Hanoi, October 17, 2026 – FPT Corporation organized FPT TRANSFORM 2026, an international technology event dedicated to artificial intelligence (AI), cloud computing, and data. At the event, FPT introduced three new solutions: FPT AI EOS, FPT Kubernetes Engine with GPU, and FPT Token Factory, designed to help enterprises integrate AI into business workflows, manage computing infrastructure, and utilize AI models.
Held at InterContinental Landmark 72 under the theme “Ship AI Live,” FPT TRANSFORM 2026 focused on moving AI from experimentation to real-world deployment. The event featured more than 50 in-depth sessions, 10 hands-on technical workshops, and over 200 solutions showcased and demonstrated. According to the organizers, more than 1,700 tickets had been sold ahead of the event.
FPT introduces a range of new technology solutions at FPT TRANSFORM 2026.Deploying AI applications in real-world environments requires enterprises to address several challenges simultaneously, including integration with existing systems, control over data and outputs, support for growing numbers of users, and management of operating costs. FPT's three new solutions are designed to support different aspects of this process.
Mr. Le Hong Viet, CEO of FPT Smart Cloud, FPT Corporation.Mr. Le Hong Viet, CEO of FPT Smart Cloud, FPT Corporation, said the solutions were developed to give enterprises greater flexibility in choosing AI deployment approaches that match their needs and capabilities.
“By adding FPT AI EOS, FPT Kubernetes Engine with GPU, and FPT Token Factory to our ecosystem, we want to enable every organization to choose the level of AI autonomy that best suits their needs and capabilities, from applications and models to infrastructure. Our ultimate goal is to give enterprises greater control over technology and data, creating a foundation for innovation, deploying AI at scale in real-world operations, and generating business value,” Mr. Le Hong Viet shared.
FPT AI EOS Helps Integrate AI into Business Workflows
FPT AI EOS is developed as an AI agent orchestration platform, enabling enterprises to integrate AI more deeply into business processes and facilitate collaboration between AI and human employees.
The platform can connect with software systems used for customer support, contract processing, and sales operations. Digital workers can track tasks, interact with existing enterprise systems, and escalate decisions requiring human oversight to authorized personnel for approval.
Ms. Dao Bich Ngoc, Product Owner of FPT AI EOS at FPT Smart Cloud, FPT Corporation, demonstrates the newly launched FPT AI EOS platform.At the event, FPT demonstrated a digital worker named Tin, a fictional AI-powered employee operating in a live environment to support business operations.
Tin can prepare quotations, track customer responses, generate contracts, consolidate feedback from Sales, Legal, and Finance departments, and update the status of individual deals.
However, actions such as sending information to customers, accepting contractual terms involving risks, or signing contracts still require approval from responsible personnel.
FPT Kubernetes Engine Supports AI Infrastructure Operations
FPT Kubernetes Engine with GPU (graphics processing units used for AI computing) adds a managed Kubernetes layer for AI workloads to the FPT AI Factory ecosystem.
Kubernetes is a technology used to manage software deployment and operations. The solution provides an environment for training, fine-tuning, and deploying AI models, reducing the need for enterprises to independently configure and manage entire infrastructure clusters.
Mr. Nguyen Khuong Duy, Deputy Director of the Cloud Technology Division and Director of the NCP Product Center at FPT Smart Cloud, FPT Corporation, presents FPT Kubernetes Engine.FPT manages the infrastructure layer and Kubernetes control plane, including installation, upgrades, monitoring, automatic scaling, and GPU-related software components.
Customers retain control over their applications, models, and AI workloads.
The service supports deployment on GPU-enabled virtual machines or physical servers, with cluster creation and management available through a service portal, Terraform, or application programming interfaces (APIs).
According to FPT's internal testing results, a GPU-enabled Kubernetes cluster can be provisioned in approximately five minutes, workloads can be deployed into the operating environment in around 30 seconds, and manual AI infrastructure operations can be reduced by up to 90%.
These results were obtained under FPT's specific testing conditions and configurations and do not constitute performance guarantees for all systems.
FPT Token Factory Provides Unified Access to Multiple AI Models
FPT Token Factory provides a unified gateway to multiple AI models, including models operated directly by FPT and those offered by other providers.
The platform focuses on model access and inference, meaning the use of existing AI models to process requests and generate outputs, rather than training models from scratch.
Mr. Tran Viet Thanh, Deputy Director of the Cloud Infrastructure Development Center at FPT Smart Cloud, FPT Corporation, introduces FPT Token Factory.For AI models operated by FPT, data is processed on FPT infrastructure.
For models accessed through third-party providers, FPT Token Factory serves as an access gateway and usage management platform. Data processing in these cases depends on the policies and infrastructure of the respective providers.
By bringing multiple AI models and deployment options together through a single gateway, the platform enables enterprises to select models based on specific workload requirements while managing costs and resources more centrally.
Enterprises can use FPT Token Factory independently or integrate the platform with other solutions.
The three solutions can be deployed separately or combined according to business needs.
For example, an enterprise can develop and operate digital workers using FPT AI EOS, select suitable AI models through FPT Token Factory, and deploy workloads on FPT Kubernetes Engine with GPU when greater control over infrastructure, performance, and scalability is required.
Alongside the product launches, hands-on workshops at FPT TRANSFORM 2026 focused on practical applications, including AI-assisted programming, coordinating multiple AI agents for software development and testing, optimizing large language model processing speed, deploying cloud applications, and securing systems.
At the demonstration area, attendees had opportunities to experience the products firsthand and discuss enterprise deployment requirements directly with engineers and system architects.