Hong Daeui, CEO of Mondrian AI: AI Utilization Matters More Than Securing Models

Expanding NeoCloud and Regional AX, Targeting KRW 10 Billion in Revenue This Year
Enterprise AI Management with Multi-Model Agents, Reducing Dependence on Specific Models

 

Mondrian AI is accelerating its expansion into the enterprise and regional AI transformation (AX) markets, led by its AI agent and NeoCloud businesses. Rather than competing to directly secure GPUs and AI models, the company is focusing on connecting existing infrastructure and models with enterprises and developers so they can be applied to real-world operations.
 

In an interview held on September 12 during “Agent Field Trip: Incheon” in the Dongincheon area, Mondrian AI CEO Daeui Hong discussed changes in the AI agent market and outlined the company’s strategies for NeoCloud and regional AX. As the supply of AI infrastructure and models continues to grow, he emphasized that the ability to translate these resources into practical applications is becoming increasingly important. Hosted by GDG Incheon, the event invited participants to explore the Dongincheon area, identify local challenges, and develop services using AI agents. Mondrian AI supported the event with Runyour Agent, its platform that provides access to multiple generative AI models.
 

As a founder and organizer of GDG Incheon and GDG Songdo, Hong viewed the event as an opportunity to examine how significantly AI agents could lower the barriers to software development. “The era when software development belonged exclusively to developers is coming to an end,” he said. “We wanted to see whether people without coding knowledge could now turn the problems and ideas they discover into working services within just a few hours.”

 

Managing Multiple Models Becomes a Key Enterprise Challenge

As AI adoption expands across organizations, managing multiple models within a single operating environment has become an increasingly important challenge. Enterprises must manage not only model selection, but also token usage, costs, access permissions, and outputs in an integrated manner. Hong also identified dependence on a specific AI model or provider as an issue that enterprises must address. With new models emerging rapidly, organizations need the flexibility to select the right model for each task rather than becoming locked into a single provider. “Organizations should be able to select the models they need at any time without being anchored to a particular model provider,” he said. “Users should be able to focus on their work while choosing models according to their needs.”
 

Runyour Agent is a multi-model agent platform that enables organizations to use multiple large language models (LLM) within a single environment. It is designed to provide integrated management of token usage and costs by organization, as well as project-level access permissions. “Using AI as an individual has become easy, but the situation changes when multiple people use multiple models simultaneously,” Hong explained. “The key question will be how organizations manage costs, outputs, performance, and efficiency across the entire enterprise.”

“In the future, organizations will also need roles that manage AI models and usage in much the same way that they manage human resources,” he added. “We are building an environment where multiple users and models can be managed at the organizational level.”

 

Focusing on AI Distribution Rather Than Asset-Heavy Competition

Alongside its enterprise agent business, Mondrian AI has positioned NeoCloud as another key growth driver. Hong described the traditional cloud and AI data center markets as asset-heavy businesses that require substantial capital investment to secure data centers and GPUs. Although investment in AI data centers and the supply of GPUs and AI models are expanding rapidly, Mondrian AI is placing greater emphasis on connecting existing infrastructure and models with real-world demand rather than competing to acquire more assets.
 

“Competition is intensifying rapidly across both GPU resources and AI models,” Hong said. “Instead of joining that competition directly, we are focusing on how existing resources and models can be delivered effectively to users.” NeoCloud is being developed in line with this approach. It combines GPU resources, AI development and operations platforms, and agent services, enabling enterprises to apply the resources and models they need directly to their operations. Hong described this as an issue of AI “distribution.” Even if the supply of infrastructure and models increases, widespread adoption will remain difficult if enterprises and developers cannot access and use them easily.
 

He noted that Korean AI models face the same challenge. “Many domestic AI models are being developed, but there are still not enough channels capable of delivering them effectively to general users,” he said. “Beyond building strong models, we need to open distribution channels that allow developers and the public to use them in practice.” As the government moves beyond AI model development toward services that deliver tangible value to the public, platforms that connect models with users will play an increasingly important role.

