Christine Ahn, NVIDIA Chief Commercial Officer & Lead Client Service Partner is interviewing with ChosunBiz on the 29th of May. /Courtesy of Deloitte Anjin

It is no exaggeration to say that the future global hegemony competition rests on artificial intelligence (AI). The world is entering a new battleground over AI hegemony. Countries without AI sovereignty will have limited ability to shape their own AI future.

AI hegemony directly relates to the issues of security and economic sovereignty. Depending on who holds the AI infrastructure, the power map can change. The new government’s appointment of Ha Jung-woo, former head of the Naver Cloud AI Innovation Center, as chief AI officer signifies more than just talent acquisition. Positioning a technology expert who understands the flows of cloud, chips, reasoning, and generation at the forefront of national strategy is proof that our government recognizes the era where ‘AI cannot be run on someone else’s server.’

On the 29th of last month, I met with Christine Ahn, NVIDIA Chief Commercial Officer & Lead Client Service Partner, at the Deloitte Anjin headquarters in Yeouido, Seoul, to inquire about the current state of the global AI industry. Christine Ahn noted that governments around the world are rushing to establish their own AI factories, that is, private cloud infrastructures. She also introduced the latest trends in the AI industry, discussed the ‘Zora AI’ service created by Deloitte, and provided details about the partnership with NVIDIA.

The following is a Q&A.

─What is the reason for your visit?

I took part in a keynote speech at ‘Deloitte ConnecT in Korea,’ hosted jointly by the Deloitte Korea group and seven major big tech companies from Deloitte Global.

More than 200 C-level executives from major domestic corporations attended, where we shared AI transformation strategies and case studies alongside partners from the big tech alliance of Deloitte Global including AWS, Dataiku, NVIDIA, Salesforce, SAP, ServiceNow, and Workday.

Deloitte prides itself on being a leading company in global technology services. To proactively respond to global AI trends, we developed the AI platform ‘Zora AI’ in collaboration with NVIDIA. Deloitte actively supports the introduction of generative AI for corporations worldwide, from high-performance computing (GPU) infrastructure to services.

─What are the latest trends in the AI industry?

The first point is that generative AI is evolving into Agentic AI, or ‘acting AI.’ As I explained yesterday (May 28), in the past, users would input simple questions like ‘What is the weather in Seoul today?’ and receive a brief response. Now, the questions themselves are becoming increasingly complex, resulting in responses derived through complex reasoning.

Such Agentic AI requires a much higher level of computation than traditional generative AI. Metrics like cloud usage, inference capacity, and token consumption can increase by more than 20 times. This leads to a completely different dimension of computing demand compared to existing data processing.

The second trend is the growing importance of AI factories and sovereign AI. About 90% of existing data centers operate on CPU-based systems. However, to properly run Agentic AI, CPUs are insufficient. A new AI infrastructure centered on GPUs is required to develop, train, and operate AI systems.

The issue is that existing data centers cannot handle this GPU-centric AI. GPU servers can weigh over a ton, and traditional air conditioning systems cannot cool them effectively. Instead, liquid cooling systems are necessary.

So we have created a model that provides AI factories as a service. This includes everything from chips to storage and network configurations. Deloitte is helping each corporation design customized structures for building these AI factories.

The third is physical AI and robotics. This is not a story for the distant future. Robotics is rapidly growing, and companies like JP Morgan and Citibank are seeing robotics as a next-generation trillion-dollar industry. In the future, robots will be present everywhere, whether in homes or factories. At Deloitte, we are equipped with solutions to match these three trends.

─What conversations did you have with senior executives from Korean corporations?

I understand that many high-ranking individuals with decision-making authority attended. Interestingly, their understanding of concepts such as Agentic AI, AI factories, and physical AI varied significantly.

─Globally, which corporations or industries have a high level of AI adoption?

Rather than pinpointing specific corporations, I would prefer to discuss industry sectors. In my view, the automotive industry, particularly in the robotics field, stands out for its technological maturity. For instance, Tesla, which has a very close partnership with NVIDIA, is leading the way in autonomous driving.

BMW cannot be overlooked. BMW is providing innovation through AI at manufacturing plants worldwide. In particular, it is simulating the car production process using digital twins. These are very advanced examples of how physical AI is being utilized not only in autonomous driving but also in the automotive manufacturing process.

The next advanced sector is the Financial Services Industry (FSI). The Financial Services Industry generates vast amounts of data. For instance, every time we make a purchase, extensive data such as customer information, purchase history, and payment information is generated. This data is also used for purposes like fraud detection and cyberattack responses.

To perform these functions effectively, high-performance computing capabilities are needed, and currently, the fastest chips in the world are NVIDIA’s chips. In fact, almost every bank has a server equipped with NVIDIA’s GPU technology somewhere in their data centers.

