Neon Announces Up to 25% Reduction in Compute Prices

Company

Neon significantly cuts compute costs by up to 25% across plans, leveraging Databricks' global infrastructure for enhanced efficiency and superior price-performance.

Nov 03, 2025

Major Compute Price Reduction on Neon

Reducing the cost of compute by 25% as we scale on Databricks

Databases are frequently one of the largest infrastructure expenses for any company. From day one, Neon's mission has been to make databases radically more efficient through the separation of storage and compute, enabling instant autoscaling and superior unit economics.

Now, with Neon running on Databricks' global infrastructure, we've taken another significant step forward: compute on Neon is now up to 25% cheaper across all plans.

What's Changing?

Starting today:

  • Launch Plan: $0.14 \u2192 $0.106 per CU-hour
  • Scale Plan: $0.26 \u2192 $0.222 per CU-hour

This continues a series of strategic cost reductions since joining Databricks earlier this year:

  • August: Minimum spend dropped to $5/month, and storage fell from $1.75 \u2192 $0.35 per GB-month.
  • September: Enterprise features such as Private Link, SLA, SOC 2, HIPAA, SSO, and Point-in-Time Recovery (PITR) became integral parts of the Scale Plan, eliminating the previous $700/month surcharge.
  • October: The Free Plan's compute allowance doubled from 50 \u2192 100 CU-hours.

With these continuous improvements, Neon now provides the best price-performance ratio across every stage of development, from prototyping with the Free Plan to full production workloads on the Launch and Scale Plans.

Real-world Examples

Comparing the total ownership costs of databases is notoriously difficult due to their highly scenario-dependent nature. It becomes even more challenging when comparing an autoscaling system like Neon (which can scale up to larger resources and down to zero in hundreds of milliseconds) with fixed-capacity databases.

In this section, we'll price a few real-world scenarios and compare Neon with two other popular database options: Aurora Serverless v2, known for its slower autoscaling, and Supabase, which offers fixed capacity.

Scenario 1. Entry-Level Database (0.25CU, 1GB, 9h/day usage)

This scenario is for building a low-stakes application with relatively low usage. The database operates on the leanest compute resources: 0.25 CU on Neon, Micro on Supabase, and 0.5 ACU on Aurora. Neon's autoscaling capability can swiftly scale the database down to zero when inactive and scale up in mere milliseconds when load demands it. The cost of running this database on Neon is only $7.66/month, less than half of Aurora Serverless and 30% of Supabase. Beyond cost savings, you also benefit from point-in-time recovery, snapshots, and data durability.

  • Entry-Level Workload: Small, production-grade database suitable for an internal app or side project.

    Chart Details

    FeatureNeonSupabaseAurora Serverless
    Best Price$0/mo$0/mo$0/mo
    Production
    \u2003Compute$0$0$0
    \u2003Storage$0$0$0
    \u2003I/O$0
    Non-Production
    \u2003Dev/Staging$0$0$0
    Add-ons
    \u2003SOC2$0$0$0
    \u2003HIPAA$0$0$0

    Workload Details: Database needs minimal CPU and Memory, is only actively queried ~9 hours/day, and can run on the smallest size on each platform: 0.25 CU on Neon, Micro on Supabase, and 0.5 ACU on Aurora.

    • Storage (GB): 1 GB
    • Scale-to-zero enabled: Pause compute when inactive
    • Dev Branches: 0
    • SOC2: Yes
    • HIPAA: Yes

    Note: If your entry-level project requires SOC2 Compliance, the savings are extreme because we don't charge a high fixed fee for premium features, whereas Supabase requires you to opt into their $599/month business plan.

Scenario 2. Launching a New Product (1 to 4CU, 20GB, always on)

For a lean startup launching a new product, guaranteeing responsiveness is key. On Neon, you would create a Launch Plan database with 1 to 4 CU autoscaling, comparable to 2 to 8 ACU on Aurora (where 1 CU in Neon offers similar resources to 2 ACU in Aurora), or an XL instance on Supabase. In this scenario, Neon's cost is less than 40% of Aurora and less than 70% of Supabase.

