.png)
Neptune Serverless launched in general availability in October 2022. It applies the serverless capacity model to graph databases: instead of choosing a fixed instance type (db.r5.large, db.r6g.xlarge), you set a minimum and maximum NCU range and Neptune automatically scales the database’s compute capacity between those bounds in response to actual query load. You pay per NCU per second for the capacity in use, not for a fixed instance running 24/7.
The appeal is clear. Graph workloads — social network queries, fraud detection traversals, recommendation engine lookups — often have highly variable access patterns. A recommendation service might process thousands of concurrent graph traversals during a product launch and near-zero queries at 3am. A provisioned instance sized for peak means paying for peak capacity at 3am. Neptune Serverless pays for the actual NCUs consumed at each moment.
The nuance: Neptune Serverless is not universally cheaper. For workloads with consistent sustained query load, provisioned instances are significantly cheaper because the fixed per-hour instance rate is lower than the equivalent NCU cost at sustained high capacity. AWS’s own Database Blog explicitly states this. This guide explains exactly when Serverless saves money and exactly when it does not — with specific dollar examples.
See exactly what you’re overpaying in under 60 seconds. Try the Calculator for free →
What Is a Neptune Capacity Unit (NCU)?
A Neptune Capacity Unit (NCU) is the billing unit for Neptune Serverless. Each NCU provides approximately 2 GB of memory with a corresponding virtual CPU and networking capacity — similar to the resource allocation in provisioned Neptune instances.
Neptune Serverless scales the compute capacity of each database instance in real time. At any moment, the instance has a floating-point NCU value between the configured minimum and maximum. The billing is the current NCU count multiplied by the per-NCU hourly rate of $0.1098, calculated per second and invoiced hourly.
NCU equivalences with provisioned instance families (approximate, for planning purposes only):
4 NCUs = ~8 GB memory (roughly equivalent to db.r5.large at 16 GB is closer to 8 NCUs). 8 NCUs = ~16 GB memory. 16 NCUs = ~32 GB memory (roughly equivalent to a db.r5.xlarge). 32 NCUs = ~64 GB memory (roughly equivalent to db.r5.2xlarge). Source: AWS Database Blog (NCU equivalences are approximate — 1 NCU = ~2 GB memory per AWS official documentation).
Important: AWS’s own Database Blog explicitly states: ‘For steady state workloads, provisioned instances can be more cost effective. For example, choosing a provisioned db.r5.xlarge instance is more cost-effective than a serverless instance that remains at NCU=16 by setting the Min=Max NCU=16.’ This means if you set your minimum and maximum NCUs to the same value and run at that value consistently, you are almost certainly paying more than a provisioned instance would cost.
Neptune Serverless Pricing: Every Component
All rates: US East (N. Virginia), June 2026. Verify at aws.amazon.com/neptune/pricing — rates change.
| Component | Rate | Notes |
| NCU compute | $0.1098/NCU-hr | Billed per second. Floating-point NCU count between min and max. 1 NCU = ~2 GB memory. |
| Minimum NCU | 1.0 NCU | Reduced from 2.5 to 1.0 on March 1, 2023. At 1 NCU continuous: ~$80/month. Source: AWS official announcement. |
| Maximum NCU | 128 NCUs | Maximum configurable cap. Prevents unexpected cost spikes. Queries may be throttled if max is hit. |
| NCU increment | 0.5 NCU | Scales in 0.5 NCU steps. Fine-grained scaling means smooth cost curve. |
| Storage (Standard) | $0.10/GB-month | Plus $0.20/million I/O requests. Same rate as provisioned Standard. Storage independent of NCU scaling. |
| Storage (I/O-Optimized) | $0.225/GB-month | I/O included. Higher storage rate but no per-request I/O charges. For I/O-heavy traversals. |
| Backup storage (free tier) | Free to 100% of DB size | Automated backups free up to total DB storage size. Manual snapshots charged from creation. |
| Backup storage (above free) | $0.021/GB-month | Above free tier threshold. |
| T4g/T3 CPU credits (if applicable) | $0.15/vCPU-hr | Applies only if T4g or T3 serverless instances are used and exceed baseline CPU. Verify config. |
| Data transfer (same region) | $0.00 | No charge for data transfer between Neptune and other AWS services in the same region. |
| Free tier | None | Neptune has no free tier. Minimum cost: ~$80/month at 1 NCU continuous. 30-day trial for new customers only. |
NCU rate $0.1098/hr verified from wring.co (March 2026) and oreateai.com (February 2026) both citing AWS official. 1 NCU minimum from AWS official announcement March 1, 2023. Storage rates from AWS official Neptune pricing. Verify all at aws.amazon.com/neptune/pricing — rates change.
