What tech stack does JPMorgan Chase use?

Java remains JPMorgan Chase's dominant backend language, used across trading systems, payments processing, and core banking infrastructure — confirmed by numerous Software Engineer III (Java/Kotlin) job postings and open-source projects on github.com/jpmorganchase. Python has become the primary language for data engineering, AI/ML workflows, and quantitative analysis. Kotlin is increasingly used for both Android mobile development and backend services. Node.js appears in job postings for API gateway and frontend-adjacent services.

On infrastructure, AWS is the primary public cloud partner: JPMorgan has migrated more than 6,000 applications and 1 exabyte of data to AWS, including its data lake (migrated from Hadoop to AWS S3 and Glue using a federated data mesh architecture). The firm runs AWS Graviton-based compute instances for cost and performance efficiency, and serves Chase.com through AWS CloudFront. A hybrid multi-cloud strategy with Azure and Google Cloud mitigates concentration risk — an explicit regulatory concern for systemically important financial institutions. Kubernetes and Docker underpin the microservices architecture, and Apache Kafka handles real-time event streaming across trading and payments. ServiceNow is confirmed for ITSM via senior ServiceNow architect and developer job postings on JPMorgan's careers site.

JPMorgan's engineering blog (Next at Chase) has published extensively on the firm's platform engineering patterns, including its API gateway strategy, internal developer platform investments, and the shift from monolithic legacy mainframe systems to microservices — a migration that spans decades of Java-era rewrites of COBOL-era banking infrastructure.

On the frontend, React and TypeScript power Chase.com and a wide range of internal tools — the site now serves through AWS CloudFront with an average of 15 releases per week, reflecting a mature continuous deployment pipeline. Mobile apps are built with Swift (iOS) and Kotlin (Android). JPMorgan's open-source GitHub portfolio (github.com/jpmorganchase) includes React component analytics tooling, Python serialization libraries, and open-source contributions to the broader financial technology ecosystem.

For data, the firm has built a data mesh architecture that enables federated data ownership across business domains while centralizing infrastructure on AWS S3 and Glue. This replaced a monolithic Hadoop-era data lake and supports the firm's ambition to have every data product independently versioned, discoverable, and governed to regulatory standards. AI is now a firm-wide priority: CDAO Teresa Heitsenrether leads the AI Center of Excellence, deploying LLM-powered agents for trading research, compliance document analysis, customer service automation (in Chase contact centers), and code generation for software engineers. JPMorgan filed over 3,500 AI-related patents as of 2024, signaling deep internal AI investment and development rather than pure vendor procurement.

The generative AI program is noteworthy in its scope: internal AI assistants are deployed to consumer-facing Chase agents, investment banking research analysts (for document summarization), and software engineers (for code completion and review). The firm has partnered with multiple AI infrastructure providers while maintaining a build-first philosophy for mission-critical AI capabilities — particularly in trading and risk management.

JPMorgan's commitment to AWS, multi-cloud Kubernetes orchestration, and Java/Python creates clear integration angles for vendors. Cloud security, secrets management, container security, and FinOps tools that operate natively on AWS and Kubernetes are in active demand. Data governance, data cataloging, and data quality platforms that layer onto AWS Glue and S3 fit the firm's data mesh architecture and are natural conversation starters with the CDAO organization.

For AI vendors: JPMorgan's generative AI program is large and active, but the firm prefers enterprise contracts with enterprise-grade security, audit trails, data residency controls, and model explainability — purely consumer-grade AI tools will not pass TPRM review. The firm also has a strong build-versus-buy instinct for mission-critical AI (it prefers to fine-tune and deploy its own models where possible), so AI vendors must position around either infrastructure (serving, evaluation, observability) or domain-specific applications where the firm lacks internal expertise.

ServiceNow's confirmed deployment means ITSM-adjacent vendors face both a strong incumbent and a potential integration hook — products that extend ServiceNow workflows (security operations, HR service delivery, asset management) are logical expansion plays. Build-versus-buy posture at JPMorgan skews toward build for mission-critical systems (trading infrastructure, core banking, risk management) and buy for operational and productivity tooling. The key implication: don't compete with JPMorgan's internal engineering organization on core financial software; instead, sell the horizontal platform capabilities (security, observability, developer experience, data management) that the firm buys across all business lines.