The AI Factory built for insurance.
A $200M+ NVIDIA-powered supercompute hub in Vietnam. GPU capacity, sovereign data, and 35,000+ engineers — engineered so carriers, brokers, and reinsurers can train, tune, and deploy insurance-grade AI without ever leaving regulated boundaries.
Shaping the future of AI and cloud on a global scale.
FPT and NVIDIA are co-investing in one of Southeast Asia's most ambitious AI infrastructure programs — a purpose-built AI Factory in Vietnam that combines NVIDIA DGX H100 supercomputing, NVIDIA AI Enterprise software, and FPT's insurance domain expertise.
For insurance leaders, this means a rare thing: a hyperscale AI environment that is sovereign, regulated-ready, and staffed by insurance engineers — not a generic cloud tenancy.
Read the FPT × NVIDIA announcement →What makes it an AI Factory, not a data center.
Compute alone is a commodity. The AI Factory pairs frontier-grade GPUs with sovereign controls, the largest AI-native engineering workforce in the region, and a productized model lifecycle tuned for insurance workloads.
NVIDIA DGX H100 Supercluster
Purpose-built GPU supercomputing anchored by NVIDIA H100 systems and the NVIDIA AI Enterprise stack — the same class of infrastructure powering the world's largest frontier labs.
Sovereign & Regulated Data
In-country data residency for regulated insurers. Models train and infer inside jurisdictional boundaries — critical for NAIC Model Bulletin, EIOPA, MAS, and APRA compliance.
35,000+ Engineers & AI Specialists
A dedicated AI hub in Vietnam plus a globally distributed engineering force — the largest AI-native delivery organization in Southeast Asia, fluent in insurance domain.
Model Factory & MLOps
End-to-end lifecycle: from foundation model fine-tuning and RAG pipelines to eval harnesses, guardrails, and continuous monitoring — production-ready in weeks.
Six ways insurers plug directly into the AI Factory.
Every capability is delivered as a managed service — combining GPU compute, pre-trained insurance models, and an FPT engineering pod. You get a production outcome, not a science project.
Domain-Tuned Underwriting LLMs
Underwriters spend 60–70% of their day rekeying data from broker submissions, loss runs, and unstructured PDFs — while risk appetite decisions still rely on tribal knowledge locked in senior heads.
Fine-tune foundation models on your submission history, appetite guidelines, and rating manuals inside the AI Factory. Private RAG pipelines surface precedent risks and auto-populate the workbench — with an FPT underwriting pod handling every exception.
From partnership to planetary-scale sovereign AI.
Strategic NVIDIA partnership announced
FPT commits to building one of Southeast Asia's largest AI factories in collaboration with NVIDIA.
AI Factory goes live in Vietnam
$200M+ investment brings DGX H100 supercomputing, NVIDIA AI Enterprise, and Omniverse online.
Insurance model catalog launches
Domain-tuned underwriting, claims, and actuarial models productized for global carriers, brokers, and reinsurers.
Sovereign AI expansion
Regional AI factory nodes for EU, Japan, and North America — regulated inference at the edge of every jurisdiction.
Every line of business. Every model. One factory.
Book a tour of the AI Factory.
Bring your hardest underwriting, claims, or actuarial workload. We'll show you the compute, the models, and the engineering pod that will run it in production.