller since it dislocates other AI workloads.”

“Oracle is likely walking away from or delaying some born-in-the-cloud demand to make this deal, but this deal gives Oracle much better long term value than simply being the location where a startup (that may fail) will train their model(s),” according to Bernstein.

The partnership with Microsoft has been evolving over the years, with both vendors announcing an Oracle Database @ Azure offering and Azure customers with commitment dollars having the ability to use that money toward Oracle services consumption as examples of the partnership at work.

The KeyBanc report said that Oracle DB @ Azure revenue could start to come in during the first half of 2024, contributing to Oracle’s 2025 fiscal year revenue.

While some investors could see the Microsoft Bing deal as a sign of Oracle having too much capacity, Bernstein’s report said that the vendor’s leaders have dismissed that assertion in public statements.

Plus, “Microsoft would not be doing this if Oracle's OCI performance were poor—or even not on par with Microsoft's infrastructure itself because that would put Bing's performance at risk, slowing it down,” according to the report. “This deal signals that Oracle is a competitive solution for generative AI.”

The partnership could have benefits outside of AI as well, according to Bernstein. The Bing deal “is one more step” toward “building the best and most seamless multi-Cloud,” an environment that “is otherwise very complex and in some cases even painful and expensive.”