AWS Partner Network (APN) Blog
How Nubank saved $15M on Amazon DynamoDB costs with PointFive
By: Amir Hazoom, Partner Solutions Architect – AWS
By: Rob McCauley, Specialist Solutions Architect – AWS
By: Mike Rosenberg, Director of Cloud Financial Management – Nubank
By: Amir Hozez, Co-Founder & CTO – PointFive
 |
| PointFive |
 |
Nubank saved over $15 million on Amazon DynamoDB costs by partnering with PointFive while scaling from 99 million to 119 million customers in 16 months. This post describes how Nubank and PointFive achieved measurable savings while continuing to scale by building on Amazon Web Services (AWS). Nubank uses DynamoDB to handle billions of real-time financial transactions with consistent, single-digit millisecond performance. PointFive’s agentic optimization analyzed thousands of DynamoDB tables, and Nubank turned the findings into verified annualized savings.
Nubank partnered with PointFive to manage cloud efficiency and maintain its low cost-to-serve. Running on AWS, they uncovered and remediated inefficiencies across thousands of DynamoDB tables. This drove measurable savings and reinforced the value of financial operations (FinOps) maturity for teams running on AWS.
The following diagram illustrates how DynamoDB infrastructure, PointFive’s agentic optimization, and Nubank’s FinOps discipline work together on AWS to deliver verified savings at scale:
Figure 1 – How DynamoDB infrastructure, agentic optimization, and FinOps discipline work together on AWS to deliver verified savings at scale
The urgency of cloud optimization
Cloud spending continues to accelerate. The FinOps Foundation’s State of FinOps 2026 report finds that 78% of FinOps practices now report into the CTO or CIO organization, signaling that cloud cost management has become a boardroom priority. The next challenge is operationalizing it at enterprise scale. At the same time, enterprises are adopting AI workloads on AWS. For Nubank, optimizing cloud spend on existing services freed budget to reinvest in AI services such as Amazon Bedrock and Amazon SageMaker AI.
The DynamoDB 2024 pricing update reduced on-demand throughput costs by up to 50%, giving customers a strong reason to revisit their database configurations. Combined with AI-driven optimization, this is the right time to act.
DynamoDB as Nubank’s foundation
DynamoDB provides the backbone for Nubank’s real-time financial systems. As a fully managed, serverless database, it delivers the performance, security, and resilience that real-time financial transactions require at scale.
DynamoDB capabilities that supported Nubank’s success include single-digit millisecond performance at high scale and multi-Region strong consistency for resilience. The 2024 pricing updates further reduced on-demand throughput costs by up to 50%.
“It’s a bulletproof database that grows as fast as we do.”
– Mike Rosenberg, Nubank
The challenge: Visibility and waste at scale
With thousands of DynamoDB tables spread across engineering teams, identifying unused or inefficient resources manually was unfeasible. Even marginal inefficiencies could add up to significant costs at Nubank’s scale.
“No one raindrop believes it’s responsible for the flood. PointFive helped us see each table, the raindrops, that added up to costs without business value.”
– Mike Rosenberg, Nubank
The following screenshot shows how PointFive surfaces DynamoDB optimization opportunities, with specific savings estimates and recommended actions for each table:
Figure 2 – PointFive displaying DynamoDB optimization opportunities for Nubank, including inactive table detection, storage class recommendations, and capacity right-sizing, each with estimated savings and a direct remediation path
The solution: A three-way partnership in action
PointFive deploys from AWS Marketplace with no infrastructure changes required. Within 24 hours of onboarding, PointFive identified actionable optimization opportunities for Nubank’s DynamoDB environment. Its agentic, architecture-level analysis found inefficiencies that rule-based tools missed.
Three categories drove the majority of Nubank’s savings:
- Inactive table cleanup – Tables no longer in active use were identified and removed, leading to immediate savings.
- Storage class shifts – Low-access tables were moved to the DynamoDB Standard-Infrequent Access (Standard-IA) table class, cutting storage costs by up to 60% on affected workloads according to Nubank.
- Capacity right-sizing – Each remaining table was matched to the most cost-efficient throughput model for its actual usage pattern.
Each recommendation flowed directly into Nubank’s existing engineering workflow through a Jira integration. Findings became finished work, not tickets in a backlog.
“For banks and financial services customers, disaster recovery and data durability are required. DynamoDB capabilities like multi-Region strong consistency and point-in-time recovery give customers like Nubank essential resilience while freeing budget to reinvest in further innovation.”
– Rob McCauley, Amazon DynamoDB
Measurable results
The Nubank and PointFive collaboration delivered the following outcomes:
- Over $15 million in verified annualized savings while scaling by 20 million new customers over 16 months, as reported by Nubank
- Ten-day return on investment (ROI) from initial deployment to first realized savings, based on Nubank’s data
- Reinvestment in resilience: A portion of recovered spend went toward DynamoDB point-in-time recovery, strengthening Nubank’s durability posture
The following graphic summarizes the measurable outcomes Nubank achieved over 16 months:
Figure 3 – Nubank’s measurable outcomes from the AWS and PointFive collaboration include over $15 million in verified annualized savings, 119 million customers served, and ROI achieved within 10 days of deployment
“Optimization isn’t just a one-time fix. It’s a posture. Our solution continuously learns from DynamoDB telemetry to keep customers efficient as they grow.”
– Amir Hozez, Co-founder and CTO, PointFive
The takeaway: Better together
This collaboration demonstrated that enterprises can optimize cost and strengthen resilience simultaneously. DynamoDB provided the technical foundation, PointFive’s agentic analysis surfaced the opportunities, and Nubank’s FinOps discipline turned them into outcomes. The result is more than savings—it’s headroom to reinvest in resilience and in the AI workloads that will define the next wave of cloud innovation. This pattern is repeatable: Find the waste, fund the future.
Related resources
-
- Accelerate vendor evaluation with AWS specializations – APN Blog
- Best practices for cost optimization – AWS Well-Architected Framework
About PointFive
PointFive is an AWS Advanced Tier Independent Software Vendor (ISV) Partner with the AWS Cloud Operations Competency and Cloud Financial Management designation. PointFive helps enterprises identify and eliminate cloud waste across every layer of their stack, from cloud infrastructure to AI workloads, through agentic optimization available in AWS Marketplace.