Oracle’s latest earnings once again proved to the market that in the AI-driven frenzy, future promises matter far more than current results. The software giant’s shares surged 40% on Wednesday, putting it on a direct path toward the trillion-dollar market cap club.

Although Oracle reported revenue and EPS that missed expectations, CEO Safra Catz ignited investor enthusiasm on the earnings call with three bold forecasts for the company’s AI-driven growth trajectory.

Record-breaking backlog (RPO) surges by $300 billion in just one quarter

The cloud company disclosed that its remaining performance obligations (RPO)—a key indicator of future revenue from signed but undelivered contracts—soared to $455 billion in Q2, up from just $138 billion the previous quarter. That means Oracle added roughly $300 billion in incremental orders in only three months.

Catz also noted that several new multi-billion-dollar deals are set to be signed in the coming months, which could push RPO above $500 billion. Among the most significant drivers is OpenAI, with Meta and Elon Musk’s xAI also contributing.

According to media reports, Oracle’s agreement with OpenAI involves supplying 4.5 gigawatts of data center capacity valued at roughly $30 billion annually. The deal is linked to “Stargate,” a large-scale AI infrastructure project jointly backed by Oracle, OpenAI, and SoftBank. To meet OpenAI’s massive compute demand, Oracle plans to build multiple new data centers across the U.S.

Aggressive upgrades to OCI revenue guidance

Buoyed by strong demand, Catz raised Oracle Cloud Infrastructure (OCI) revenue guidance for this fiscal year to 77% growth, up from a prior forecast of 70%, or about $18 billion. Longer-term, she projected OCI revenue could expand to $32 billion in 2027, $73 billion in 2028, $114 billion in 2029, and $144 billion in 2030—nearly doubling each year for the rest of the decade.

3. Surging capital expenditures to convert backlog into revenue

Oracle lifted its capital spending outlook for this fiscal year to $35 billion, with most of the funds earmarked for revenue-generating data center equipment rather than land or buildings. That represents nearly half of projected fiscal 2026 revenue—an extraordinary bet on the company’s ability to turn its unprecedented backlog into revenue and profit growth.

The numbers are staggering. Oracle generated only $15 billion in revenue this quarter, yet its $455 billion backlog now exceeds the combined RPO of the cloud giants. For comparison: Microsoft reported $368 billion, Amazon $189 billion, and Google $108 billion. Normally, cloud providers hold backlogs worth around 10 times their quarterly revenue—for example, Google Cloud’s Q2 revenue of $13.6 billion against $108 billion in RPO.

Oracle’s jump likely stems from its 10-year deal with OpenAI, valued at around $30 billion annually, which alone could explain much of the $300 billion quarterly increase. Still, RPO accounting remains opaque, making it hard to gauge exactly how much of these commitments are already included.

How much backlog converts into real revenue?

Google Cloud CEO Thomas Kurian recently said that 55% of its $106 billion in RPO is expected to be recognized as revenue within two years. If Oracle’s $455 billion backlog converts at a similar rate, it could mean as much as $250 billion in cloud revenue over the next two years. That would make Catz’s ambitious OCI revenue forecasts plausible, but only if Oracle can scale quickly enough to deliver.

While the OpenAI deal is transformative, it also raises concentration risk. Morgan Stanley estimates that OpenAI could account for as much as 30% of Oracle’s revenue by fiscal 2029. OpenAI itself is planning aggressive capital spending—reportedly $115 billion through 2029, up $80 billion from earlier forecasts—plus an additional $20 billion in equity incentives to secure talent.

But the sustainability of such spending is uncertain. OpenAI currently generates about $1 billion in monthly revenue, and it is far from clear whether that can support its massive expansion plans. Without diversifying its customer base, Oracle’s fortunes will remain closely tied to OpenAI’s financial health.

Margin pressure looms

As Oracle pivots away from its legacy software licensing model toward AI infrastructure, profitability is under pressure. Cloud services, especially AI infrastructure, carry thinner margins due to intense price competition and heavy capex requirements. With cloud revenue growing as a share of Oracle’s total, overall margins are likely to remain squeezed.

To offset the drag, Oracle has initiated cost controls, including layoffs—a sign that the shift to low-margin AI business is forcing efficiency measures.

Amid Fed rate-cut optimism and unrelenting AI hype, Oracle’s massive backlog and bold guidance have electrified investors, fueling speculation that the company is set to join the trillion-dollar club. But with customer concentration, ballooning capex, and pressure on margins, the AI frenzy driving Oracle’s stock surge could also sow the seeds of a bubble.