New Way Now: PayPal paving the way for the future of agentic commerce with Google Cloud

Srinivasan Manoharan, Director of AI/ML Platform at PayPal

Executive summary: Digital payments leader PayPal is pioneering the next era of digital commerce, moving from simple chatbots to production-grade agents with Google Cloud. After migrating over 300 petabytes of data to BigQuery in one of the largest data transformations in the industry, PayPal built its entire AI stack on Google Cloud’s AI infrastructure, including Google Kubernetes Engine (GKE) to gain the power and flexibility for scaling complex, multi-step agent workflows. In addition, PayPal developed a unified large language model (LLM) stack on Gemini Enterprise Agent Platform with all the models, tools, and infrastructure needed to build, scale, govern, and manage agents. This unified approach has already made an impact, delivering 99.999% availability for critical AI workloads, reducing model deployment from weeks to minutes, and empowering 10,000 engineers to focus on creating agentic commerce capabilities over 400 million users can trust — everyday, everywhere.

Challenge: In the race to be ready for agentic commerce, PayPal realized its on-premises infrastructure couldn’t provide the performance and scale required to power autonomous AI agents. At the same time, fragmented development was draining productivity and lacked the unified governance needed to meet PayPal’s rigorous security and compliance standards.

Solution: To continue its legacy as a digital payments pioneer, PayPal reimagined its technology foundations and infrastructure, building its AI stack on Google Cloud. GKE was central to this approach, allowing PayPal to dynamically scale and cut release lifecycles from weeks to minutes. PayPal also developed a unified LLM stack on Agent Platform to enable developers to build, scale, govern, and manage agents at global scale without having to worry about managing infrastructure. PayPal also uses open standards like the Model Context Protocol and Google’s Agent Payments protocol to ensure every agent action is auditable, authorized, and secure.

Results: With Google Cloud, PayPal has fundamentally transformed its path to agentic commerce. The results are clear: 99.999% availability while serving 500+ models and handling billions of requests, model deployment in minutes instead of weeks, and 10,000 engineers building agentic experiences that 400 million users can trust — everyday, everywhere.

Key takeaways and highlights from our interview with Srinivasan Manoharan, Director of AI/ML Platform at PayPal:

→ “With Google Cloud infrastructure, we have achieved five nines of availability for our most critical AI use cases. We serve about 500-plus models in production, handling nearly one billion requests every single day. That’s the scale we are operating at, and the platform doesn’t blink.”

→ “We built a unified LLM stack. That meant we ran our open source models on top of our inferencing stack on GKE, and we ran our close-sourced models on top of Gemini Enterprise Agent Platform. Developers get choice – Claude, Gemini, Open Source – through one consistent governed access layer. That’s why we love Model Garden.”

→ “Google Cloud handles the infrastructure heavy lifting. For developers, the TCO is zero. Model deployment that used to take weeks now takes minutes. That’s 10,000 engineers across PayPal focused entirely on applying AI for their business problems and not managing infrastructure.”

→ “We can’t put agents into production if we do not have clear visibility into how they behave. MCP lets agents communicate with payment systems in a structured, auditable way. Every action gets logged, no black box. AP2, on the other end, defines us the rules, the maximum transaction amount, approved merchants and time window.”

→ “Today, we support hundreds of internal tools through our MCP layer, enabling agents to take real actions across our internal platforms. Agent Runtime simplifies all of that for us. PayPal evolves from a payment processor into a trust infrastructure. That’s what building on Google Cloud has made possible for PayPal.”

Google Cloud products used: Google Kubernetes Engine (GKE), Gemini Enterprise Agent Platform, BigQuery, Cloud GPUs

Learn more:

→ PayPal Showcases AI-Powered Innovation at Google Cloud Next ’26, Built on the Foundation of One of the Largest Data Migrations in History https://newsroom.paypal-corp.com/2026…

→ PayPal’s historically large data migration is the foundation for its gen AI innovation https://cloud.google.com/blog/product…

→ Google and PayPal Forge Multiyear Partnership to Revolutionize Commerce https://www.googlecloudpresscorner.co…

The video highlights several benefits for PayPal’s 10,000 engineers resulting from their transition to Google Cloud’s AI infrastructure:

• Increased Productivity: Engineers are empowered to focus entirely on applying AI to business problems rather than managing underlying infrastructure (2:11 – 2:19).
• Faster Deployment: The release lifecycle for models has been significantly reduced, dropping from weeks of regression testing to just minutes (1:14 – 1:22; 2:06 – 2:09).
• Unified Access: Developers benefit from a consistent, governed access layer that provides the choice between open-source models, Claude, and Gemini, ensuring they have the right tools to build and scale agents effectively (1:57 – 2:03).

Key Pillars of the Partnership

How Agentic Commerce Works in Practice

Strategic Impact