The $400 billion enterprise software and cloud infrastructure giant Oracle is in the hot seat with a fiscal third quarter earnings drop on Tuesday amid a spotlight on its heavy borrowing and negative free cash flow.

To set the scene, at the top line analysts are expecting about 20% growth in quarterly revenues to roughly $17 billion, right in line with Oracle’s guidance of 19% to 21% growth from the prior year. Earnings per share, excluding certain items, are expected to be up about 16% to $1.71. But under the hood? There’s a lot more going on and those issues wiggling around have helped send its stock down about 20% so far in 2026.

How Oracle’s stock fares after it reports results on Tuesday will depend largely on which storyline Wall Street chooses to focus on.

First up, job cuts. Last quarter, Oracle disclosed a 2026 restructuring plan that it expected would cost the company up to $1.6 billion primarily due to “employee severance costs.” Of that $1.6 billion, Oracle has recognized about $826 million in charges against the plan—that means Oracle still had about $788 million to go. Bloomberg reported last week that Oracle was eying layoffs in the thousands to rebalance its workforce and to lean further in on its shift from an enterprise software licensing company into a cloud infrastructure provider that competes with Microsoft and Amazon.

Meanwhile, Oracle has also turned to bonds to raise capital like the other hyperscalers, finishing its most recent full fiscal year with $92.6 billion in total debt outstanding. In the first half of its current fiscal year, the figure ratcheted up to $108.1 billion following a massive September 2025 issuance of $18 billion in notes with maturities ranging from 2030 to 2065. Oracle has also disclosed a further $248 billion in future data center lease obligations not yet on its balance sheet that it is hoping will translate into customer demand and rising revenues.

Last quarter, co-CEO Clay Magouyrk sought to reassure investors about its additional capital needs in the future. Magouyrk said the company is committed to maintaining its investment grade debt rating. Moody’s rates Oracle Baa2, which is two notches above junk and lower than Amazon, Alphabet, Meta, and Microsoft.

“We’ve been reading a lot of analyst reports, and we’ve read quite a few that show an expectation of upwards of $100 billion for Oracle to go out and kind of complete these buildouts,” said Magouyrk last quarter, referring to outside estimates of the company’s planned capital expenditures. “And based on what we see right now, we expect we will need less, if not substantially less money raised than that amount to go and fund this buildout.”

As is the case with the other hyperscalers, including Alphabet and Meta, Oracle’s capital expenditures are soaring as it races to build more data centers and infrastructure for AI. Last May, Oracle’s free cash flow turned negative by $394 million after the company’s operating cash flow of $20.8 billion was overtaken by its capex of $21.2 billion. From fiscal 2024 to fiscal 2025, Oracle’s capex jumped from $6.9 billion to $21.2 billion and last quarter Oracle guided its capex would be $50 billion this fiscal year. Meanwhile, its operating cash flow grew from $18.7 billion in fiscal 2024 to $20.8 billion in fiscal 2025 and analysts estimate it’s on track to reach $22.3 billion this year. The company has said it expects the negative free cash flow trend to continue as it pursues its AI ambitions.

According to founder and executive chairman Larry Ellison, this is all in service to Oracle’s three-step transformation. Ellison told investors last quarter that the first step was Oracle making its database available inside its competitors’ clouds, including Amazon’s AWS, Alphabet’s Google, and Microsoft’s Azure. Step two was “vectorizing” the data to make it readable by AI models, which makes the data customers have in Oracle’s systems more valuable, said Ellison. Third, Oracle built what Ellison called an “AI Lakehouse,” which vectorizes all a company’s data and not just what’s in Oracle databases or applications.

“Training AI models on public data is the largest, fastest-growing business in history,” Ellison said. “AI models reasoning on private data will be an even larger and more valuable business. Oracle databases contain most of the world’s high-value private data.”