The Break-Even Analysis: When Serverless Is Cheaper and When It Is Not
AWS’s own documentation flags the key risk: setting min=max NCU at the same value (eliminating auto-scaling) and running at consistent load makes Neptune Serverless significantly more expensive than a provisioned instance. Here is the math.
Comparison: NCU=16 Serverless vs db.r5.xlarge Provisioned
Neptune Serverless at NCU=16, running continuously for a full month: 16 NCUs x $0.1098 x 730 hours = $1,282.42/month. Neptune Provisioned db.r5.xlarge (4 vCPU, 32 GB RAM) on-demand: approximately $0.696/hr x 730 = $508.08/month. Serverless at fixed NCU=16 costs 2.5x more than provisioned db.r5.xlarge. Source: NCU rate from wring.co (March 2026), provisioned r5.xlarge rate from wring.co (March 2026) citing AWS official pricing. Verify at aws.amazon.com/neptune/pricing.
This is the exact scenario the AWS Database Blog warns against: ‘choosing a provisioned db.r5.xlarge instance is more cost-effective than a serverless instance that remains at NCU=16 by setting the Min=Max NCU=16.’ The NCU hourly rate is designed for variable workloads where average NCU consumption is significantly lower than the maximum. When the database runs at maximum capacity continuously, the NCU billing is dramatically more expensive than the equivalent provisioned instance rate. Source: AWS Database Blog (2023, confirmed current in AWS documentation 2026).
When Serverless Is Cheaper: The Variable Workload Math
The same NCU=16 maximum cluster used above becomes cost-effective when average NCU consumption is low. Consider a graph application with clear day/night patterns: 10 NCUs average during 8 business hours per day, 1 NCU overnight and on weekends.
Business hours (8 hr/day x 22 days/month = 176 hours): 10 NCUs x $0.1098 x 176 = $193.25/month. Off-hours (730 – 176 = 554 hours): 1 NCU x $0.1098 x 554 = $60.83/month. Total serverless: $254.08/month. Provisioned db.r5.xlarge for the same cluster (must be sized for peak at 10 NCU equivalent): $508.08/month. Serverless saves approximately $254/month — 50% cost reduction for this variable workload pattern. Source: math derived from verified rates.
The key variable is the ratio of average NCU consumption to peak NCU. The higher the peak-to-average ratio (the more variable the workload), the greater the serverless saving. The lower the ratio (the steadier the workload), the more provisioned wins.
Break-Even NCU Utilization Rate
Serverless becomes cheaper than provisioned when the average NCU consumption across the full billing period is low enough that the total NCU cost is less than the provisioned instance cost. The break-even point varies by instance family. For db.r5.xlarge at $0.696/hr: serverless at sustained NCU consumption breaks even at 0.696/0.1098 = 6.34 NCUs average. Below 6.34 NCUs average over the month, serverless is cheaper. Above 6.34 NCUs average, provisioned is cheaper. Source: derived from verified rates.
Practical implication: if CloudWatch shows your Neptune Serverless cluster averaging above 60-65% of its configured maximum NCUs over the past 30 days, provisioned instances likely deliver better cost. If average utilization is below 40% of the maximum, serverless is likely cheaper. Source: analysis derived from break-even math above.
How Neptune Serverless Scaling Works
Neptune Serverless continuously monitors several resource dimensions on each database instance: CPU utilization, memory utilization, network throughput, and query queue depth. When the load reaches the current capacity ceiling or when Neptune detects performance degradation, the instance scales up automatically. When load declines, capacity scales down toward the configured minimum — but with a deliberate step-down pattern. Source: AWS official Neptune Serverless scaling documentation.
Scale-Up Behavior
Neptune Serverless scales up within seconds when query demand spikes. The scale-up is designed to be fast enough to handle sudden bursts of graph queries without visible latency degradation. For fraud detection systems that receive sudden bursts of transaction verification requests, or recommendation engines that spike during marketing campaigns, the fast scale-up means the serverless cluster handles the burst without pre-provisioning.
