Efficient Management of Token Consumption and Operating Costs as Key Competitive Advantage

Cloud Dependency Limitations, On-Premises Cost and Control Effective

"AI success is determined by data and infrastructure, not by the model"

Dell Technologies held the 'Dell Technologies Forum 2026' at COEX in Samseong-dong, Seoul on the 25th, presenting solutions for converting AI vision into tangible business results, including a data-centric strategy and AI-optimized infrastructure, on-premises-based hybrid AI environments, and next-generation work environments adapted to the age of AI agents.

Until last year, companies' interest remained focused on 'how to adopt AI,' but this year's forum had a clear agenda: "How to convert AI vision into business results."

Kim Kyung-jin, chairman of Dell Technologies Korea, President Yoo Sang-mo, and Matt Dunfee, senior vice president of Dell headquarters, emphasized through keynote speeches that AI has become a critical infrastructure that directly affects actual business performance rather than a future technology.

At the same time, they diagnosed that for companies to achieve tangible returns from AI investments, a realistic approach to data, infrastructure, security, and cost structure is necessary.

"More Important Than AI Adoption Is Performance," Dell's AI Execution Strategy

Chairman Kim Kyung-jin, at the opening of the event, said that when generative AI first emerged, AI was discussed as a future possibility, but now it is an era where it is difficult to imagine daily life and work without AI.

He explained that going forward, the focus of interest is shifting from AI adoption itself to how to reduce investment costs and power consumption while effectively utilizing AI for business and industrial development.

This shift is confirmed in companies' concerns.

President Yoo Sang-mo diagnosed that "AI is at the core of why semiconductors and power have become national issues."

He emphasized that AI is no longer a tool for assisting with work but has become a new foundation for corporate operations and daily life, and that power and cooling issues have become strategic matters at the corporate management level, not just concerns of IT departments.

According to Dell's 'Modern Enterprise Readiness Survey 2026,' 87% of corporate executives are experiencing difficulty keeping pace with rapid change, and concerns about AI data security and the establishment of execution roadmaps are also growing.

In particular, Senior Vice President Matt Dunfee identified differentiated data and operating model innovation as the key factors that determine AI adoption success.

Many companies are creating AI pilot projects and productivity improvement cases but are failing to link them to actual financial results.

He emphasized that simply adding AI to existing work processes will not yield expected results and that work methods themselves must be redesigned.

Cloud Alone Is Not the Answer, On-Premises Shift in the Age of AI Agents

The most notable keyword at this year's forum was undoubtedly AI agents.

As AI evolves beyond merely answering questions to performing tasks autonomously and assisting in decision-making, new changes are being required in enterprise infrastructure.

Senior Vice President Matt Dunfee predicted that an era will arrive where every employee manages dozens to hundreds of AI agents.

The issue is cost.

While GPU procurement was the competitive advantage in the generative AI era, in the agent era, how efficiently token consumption and operating costs can be managed becomes the core competitive advantage.

Through actual examples, he explained that even for identical AI tasks, significant cost differences can occur depending on which model is chosen and where it is executed.

Dell's message on this day was clear.

As the scale of AI utilization expands, relying solely on public cloud approaches can reveal limitations in terms of cost-effectiveness.

In an agent environment where token usage explodes, operating AI within the data center through on-premises approaches can be a more effective choice in terms of cost and data control.

Senior Vice President Matt Dunfee emphasized that "it is very important to determine the economics of which models to place in the cloud and which models to place within the enterprise."

Dell is taking the lead with 'Dell AI Factory,' which integrates servers, storage, data platforms, and AI software to respond to these environmental changes.

According to the company, over 5,000 customers worldwide currently use AI Factory to operate AI workloads.

In the age of AI agents, the personal work environment is becoming important in addition to data centers.

In environments where multiple AI agents operate simultaneously, PCs are expanding their role from being simple business terminals to becoming AI execution platforms.

According to Dell's survey, 90% of domestic respondents answered that the latest devices improve collaboration quality, and 87% responded that outdated PCs become obstacles to AI application adoption.

This suggests that in the future, AI PCs and high-performance work devices will become core components of enterprise AI strategy.

AI competitiveness starts in the data center, but actual productivity innovation is completed on the PC on employees' desks.

Ultimately, the message Dell Technologies Forum threw this year is clear. Enterprise competitiveness in the AI era is not simply about adopting the latest models. Success or failure depends on how efficiently data is utilized, how effectively costs are controlled, and how AI is established in business processes while ensuring security and governance.

Dell presented data center modernization and on-premises-focused hybrid AI, along with work environments optimized for the age of AI agents, as the answer, emphasizing that it is now time to discuss 'AI's profitability' rather than 'AI's future.'