Will Oracle's OCI IaaS revenue growth decelerate below 40% YoY in Q3 FY2026?

OCI IaaS growth has been accelerating for four consecutive quarters (51% to 52% to 54% to 66%), and growth is capacity-constrained rather than demand-constrained -- GPU capacity delivered grew 50% QoQ while GPU revenue surged 177% YoY. A deceleration from 66% to below 40% in a single quarter would require a dramatic structural break. Even the committee's minority position (Myth Meter STRETCHED assessment) focused on whether 40%+ CAGR is sustainable through FY2028, not whether a single-quarter collapse below 40% is likely. With 147 live OCI regions and 64 more planned, capacity expansion continues to drive the growth rate.

While the acceleration trend is strong, the 12-month RPO grew only 40% versus 433% total RPO growth -- this gap is the most concerning signal for near-term revenue conversion. If mega-contracts (particularly OpenAI, estimated at ~58% of $523B RPO) are backloaded in their revenue recognition, Q3 could see slower conversion than the headline RPO suggests. Additionally, Q2's 66% growth may have been partially amplified by large contract activations that front-loaded revenue. The below-market GPU pricing ($10/hr vs $14-18/hr) also means revenue per GPU deployed grows less rapidly than capacity additions would suggest. Still, dropping below 40% from 66% in one quarter is an extreme outcome.

The committee established REVENUE_DURABILITY as CONDITIONAL at medium confidence, meaning the growth engine depends on OpenAI contract performance. However, even the downgrade trigger was defined as 'below 50%' confirming the historical deceleration pattern, with 'below 40%' being the escalation toward FRAGILE. The committee itself treated below-40% as a tail scenario, not a base case. The CapEx-to-revenue ratio of 0.58x (4-6x hyperscaler peers) and below-market pricing are structural concerns, but they affect profitability more than top-line growth. GPU capacity is still being deployed at +50% QoQ, providing a mechanical floor on revenue growth even if pricing is compressed.

The growth trajectory is strongly upward: 51%, 52%, 54%, 66% over four quarters. Growth is mechanically driven by capacity deployment (+50% QoQ GPU additions) rather than dependent on new contract wins. The question asks about a single quarter dropping 26+ percentage points -- this would require either a massive supply chain disruption halting capacity deployment or a contract cancellation/modification of OpenAI-scale. Neither scenario is base case. Multicloud consumption growing 817% YoY and Autonomous Database growing 43% YoY indicate the growth is broadening beyond pure GPU buildout, providing additional support.

I weight the minority position (EXPECTATIONS_PRICED = STRETCHED) and the unresolved RPO conversion debate more heavily than the consensus. The 12-month RPO growing only 40% while total RPO grew 433% is a significant disconnect that directly affects quarterly growth rates. If the mega-contracts have longer conversion cycles than the market assumes, revenue recognition could disappoint even as bookings remain strong. The Stress Scanner's finding that CapEx-to-revenue ratio is 0.58x (4-6x peers) suggests Oracle may be 'buying' revenue growth through aggressive capacity deployment at sub-market returns. If even one quarter's capacity deployment slips or OpenAI contract terms shift, the growth rate is more fragile than the acceleration trend suggests. Still, below 40% remains a tail outcome given current momentum.

The base rate for this question should be anchored on: (1) the current 66% growth rate, (2) four quarters of acceleration, and (3) the capacity-constrained nature of growth. Historical cloud infrastructure deceleration is real (AWS 50%+ to 17%), but that deceleration played out over many quarters, not a single-quarter collapse. OCI's $4.1B quarterly base is still small relative to AWS ($27B+) and Azure ($18B+), meaning the law of large numbers is less binding. The key risk factor is Q2's 66% potentially including front-loaded contract activations, but even reverting to Q1's 54% growth would still be well above 40%. To breach 40%, multiple negative factors would need to converge simultaneously.

Four straight quarters of acceleration (51% to 66%) with capacity-constrained growth. GPU capacity growing 50% QoQ mechanically supports revenue growth. A drop to below 40% from 66% is extremely unlikely in one quarter absent a major disruption. The committee treated below-40% as a tail risk escalation trigger, not a likely outcome.

The acceleration trend and capacity constraints make deceleration below 40% unlikely, but the RPO conversion gap (12-month RPO +40% vs total +433%) and ~58% OpenAI concentration introduce meaningful tail risk. Historical deceleration patterns affect all cloud providers eventually. Below-market pricing compresses revenue per unit. Still a low probability event given the strong momentum.

The data strongly favors continued high growth: accelerating IaaS trajectory, capacity-constrained not demand-constrained, broadening multicloud adoption (+817% YoY). The threshold of 40% is far below current trajectory (66%). Some risk from OpenAI concentration and RPO conversion timing, but insufficient to make a 26pp decline probable in one quarter.