Scale-Down Behavior
Scale-down is more gradual than scale-up. Neptune releases CPU capacity before memory, allowing the database engine to retain cached graph data in memory during temporary lulls even while reducing CPU resources. This step-down architecture prevents oscillation — where the cluster repeatedly scales up and down in response to natural query volume variance — and ensures that cached data remains available for queries that arrive shortly after a quiet period.
Reader Instance Scaling
In a Neptune Serverless cluster with read replicas, each reader instance can be configured to scale independently from the writer instance or to scale together with it. Independent scaling is appropriate when read and write workloads have different patterns — for example, a writer handling consistent ingestion while readers serve bursty application query traffic. Coupled scaling keeps all instances at similar capacity levels, which simplifies capacity management for uniform workloads. The scaling configuration is set per instance using the promotion tier setting.
Neptune Serverless vs Neptune Provisioned: Complete Comparison
Both Neptune Serverless and Neptune Provisioned are fully managed, support the same query languages and features, and share the same storage layer. The choice is entirely about compute billing and capacity management.
| Factor | Neptune Serverless | Neptune Provisioned | Choose based on |
| Billing model | Per NCU per second. Floating cost. | Per instance-hour. Fixed cost. | Variable workload: Serverless. Steady: Provisioned. |
| Capacity management | Automatic. Set min/max bounds only. | Manual. Choose instance type. Add replicas. | Low ops overhead: Serverless. Precise control: Provisioned. |
| Cost at sustained high load | Higher. NCU rate x sustained NCUs. | Lower. Fixed instance rate, typically 2-3x cheaper than equivalent NCU cost. | Sustained high utilization: always Provisioned. |
| Cost at low or variable load | Lower. Scales down, you pay less. | Higher. Fixed rate whether idle or busy. | Variable, bursty, or dev workloads: Serverless. |
| Query languages | Gremlin, openCypher, SPARQL | Gremlin, openCypher, SPARQL | Identical. No difference. |
| Read replicas | Supported. Each replica also scales independently. | Supported. Up to 15 replicas. | Up to 15 replicas for both. |
| High availability | Supported. Multi-AZ with automatic failover. | Supported. Multi-AZ with automatic failover. | Identical HA behavior. |
| Storage | Same as Provisioned. NOT scaled by NCUs. | $0.10/GB-month + I/O or $0.225/GB-month I/O-Optimized. | Identical storage billing. |
| Discount options | Database Savings Plans only. No Reserved Instances. | Reserved Instances (up to 45%) or Database Savings Plans (up to 35%). | Deep discount seekers: Provisioned + RI. Flexibility seekers: DSP. |
| New application sizing | Best. No sizing required. Set wide bounds and observe. | Harder. Must estimate instance size upfront. | Unknown workload: start Serverless, migrate to Provisioned once pattern is known. |
| Min monthly cost | ~$80/month at 1 NCU continuous (plus storage). | ~$254/month for db.r5.large on-demand (plus storage). | Serverless is cheaper for low-traffic graphs. |
Source: wring.co (March 2026), AWS Database Blog, AWS official Neptune documentation. Provisioned db.r5.large rate ~$0.348/hr from wring.co March 2026. Verify all at aws.amazon.com/neptune/pricing.
Also read: AWS Neptune Pricing: Complete Cost Guide for 2026Â
When to Use Neptune Serverless: The Four Best Scenarios
Scenario 1: New Graph Applications With Unknown Workload
When building a new graph application, it is often difficult to estimate the required graph database instance size. The optimal instance depends on graph density (how many edges per node), query complexity (how many hops, how many concurrent traversals), and traffic patterns — all of which are hard to forecast before the application is live.
Neptune Serverless eliminates this guessing: set a low minimum (1-2.5 NCUs) and a generous maximum (32-64 NCUs), deploy the application, and observe actual NCU consumption from CloudWatch (ServerlessDatabaseCapacity metric). After 30-60 days of production traffic, you have real data on your workload’s NCU requirements. If NCU consumption is consistently high and steady, migrate to a provisioned instance at the right size. If NCU consumption is variable, stay on Serverless.
Scenario 2: Development and Test Environments
Development and testing graph databases have intermittent usage: active during business hours when engineers are running queries, idle overnight and over weekends. The 1 NCU minimum means the dev cluster costs approximately $0.1098/hr when idle — $2.64/day overnight versus the full provisioned instance rate.
