HANOI, Vietnam–FPT AI Factory, in partnership with InFlow and Visa Intelligent Commerce, has launched an agent-native commerce platform, enabling access to frontier AI models that fuel AI agents in action.

AI agents can dynamically route between models based on cost, latency, or performance. This collaboration allows seamless AI agent workflows to autonomously research, procure, and pay for services while ensuring security and compliance, pioneering the B2AI economy on a global scale.

FPT AI Factory serves as the production inference layer, demonstrating the agent-native commerce platform live where AI agents access compute services, execute high-frequency API calls, and generate sustained, programmatic demand.

The Dawn of B2AI Commerce

B2AI commerce, where AI agents research, negotiate, and complete purchases autonomously, is rapidly moving from concept to production reality. The new horizon is B2AI: an emerging commercial paradigm in which AI agents are the new customer type executing end-to-end procurement workflows under human-defined intent and outcomes.

Just as cloud infrastructure unlocked the SaaS economy, the agent-native commerce platform is now unlocking the B2AI economy, opening an entirely new commerce market.

Closing the Loop: The Full B2AI Commerce Stack

This three-way partnership assembles every layer required for an end-to-end, agent-native transaction.

InFlow provides B2AI commerce infrastructure for both sides of the transaction. AI agents (buyers) onboard, pay, and transact as discrete customers. Businesses (sellers) onboard, transact with, and monetize agent-driven demand at scale.

Visa Intelligent Commerce contributes trusted payment credentials and global network access, ensuring every agent-initiated transaction is authenticated, policy-governed, and fully auditable.

FPT AI Factory serves as the production inference layer, demonstrating the agent-native commerce platform live where AI agents access compute services, execute high-frequency API calls, and generate sustained, programmatic demand.

Together, the three companies cover the complete transaction stack: identity, payment credentials, and AI compute, making B2AI commerce operational today.

What This Unlocks for AI Innovators

For AI agent developers, the partnership removes the friction that has historically kept commerce out of autonomous workflows. Procurement, onboarding, and payment now execute without human-in-the-loop intervention, and every transaction remains traceable and policy-governed in a single, uninterrupted workflow.

For businesses offering AI services, the platform opens up an entirely new customer segment. AI agents can autonomously onboard and transact on the platform, purchasing inference services, selecting models for corresponding workloads, and sustaining operations without interruption.

Powering Continuous Agent Workflows on FPT AI Factory

FPT AI Factory (NVIDIA Cloud Partner) offers robust GPU Cloud, inference-ready AI platforms, and access to more than 25 of the latest AI models via production-ready API, including Nemotron 3 Super, Alpamayo, Qwen 3, and Llama 4, covering the full range of AI workloads from training to deployment.

The platform utilises the latest NVIDIA HGX B300, H200, and H100, fulfilling the rigorous demands of next-generation AI and high-performance computing workloads.

Within FPT AI Factory, agents dynamically route between models based on cost, latency, or performance requirements, with InFlow handling continuous usage and payment settlement across every inference request. The result is an uninterrupted agentic workflow: agents identify the capability they need, onboard programmatically, transact via Visa Intelligent Commerce’s global payment network, and execute inference at scale, all without a single manual approval step.

About FPT AI Factory

FPT AI Factory delivers an all-in-one AI Developer Cloud that combines NVIDIA-accelerated GPU Cloud services, inference-ready AI platforms, and ready-to-use AI applications. Guided by the vision “Build Your Own AI,” FPT AI Factory empowers enterprises, startups, and the tech community with the compute power, model variety, and deployment flexibility to support any AI workload while ensuring optimal price-performance and sovereign AI.