The energy industry is somewhat slow in its transition to AI. However, Germany is advancing quite well in applying AI to various infrastructures, including power grids. The telecommunications sector is similar. While it has been slow thus far, it is now rapidly transitioning and entering the stage where it is adapting how to implement AI in their businesses beyond simply introducing it.

The maturity of AI technology in the consumer goods and retail industry can be considered at a mid-level. These corporations have massive datasets related to what customers bought, which stores they visited, and what credit cards they used. Processing this data to provide personalized products or services also requires high-performance computing capabilities.

─How did the partnership between Deloitte and NVIDIA begin?

It started five years ago. The starting point was, above all, the moment our mindset at Deloitte changed. Deloitte did not view this partnership merely as ‘selling services to NVIDIA.’

At that time, we believed that the AI era would inevitably come and that we needed to respond proactively to such changes, and we considered NVIDIA as the optimal partner to accompany us on that journey, as they already possessed unmatched capabilities in high-performance computing, particularly GPU technology.

Thus, Deloitte actually purchased six GPU nodes and built an internal environment based on that. This environment was used to test new technologies, experiment with innovation, and train members. Deloitte was the first global system integrator (GSI) to attempt this.

Since starting this, Deloitte has been developing increasingly more solutions based on NVIDIA’s technology. We have created various solutions based on the software development kits (SDKs) that NVIDIA provides. Examples include interactive avatars, logistics optimization solutions, and drug discovery in the new drug development stage. These solutions are those developed by Deloitte based on NVIDIA’s technology.

Thanks to this, Deloitte has gained significant interest and support from NVIDIA’s CEO Jensen Huang and executives. Substantial support was also received from NVIDIA’s institutional sector leaders and sales leaders, resulting in a very successful partnership between the two companies.

Deloitte has since been recognized as ‘Consulting partner of the year’ by NVIDIA for four consecutive years and was selected as Consulting Partner of the Year at the NVIDIA Partner Awards EMEA (Europe, the Middle East, and Africa) 2025.

─Please introduce Zora AI.

NVIDIA holds a global technology conference every March, and at this year’s conference, our Deloitte’s Agentic AI solution ‘Zora’ was unveiled. The name Zora is derived from the ‘goddess of dawn’ in Slavic mythology.

Zora AI is a software created through collaboration between Deloitte and NVIDIA, aimed at providing advanced intelligent agents capable of performing tasks and reasoning.

Zora first started being used internally in Deloitte’s finance and accounting department, with Hewlett Packard Enterprise (HPE) being its first major client. The CEO of HPE mentioned the introduction of Zora during last year’s fourth-quarter earnings conference call.

─Recently, the importance of sovereign AI has been increasing. Why is that?

First, it is necessary to explain the concept of tokens. When we input something into ChatGPT or Google, that input is composed of tokens, which are the smallest units (1,000 Korean characters typically convert to about 500 to 700 tokens). As soon as we hit the Enter key, the tokens are bundled and sent through the cloud network, where inference takes place, and the answer is returned.

The problem is that expenses arise at each step of this process. The flow of money in the AI world originates from this ‘tokenization.’

Agentic AI, which requires high-level reasoning, uses 20 times more tokens. Thus, the costs of cloud usage are continually trending upward.

As a result, corporations have begun to contemplate sustainable alternatives, and the answer that has emerged is now the ‘AI Factory.’ This means constructing private clouds within corporations to handle data and computations internally. They aim to run AI without relying on external public clouds by utilizing their own GPUs, storage, and networks.

This trend is now expanding at the government level. Many countries are just beginning to recognize the need for their own private clouds and independent AI infrastructures. Currently, most corporations with cloud infrastructures are of U.S. nationality.

Countries like France, Japan, Saudi Arabia, and Qatar are announcing investments ranging from billions to potentially hundreds of billions of dollars. This is no longer just a matter of expense; it is an issue of securing AI sovereignty from a national security perspective and to ensure economic competitiveness. In other words, it is about not entrusting our data, computations, and future to companies from other countries. This is what sovereign AI entails.

─What differentiates Deloitte’s AI services?

Deloitte possesses the capability to organically integrate key technology elements for AI adoption through alliances with global big tech companies, such as NVIDIA, beyond merely applying technology. Especially through close collaboration with NVIDIA, we can rapidly implement business-specific AI solutions on the latest AI platforms.

The goals of corporations in establishing AI strategies are transitioning from ‘what technology to use’ to ‘maximizing effectiveness relative to cost by placing the optimal technology tailored to each task appropriately.’ Deloitte will position itself as the optimal partner to support corporations in achieving their AI strategy goals by comprehensively considering technology, infrastructure, and business processes.