For a team with 5 developers running graph workloads 8 hours per day, 22 workdays per month: active (5 NCUs x 8 x 22 x $0.1098) + idle (1 NCU x (730-176) x $0.1098) = $96.57 + $60.83 = $157.40/month. Equivalent provisioned db.r5.xlarge: $508.08/month on-demand. Neptune Serverless saves approximately $350/month for this development environment.
Scenario 3: Multi-Tenant Graph Applications
Applications serving multiple customers (SaaS platforms, enterprise analytics tools) sometimes require data isolation with separate Neptune clusters per customer for security or compliance reasons. With provisioned instances, provisioning a minimum db.r5.large per customer at $0.348/hr means $254/month per customer cluster even for customers with minimal usage.
Neptune Serverless at 1 NCU minimum per customer cluster: $80/month per cluster for light users. As a customer’s usage grows, the serverless cluster scales up automatically. Customers with heavy graph workloads naturally cost more — the billing tracks actual usage rather than a uniform fixed capacity per customer.
Scenario 4: Spiky Event-Driven Graph Workloads
Graph workloads triggered by external events — marketing campaigns that spike social graph traversals, security incident investigations that trigger network topology analysis, sports events that generate rapid social connection queries — have peak demand that is 10-100x higher than baseline. Provisioning for peak means massive over-provisioning at baseline.
Neptune Serverless handles this naturally: the cluster scales to the required NCU count within seconds of the spike, processes the burst at full capacity, and scales back down when the event ends. For a cluster that needs 64 NCUs for 2 hours per month (a quarterly business review, a product launch event) and 2 NCUs the rest of the time: (64 x $0.1098 x 2) + (2 x $0.1098 x 728) = $14.06 + $159.91 = $173.97/month. A provisioned instance sized for peak (64 NCUs equivalent is roughly db.r5.4xlarge at $2.784/hr): $2,032.32/month. Serverless saves $1,858/month for this extremely spiky pattern. Source: math derived from verified rates.
Configuring Neptune Serverless: Min, Max, and the Cost Implications
Setting Minimum NCUs
The minimum NCU determines the lowest capacity the serverless cluster will maintain. At minimum, the cluster is always ready to respond to queries without a cold-start delay. A minimum of 1 NCU ensures the database can handle low-volume queries immediately. A minimum of 2.5 NCUs is the pre-March 2023 default and is still recommended by AWS for production workloads that need consistent sub-second query response times even at low traffic levels. Source: wring.co (March 2026) citing AWS documentation.
Cost impact of minimum NCU choice: at 1 NCU minimum, continuous idle cost is $0.1098 x 1 x 730 = $80.15/month. At 2.5 NCU minimum: $0.1098 x 2.5 x 730 = $200.39/month. The difference is $120/month — material for development environments where idle cost is the primary driver.
Setting Maximum NCUs
The maximum NCU acts as a cost ceiling. Setting maximum too low means the cluster is throttled at peak, potentially degrading query performance. Setting maximum too high does not cost more (you only pay for NCUs actually consumed) but removes cost predictability since an unexpected workload spike or runaway query could temporarily scale to a very high NCU count.
Recommended approach: set maximum to 2-4x your expected peak NCU consumption, observed from load testing. For a new application where peak is unknown, set a conservative maximum (16-32 NCUs) initially. Monitor NCUUtilization in CloudWatch. If NCUUtilization consistently hits 80-90%, the maximum is too low. Increase the maximum and re-evaluate. Source: AWS Neptune Serverless capacity scaling documentation.
Cost control tip: the maximum NCU setting is your only hard cost cap for compute in Neptune Serverless. Unlike some serverless services that have no configurable spending cap, Neptune Serverless respects the maximum you set — queries may be throttled at peak if the maximum is hit, but compute charges will not exceed (max NCU x $0.1098 x hours). Set the maximum at a level where the worst-case hourly cost is acceptable to your budget. Source: AWS official Neptune Serverless documentation.
Storage for Neptune Serverless: Standard vs I/O-Optimized
Neptune Serverless uses the same underlying distributed storage as Neptune Provisioned — the cluster volume that automatically grows in 10 GB increments and can reach up to 128 TiB. Storage charges are identical regardless of whether the compute layer is serverless or provisioned. Storage scaling is completely independent of NCU scaling. Source: AWS Database Blog.
Standard Storage for Serverless
Storage: $0.10/GB-month. I/O: $0.20 per million requests, billed separately. For graph applications with moderate I/O — property graph lookups with focused traversals, social graph queries that follow a bounded number of edges per query — Standard storage is typically cheaper because total I/O volume stays moderate relative to data size.