 

Connecting Regional Data Centers with Industrial AX

Mondrian AI is also extending its NeoCloud strategy into regional AX. Centered on Incheon, where the company is headquartered, the company is pursuing projects that connect regional data centers with industries such as manufacturing, biotechnology, and logistics. Hong expects AX demand, which has largely been driven by major enterprises, to expand to mid-sized companies and SMEs as the cost of adopting generative AI and AI agents continues to decline.


“In the past, AX investments often imposed a greater financial burden than the benefits they delivered for mid-sized companies and SMEs,” he explained. “We are now moving toward an environment where companies can achieve greater results with significantly less investment.” “Regional companies often find it difficult to secure both AI resources and cloud expertise on their own,” he added. “Mondrian AI can support these capabilities and help accelerate AX among regional enterprises.”


Hong also emphasized that regional data centers must move beyond simply installing servers and GPUs. For AI to be applied in real industrial environments, enterprises need software capabilities that can transform their data into an AI-ready format and connect it with models and services. “Beyond the physical construction of data centers and infrastructure, converting conventional data into a form that AI can use requires specialized software and technical expertise,” he said. “Providing these capabilities to regional industries is another role that Mondrian AI can play.” Incheon has a diverse industrial base encompassing biotechnology, manufacturing, logistics, aviation, and port operations. Mondrian AI plans to expand regional AX initiatives in Incheon and beyond by connecting regional data centers with industry-specific demand.

 

Expanding Across AI Cloud, Agentic AI, and Industrial AX

Mondrian AI plans to expand its business across three key areas: AI cloud, agentic AI, and enterprise and public-sector AX. Its AI cloud business provides enterprises and developers with easier access to regional data centers and GPU servers. Its agentic AI business focuses on improving operational efficiency across industries such as manufacturing. The company’s AX initiatives apply its proprietary cloud and platform technologies to meet industry-specific needs.
 

“We are expanding our AI cloud so that general users can easily access data centers and GPU servers being established across different regions,” Hong said. “Our agentic AI business is focused on improving efficiency for manufacturing companies and other industries, while our AX business is expanding across sectors based on our cloud and platform technologies.” Mondrian AI surpassed KRW 5 billion in revenue in 2025 and achieved operating profitability. As of the second quarter of this year, the company had secured approximately KRW 5 billion in contracts and is targeting more than KRW 10 billion in annual revenue. While its business had primarily focused on regional markets through last year, Mondrian AI is now expanding its customer base into Seoul and the broader capital region. The company also views overseas expansion as a mid- to long-term growth opportunity, although the target markets and timeline have not yet been finalized.
 

Mondrian AI plans to maintain a balanced business structure by operating both capital-intensive, asset-heavy businesses—including GPUs and data centers—and relatively asset-light platform and software businesses. Its strategy is to expand its infrastructure operations while securing recurring revenue from agents and platforms, strengthening the overall stability of the business. “For an AI company to grow sustainably, it needs to pursue relatively asset-light businesses alongside capital-intensive operations rather than relying exclusively on heavy infrastructure investment,” Hong said. “We will continue to expand our business portfolio based on our cloud and platform capabilities.”
 

In addition to Runyour Agent, Mondrian AI is expanding its AI infrastructure and platform business through services including Runyour AI and Runyour Cloud. The company also offers specialized solutions designed to support AX across industries such as manufacturing. As the supply of GPUs and AI models continues to increase, the focus of market competition is likely to shift from securing resources to enabling their effective use and operation. Mondrian AI’s future growth will depend on its ability to connect its NeoCloud and AI agent businesses with enterprise and regional AX demand—and achieve its target of surpassing KRW 10 billion in revenue this year.




Source: TechWorld
https://www.epnc.co.kr



👉 Read the full article on TechWorld