What tech stack does Visa use?

VisaNet — Visa's proprietary global payments network — is the crown jewel of the stack. It processes more than 65,000 transaction messages per second with 99.999%+ uptime using an active-active automated failover architecture across multiple global data centers, with no single point of failure. The core network relies on high-performance languages including Java, C/C++, Scala, and Go suited to microsecond-latency transaction processing. Python and JavaScript are used across data engineering, API layers, and internal tooling. Microservices communicate via REST/HTTP/JSON, with PACT used for contract testing across service boundaries. CI/CD pipelines run on Maven and Jenkins.

For infrastructure, Visa runs OpenStack to give internal developers private cloud IaaS capabilities without managing bare metal — enabling engineers to provision and deploy without worrying about the underlying hardware. Kubernetes orchestrates containerized microservices across this layer. AWS is used for select non-core workloads and experimentation. The Pismo acquisition (completed January 2024) added a cloud-native, API-first issuer processing and core banking layer, which Visa is now offering as a managed service to financial institutions globally. Job postings confirm MySQL and relational databases in use alongside distributed data infrastructure.

Visa runs one of the largest financial data processing environments in the world: Apache Kafka for real-time event streaming across billions of daily transactions, Apache Spark for large-scale batch and near-real-time analytics, and Hadoop/Hive for the data warehouse layer. DataHub is used for enterprise data governance — cataloging and lineage across Kafka, Spark, multiple databases, and BI tools (per a published DataHub customer story). Visa's AI infrastructure underpins VisaNet +AI and Visa Advanced Authorization (VAA), which analyze each transaction in real-time for fraud risk using models that run inference in milliseconds at 65K+ TPS scale. Job postings signal active investment in agentic AI solutions using tools including Cline, Claude (Anthropic), and GitHub Copilot — consistent with the June 2026 OpenAI Intelligent Commerce integration.

For frontend, Visa's developer portal and internal web applications are primarily React-based. Mobile applications use Swift (iOS) and Kotlin (Android) for native performance. On the GTM side, Salesforce CRM is in use for commercial and enterprise relationship management, providing an integration touchpoint for revenue intelligence platforms. Visa's developer ecosystem runs through the Visa Developer Platform (developer.visa.com), exposing REST APIs for tokenization, authorization, Visa Direct, and fraud services.

Visa's commitment to private/hybrid cloud (OpenStack + Kubernetes) over pure public cloud means infrastructure vendors that support on-premise or hybrid deployment models have a significant advantage over AWS/GCP/Azure-only SaaS. Visa cannot easily move VisaNet to a hyperscaler due to regulatory, latency, and data-sovereignty constraints — so any tooling (observability, security, data governance) must support private/hybrid deployment as a first-class mode.

The Kafka + Spark + Hadoop data stack is a natural integration point for data pipeline, streaming analytics, and AI/ML platform vendors. DataHub signals a mature data governance posture — Visa is receptive to enterprise data catalog, lineage, and compliance tooling that integrates with these systems. The Pismo acquisition signals that Visa is building an ecosystem of cloud-native financial infrastructure — vendors that interoperate with Pismo's API-first platform have a natural entry point. The Salesforce CRM signal means revenue operations and sales intelligence tooling has a clear integration path into Visa's existing GTM infrastructure. The active use of GitHub Copilot and agentic AI tools (Cline, Claude) in engineering signals receptivity to developer productivity and AI coding tooling — a high-velocity investment area given Visa's 34,100-person engineering-heavy workforce.