I/O-Optimized Storage for Serverless
Storage: $0.225/GB-month. I/O: included, no per-request charge. I/O-Optimized is worth considering for Neptune Serverless clusters running traversal-heavy workloads where each query follows many edges — fraud ring detection traversals across multiple hops, network topology analysis over complex dependency graphs, multi-hop recommendation queries. For these patterns, I/O charges under Standard storage can exceed the storage rate differential of I/O-Optimized.
The break-even for I/O-Optimized: if your I/O charges under Standard would exceed $0.125/GB-month (the premium over Standard storage), I/O-Optimized is cheaper. This translates to approximately 625,000 I/O requests per GB of storage per month as the crossover. Monitor Neptune’s CloudWatch VolumeReadIOPs and VolumeWriteIOPs metrics to evaluate your actual I/O rate before switching configurations.
Also read:
Database Savings Plans for Neptune Serverless
Reserved Instances are not available for Neptune Serverless — Reserved Instances apply only to Neptune Provisioned instances. The commitment-based discount mechanism for Neptune Serverless is Database Savings Plans.
Database Savings Plans cover Neptune Serverless with up to 35% savings on a 1-year commitment. The commitment is a dollar-per-hour spend level rather than a specific NCU range or instance type. This means the DSP discount applies regardless of how the serverless cluster scales during the commitment period — whether it is at 1 NCU or 128 NCUs, the committed hourly spend is counted against the DSP.
Sizing a DSP commitment for Neptune Serverless: the correct commitment level is the floor of your consistent hourly Neptune Serverless spend — the level your costs never drop below across all hours of the month. For a cluster that averages 4 NCUs at $0.44/hr most hours but drops to 1 NCU at $0.11/hr overnight and weekends, the consistent floor is approximately $0.11/hr. A DSP commitment at that floor captures 35% savings on the baseline spend while the variable portion above the floor remains on on-demand NCU pricing.
Database Savings Plans can cover multiple Neptune clusters and other database services simultaneously under one commitment. For teams running both Neptune Serverless clusters and other managed databases (ElastiCache, RDS, DynamoDB), a single DSP commitment can be sized for the combined database infrastructure floor spend. Source: Usage.ai live Database Savings Plans guide.
Also read: AWS Database Savings Plans: Complete Guide
Neptune Serverless Features and Limitations
What Neptune Serverless Supports
All three graph query languages: Gremlin (Apache TinkerPop), openCypher, and SPARQL 1.1. Source: AWS official Neptune Serverless product page. Read replicas for horizontal read scaling and high availability. Multi-AZ deployment with automatic failover to replicas in a different Availability Zone. Neptune ML integration (graph neural network training via SageMaker). Neptune Global Database (cross-region replication). Point-in-time recovery (PITR) for backup and restore. Performance Insights for query-level monitoring. All standard Neptune security features: VPC isolation, encryption at rest and in transit, IAM authentication, and advanced auditing. Source: AWS official Neptune Serverless documentation.
What Neptune Serverless Does Not Support
Neptune Serverless does not support Custom Endpoints for routing specific query types to specific instance groups. It does not support Neptune Labs experimental features. It has some limitations with specific instance types that provisioned supports but Serverless does not — verify the current limitations in the AWS Neptune Serverless User Guide (docs.aws.amazon.com/neptune/latest/userguide/neptune-serverless.html) before migrating a provisioned cluster to Serverless. Source: AWS official Neptune documentation.
Neptune Serverless Engine Version Requirements
Neptune Serverless requires Neptune engine version 1.2.0.0 or higher. If you have an existing Neptune provisioned cluster running an older engine version, upgrade the engine version before migrating to Serverless. Engine version upgrades require a cluster reboot. Source: AWS official Neptune Serverless documentation.
See exactly what you’re overpaying in under 60 seconds. Try the Calculator for free →
How Usage.ai Handles Neptune Serverless Cost Optimization
Usage.ai analyzes Neptune Serverless clusters using 30 days of CloudWatch ServerlessDatabaseCapacity and NCUUtilization data. The platform identifies two primary optimization signals: clusters where average NCU utilization relative to the configured maximum is consistently high (suggesting the cluster would be cheaper on provisioned instances), and clusters where DSP commitment would deliver savings on the stable floor of NCU spend.
