The AI race is accelerating, yet the gap between experimentation and real-world deployment remains a major challenge for businesses. According to the CIO Vietnam Survey 2026, nearly 75% of enterprises have begun experimenting with or deploying AI, but only 6.3% report having demonstrated clear financial value across their AI portfolios.
At CIO Summit 2026, Le Hong Minh - Founder and Chairman of VNG Corporation, and Vu Thanh Tung - CEO of GreenNode, shared their perspectives on the next phase of AI adoption among Vietnamese enterprises. Drawing on their organizations’ experiences over the past several years, the speakers explored new approaches to enterprise AI from empowering leadership and a culture of innovation to building the infrastructure needed to deploy AI securely, efficiently, and with greater control over data.
The “Ask CIO, CEO Anything About AI” discussion at CIO Summit 2026, held in Ho Chi Minh City on October 1. Photo: GreenNode
AI Adoption Must Reflect Each Business’s Needs
During the discussion, Le Hong Minh emphasized that the approach to AI adoption should reflect the specific needs and operating models of each business. Pressure to measure return on investment (ROI) and demonstrate tangible results is prompting many CIOs to carefully evaluate how they implement AI.
In sectors such as banking, a top-down approach is often more suitable, starting with specific use cases that have demonstrated clear value and ROI. By contrast, with a bottom-up approach where employees are primarily given access to AI platforms and tools, the challenge is determining how to measure the impact and ROI of AI investments.
However, Minh noted that AI investment remains manageable when viewed in the context of an enterprise’s overall technology budget. In the past, banks had to commit substantial resources to core banking systems, data centers, and servers. Today, a significant share of AI-related costs comes from token consumption. Yet these costs remain relatively small compared with overall technology spending, particularly when weighed against the productivity gains AI can deliver.
“With AI, the biggest cost today is tokens. Compared with the productivity gains across teams, I believe these costs remain manageable and represent a relatively small share of overall technology investment.”
Le Hong Minh (right) and Dao Ngoc Dung (left) in discussion during the panel. Photo: GreenNode
AI’s Value Starts with People
Rather than taking a purely top-down approach, VNG has adopted a bottom-up model that encourages employees to explore AI independently and apply it to their work. According to Minh, AI’s greatest value lies not simply in integrating it into existing systems, but in unlocking the potential of individuals across the organization.
For this reason, VNG prioritizes building AI capabilities across the organization rather than focusing solely on how much revenue or profit AI can generate. As employees become more confident in applying AI to their work, individual capabilities can gradually develop into an enterprise-wide strength.
On the risk front, Minh expects security to become a major concern over the next two to three years, as offensive capabilities evolve faster than defensive measures. At the same time, as AI becomes increasingly accessible, businesses will need the flexibility to select models based on cost, data type, and deployment requirements, including on-premises options.
This also reflects VNG’s long-term direction: building AI capabilities and infrastructure first to meet internal needs, then expanding these capabilities to serve customers in Vietnam.
Overcoming Infrastructure Challenges to Scale Enterprise AI
According to Vũ Thanh Tùng, CEO of GreenNode, building the right infrastructure is essential to bringing AI into enterprise operations and enabling employees across the organization to use it effectively.
Vietnamese businesses face several interconnected infrastructure challenges:
Rather than requiring each business to build and manage the entire infrastructure stack independently, GreenNode is developing an ecosystem comprising four layers:
- Sovereign Infrastructure: GreenNode provides high-performance GPU infrastructure tailored to different workloads, deployed across multiple availability zones in Hanoi and Ho Chi Minh City. Located close to customers’ data centers, this infrastructure enables data to be stored and processed within Vietnam while connecting directly to enterprise private networks.
- AI Platform: The platform enables business teams to fine-tune open-source models using proprietary data and move them into production quickly, with the flexibility to scale as requirements evolve.
- Vietnam-Based Token Infrastructure: Scheduled for launch by the end of the year, this offering is designed to process tokens on infrastructure located in Vietnam, connect to enterprise private networks, and provide transparency into usage. GreenNode is focused on optimizing the inference layer to make token consumption more cost-effective for businesses and end users.
- AI Agent Orchestration and Governance: Designed for enterprises deploying AI agents, this governance layer enables organizations to monitor which agents are active, how many tokens they consume, and which data they access.
Vu Thanh Tung delivers his presentation at CIO Summit 2026 on the afternoon of October 1. Photo: GreenNode
From Business Use Cases to AI in Production
Infrastructure creates value only when it is connected to real business needs. GreenNode therefore enables business teams, including non-technical employees, to participate directly in building AI solutions.
GreenNode’s Hackathon as a Service model has been implemented at VNG through GreenNode Claw-a-thon 2026 and, more recently, at Vietnam Maritime Commercial Joint Stock Bank (MSB), attracting more than 300 participants.
At MSB, teams developed multiple AI agents to support sales operations, from checking the credit profiles of existing customers to recommending cross-selling opportunities. These agents run on GreenNode’s infrastructure in Vietnam and can be deployed in live operations once approved by the relevant business leaders.
The experiences shared by the two speakers at CIO Summit 2026 underscore a key principle: turning AI into a genuine operational capability requires businesses to develop their people and build reliable infrastructure in parallel. Leadership must create the conditions for innovation to flourish, while infrastructure provides the foundation for scaling those innovations securely, efficiently, and with greater control over data.
GreenNode supported MSB AI Hackathon as a resource sponsor and technical advisor on September 24, 2026. Photo: GreenNode