  • Startup Workload: Mid-size database with 24/7 variable load for a fast-moving startup.

    Chart Details

    FeatureNeonSupabaseAurora Serverless
    Best Price$0/mo$0/mo$0/mo
    Production
    \u2003Compute$0$0$0
    \u2003Storage$0$0$0
    \u2003I/O$0
    Non-Production
    \u2003Dev/Staging$0$0$0
    Add-ons
    \u2003SOC2$0$0$0
    \u2003HIPAA$0$0$0

    Workload Details: Database has non-stop load that varies throughout the day/week, requiring moderate compute resources with 1-4 CU on Neon, an XL instance on Supabase, and 2-8 ACU on Aurora to handle typical workload demands.

    • Storage (GB): 20 GB
    • Dev Branches: 5
    • SOC2: Yes
    • HIPAA: Yes

Scenario 3. Scaling Your Business (4 to 10CU, 100GB, always on)

Having achieved product-market fit, you're now investing in scaling both your company and product. As you enter new market segments, compliance requirements emerge. You upgrade to Neon's most comprehensive tier, which includes all features. Even in this advanced scenario, Neon remains substantially more affordable than the other two alternatives.

  • Scaling Workload: Large database with high traffic and enterprise requirements for a rapidly scaling company.

    Chart Details

    FeatureNeonSupabaseAurora Serverless
    Best Price$0/mo$0/mo$0/mo
    Production
    \u2003Compute$0$0$0
    \u2003Storage$0$0$0
    \u2003I/O$0
    Non-Production
    \u2003Dev/Staging$0$0$0
    Add-ons
    \u2003SOC2$0$0$0
    \u2003HIPAA$0$0$0

    Workload Details: Database has non-stop load that varies across the day/week. It needs substantial compute capacity with 4-10 CU on Neon, a 4XL instance on Supabase, and 8-20 ACU on Aurora to support growing user demands.

    • Storage (GB): 100 GB
    • Dev Branches: 15
    • SOC2: Yes
    • HIPAA: Yes

The scenarios presented in this section clearly demonstrate that Neon offers the lowest cost, regardless of your current development stage or scale. What's even better is the ease with which you can transition between these different scenarios with just a few clicks.

Why We Can Afford to Lower Prices

You might wonder: How can we make databases this affordable? There's no catch. As alluded to earlier, there are two fundamental reasons why we can offer the best price-performance.

First, Neon's separation of storage and compute means your database scales up when it's busy and down to zero when it's idle. Data remains safely stored on durable cloud object storage, while compute resources spin up in milliseconds. You only pay for what you use. This stands in stark contrast to other database providers offering fixed capacity, where you often need to provision for peak workloads and end up paying for significant idle time.

Second, since joining Databricks, Neon now operates on one of the world's largest and most efficient multi-cloud infrastructures. Databricks launches tens of millions of VMs daily, and this unparalleled scale directly translates into lower operational costs for Neon users.

What's Next

Choosing a database is one of the most critical architectural decisions. We are humbled by the trust Neon users have placed in us. With features like autoscaling, instant branching, and point-in-time recovery, we provide an unparalleled developer experience and can dramatically reduce your database's operational complexity. Coupled with consecutive price reductions, we now offer the leading price-performance option in the database market.

Our work is never finished. We are building Neon not just to excel in one metric, but to be the best in all aspects, making Neon and Databricks synonymous with OLTP innovation for years to come.

Posted by

Nikita Shamgunov CEO at Neon

  • Practical Guide to Database Branching Savannah Longoria

  • Just Launched: Neon Is Now a Kiro Power Carlota Soto

  • Improving DNS performance with NodeLocalDNS Luca Cittadini

Share:

Subscribe to our changelog. No spam, guaranteed. Subscribe

Share:

More from Neon

  • Product Dec 04, 2025 Practical Guide to Database Branching Savannah Longoria

  • AI Dec 03, 2025 Just Launched: Neon Is Now a Kiro Power Carlota Soto, Krishna B. Parab

  • Engineering Dec 01, 2025 Improving DNS performance with NodeLocalDNS Luca Cittadini, Lassi Pölönen