Which Hyperscaler Has the Highest RPO-to-Revenue Conversion Efficiency?
Period: Q4 2025 / Q1 2026
Overview
Microsoft and Alphabet lead hyperscaler RPO-to-revenue conversion efficiency at approximately 25% twelve-month recognition rates as of late 2025, while Oracle trails at just 12% following a massive influx of long-dated AI infrastructure contracts. However, the headline percentages mask a more nuanced story: the dollar value of near-term RPO conversion has actually increased across all four players, even as conversion rates have plummeted—because total RPO backlogs have exploded by 2-5x in under a year.
The Conversion Efficiency Scorecard
| Company | RPO Date | Total RPO ($B) | 12-Mo Recognition Rate | 12-Mo $ Recognized ($B) | RPO / Annual Rev |
|---|---|---|---|---|---|
| MSFT | Dec 31, 2025 | $631.0 | 25% | ~$157.8 | 2.1x |
| GOOGL | Dec 31, 2025 | $242.8 | ~25% (est.) | ~$60.7 | 4.1x (cloud) |
| AMZN | Dec 31, 2025 | $244.0 | ~24% (implied) | ~$59.5 | N/A (AWS only) |
| ORCL | Feb 28, 2026 | $552.6 | 12% | ~$66.3 | 8.6x |
Sources: MSFT 10-Q (Jan 2026), GOOGL 10-K (Feb 2026), AMZN 10-K (Feb 2026), ORCL 10-Q (Mar 2026)
Microsoft: Largest Backlog, Fastest Conversion
Microsoft's $631 billion RPO as of December 31, 2025 is the largest absolute backlog among the four, with a commercial RPO of $625 billion and a weighted-average duration of approximately 2.5 years. The 25% twelve-month recognition rate implies roughly $157.8 billion in near-term revenue locked in—already exceeding any single quarter's total company revenue of $81.3 billion (Q2 FY2026).
Critically, this 25% rate is a sharp decline from 40-45% just 12-18 months earlier. In September 2024, MSFT's RPO was $266 billion at a 45% conversion rate (yielding ~$119.7 billion in 12-month recognized revenue). By December 2025, the RPO more than doubled to $631 billion but the dollar amount earmarked for near-term recognition grew to $157.8 billion—a 32% increase in absolute terms. The denominator, not the numerator, changed. Massive multi-year Azure AI and Microsoft 365 Copilot commitments inflated total RPO while maintaining strong near-term fulfillment.
Alphabet: Steady Conversion Despite RPO Surge
Google Cloud's revenue backlog reached $242.8 billion by December 31, 2025—up 161% from $93.2 billion a year earlier. Alphabet expects to recognize "just over 50%" within 24 months, implying an estimated 25% annual conversion rate (~$60.7 billion over the next twelve months).
Compared to Google Cloud's $58.7 billion in CY2025 revenue, this means RPO near-term conversion roughly matches current annual run rate—a healthy signal of contracted revenue visibility. The RPO/Cloud Revenue ratio of 4.1x suggests approximately four years of cloud revenue are now contractually committed, up from approximately 2.2x a year ago ($93.2 billion RPO vs. $43.2 billion CY2024 cloud revenue).
Notably, Alphabet's conversion rate has been the most stable of the four hyperscalers, holding at approximately 50% over 24 months across multiple quarters—suggesting a more disciplined contract duration profile.
Amazon: Usage-Based Model Complicates Comparisons
Amazon's $244 billion RPO as of December 31, 2025 (up 38% from $177 billion a year earlier) applies only to contracts with original terms exceeding one year and is "primarily related to AWS." With a weighted-average remaining life of 4.1 years, the implied annual conversion rate is approximately 24%.
However, Amazon explicitly cautions that "the amount and timing of revenue recognition is largely driven by customer usage, which can extend beyond the original contractual term." This usage-based model makes AWS's conversion efficiency fundamentally different from the more predictable subscription-based RPO of MSFT and ORCL. Customers commit to minimum spend thresholds but actual consumption often exceeds contracted amounts, meaning AWS's effective conversion can exceed 100% of contracted minimums.
Oracle: Longest Duration, Lowest Conversion Rate
Oracle's RPO trajectory has been the most dramatic: from $97.3 billion in November 2024 to $552.6 billion by February 2026—a 468% increase in fifteen months. The twelve-month recognition rate collapsed from 39% to just 12%, driven by "certain significant cloud contracts" in the OCI (Oracle Cloud Infrastructure) space.
At 12%, Oracle expects to convert approximately $66.3 billion from its RPO over the next twelve months. Relative to its trailing four-quarter revenue of approximately $61 billion (FY26 Q1-Q2 + FY25 Q3-Q4), this implies RPO near-term conversion modestly exceeds current annual revenue—a bullish signal for revenue acceleration. However, the remaining 88% of RPO ($486 billion) is scheduled over 3-5+ years, creating the longest-duration backlog profile among the four hyperscalers.
The RPO-to-annual-revenue ratio of 8.6x dwarfs the competition, reflecting Oracle's strategy of locking in massive multi-year AI infrastructure commitments that will take nearly a decade to fully recognize.
The AI Contract Effect
The collapse in conversion rates across the board tells a consistent story: 2025 was the year of the mega AI infrastructure deal. All four hyperscalers signed unprecedented long-dated commitments, but the impact varied:
| Company | RPO 12-Mo Rate (Mid-2024) | RPO 12-Mo Rate (Late 2025) | Change |
|---|---|---|---|
| MSFT | 45% | 25% | -20 pp |
| ORCL | 39% | 12% | -27 pp |
| GOOGL | ~25% | ~25% | Stable |
| AMZN | ~25% | ~24% | Stable |
Microsoft and Oracle saw the most dramatic shifts, suggesting they absorbed the longest-duration AI contracts. Google and Amazon maintained steadier conversion profiles, potentially indicating shorter average contract terms or a more balanced mix of deal sizes.
Investment Implications
Microsoft leads on conversion efficiency with the highest absolute dollar recognition ($157.8 billion in 12 months) and a 25% rate that, while compressed, still reflects the fastest RPO-to-revenue flywheel among the group.
Oracle's 12% rate is a double-edged sword. It signals massive future revenue visibility (8.6x coverage) but raises execution risk: can Oracle actually build out OCI capacity fast enough to fulfill $486 billion in long-dated commitments? Capital expenditure discipline and datacenter build-out timelines become the binding constraints.
Google and Amazon offer the most predictable conversion at ~25%/~24%, with less volatility in their recognition profiles—attractive for investors prioritizing earnings visibility over backlog size.
What to Watch
- Oracle's capacity buildout: Monitor capex growth and datacenter announcements to assess whether $552.6 billion in RPO is achievable
- Microsoft's 12-month dollar recognition: If the ~$158 billion near-term figure grows in subsequent quarters, it validates conversion despite the lower rate
- Google Cloud's RPO acceleration: The jump from $157.7 billion (Sep 2025) to $242.8 billion (Dec 2025) suggests a potential rate shift ahead
- AWS usage vs. commitment: Watch for divergence between RPO growth and actual AWS revenue growth as an indicator of utilization rates
Sources: MSFT 10-Q filed Jan 28, 2026 (period ending Dec 31, 2025); ORCL 10-Q filed Mar 11, 2026 (period ending Feb 28, 2026); GOOGL 10-K filed Feb 5, 2026 (period ending Dec 31, 2025); AMZN 10-K filed Feb 6, 2026 (period ending Dec 31, 2025); MSFT 10-K filed Jul 30, 2025 (FY2025); ORCL 10-K filed Jun 18, 2025 (FY2025)