For clusters where the average NCU consumption is approaching or exceeding the provisioned break-even point (approximately 60-65% of the maximum NCUs sustained), Usage.ai surfaces the migration-to-provisioned recommendation with the specific dollar saving calculation. The platform accounts for the sequencing: confirm the target provisioned instance size from serverless NCU data, test for 30 days on the new provisioned instance, then purchase Reserved Instances or DSP for the confirmed configuration.
For Neptune Serverless clusters with stable minimum spend patterns, Usage.ai evaluates the DSP commitment sizing. For teams running multiple Neptune clusters (serverless and provisioned) or multiple database services, the platform evaluates whether a single combined DSP commitment covers the total database infrastructure floor more efficiently than separate per-cluster commitments.
If any DSP commitment for Neptune Serverless becomes underutilized — because a cluster is decommissioned, migrated to provisioned, or workload traffic drops permanently — Usage.ai provides cashback on the unused commitment in real money. Fee: percentage of realized savings only.
Frequently Asked Questions
1. What is Neptune Serverless?
Amazon Neptune Serverless is a deployment option for Neptune Database that automatically scales graph database compute capacity between a configurable minimum (1 NCU) and maximum (up to 128 NCUs) in response to actual query load. You pay $0.1098 per NCU-hour, billed per second. 1 NCU provides approximately 2 GB of memory with corresponding CPU and networking. Neptune Serverless is production-ready and supports all three graph query languages (Gremlin, openCypher, SPARQL), read replicas, Multi-AZ HA, and all standard Neptune features. Source: AWS official Neptune Serverless documentation.
2. How much does Neptune Serverless cost?
$0.1098 per NCU-hour, billed per second. At 1 NCU minimum running continuously: approximately $80/month in compute. Storage billed separately at $0.10/GB-month (Standard, plus $0.20/million I/O) or $0.225/GB-month (I/O-Optimized, I/O included). Backup storage: free up to 100% of DB size, $0.021/GB-month above. No free tier. The minimum cost for any Neptune cluster (serverless or provisioned) is approximately $80/month at 1 NCU. Source: wring.co March 2026 citing AWS official. Verify at aws.amazon.com/neptune/pricing.
3. When is Neptune Serverless more expensive than provisioned?
When average NCU utilization is consistently high and the NCU cost exceeds the provisioned equivalent rate. At sustained NCU=16: serverless costs 16 x $0.1098 x 730 = $1,282/month versus provisioned db.r5.xlarge at approximately $508/month — serverless is 2.5x more expensive. AWS Database Blog states: ‘For steady state workloads, provisioned instances can be more cost effective.’ The break-even is approximately 6.34 NCUs average sustained for db.r5.xlarge equivalence. Above that average, provisioned is cheaper. Source: AWS Database Blog, wring.co (March 2026).
4. What is the minimum NCU for Neptune Serverless?
1.0 NCU, since March 1, 2023. AWS reduced the minimum from 2.5 NCUs to 1.0 NCU on that date, reducing the minimum cost by up to 2.5x for clusters spending significant time at low utilization. Source: AWS official announcement (aws.amazon.com/about-aws/whats-new/2023/03/amazon-neptune-serverless-scales-down-1-ncu-costs/).
5. Do Reserved Instances apply to Neptune Serverless?
No. Reserved Instances apply only to Neptune Provisioned instances. The discount mechanism for Neptune Serverless is Database Savings Plans, which provide up to 35% savings on a 1-year commitment. DSP applies based on a dollar-per-hour spend commitment, not tied to a specific NCU range. Source: AWS official Neptune pricing page and Usage.ai live DSP blog.
6. What query languages does Neptune Serverless support?
Neptune Serverless supports all three graph query languages: Apache TinkerPop Gremlin, openCypher, and SPARQL 1.1. It supports the same features and SLAs as Neptune Provisioned. Applications already using Neptune Provisioned can migrate to Serverless without any query language or application code changes. Source: AWS official Neptune Serverless product page.
7. How many regions is Neptune Serverless available in?
19 AWS Regions as of November 2024. AWS expanded Neptune Serverless to 6 additional regions on November 12, 2024: Europe (Paris), South America (Sao Paulo), Asia Pacific (Jakarta), Asia Pacific (Mumbai), Asia Pacific (Hong Kong), and Asia Pacific (Seoul). The original 12 regions at GA (October 2022) plus 1 additional in December 2022. Source: AWS official announcements (October 2022, December 2022, November 2024).
