The US is nurturing AI as an independent industry and has entered a large-scale recovery competition,

while Korea is in the phase of increasing revenue efficiency by adding AI to existing platforms and enterprise services.

Ultimately, the deciding factor in this cycle lies not in the model performance itself, but in the speed of monetization and the capital recovery structure.

The question of whether artificial intelligence (AI) is making money is now a step behind.

What the market is asking now is how efficiently AI is making money. The key criteria have become how much electricity, semiconductors, servers, data centers, and labor costs must be invested first to generate the same 1 won in revenue.

Although the US and Korea are experiencing the same AI frenzy, their monetization methods are clearly different.

The US is fostering AI as an independent industry, spearheaded by large-scale models and cloud infrastructure.

Conversely, Korea is generating revenue first by attaching AI to search, advertising, shopping, telecommunications, cloud, and collaboration tools to increase the turnover and unit prices of existing businesses.

US: Simultaneous Growth and Recovery Competition Underway

The strength of US big tech lies in its speed of transforming AI from a new cost into a new revenue stream.

Meta, a US platform company, projects a 12% increase in ad impressions and a 9% rise in ad prices on an annual basis for 2025. Capital expenditure guidance for 2026 is $115 billion to $135 billion.

Meta's AI strengthens the existing system by improving ad recommendations and impression efficiency, rather than charging significant separate usage fees. While monetization is already underway, larger investments are following to achieve even greater monetization.

Microsoft is closer to a productivity-based subscription model. In the second quarter of fiscal year 2026, revenue was $81.3 billion and operating profit was $34.3 billion. During the same quarter, Microsoft Cloud revenue surpassed $50 billion for the first time.

The company announced that Azure (Microsoft's core cloud service) revenue for fiscal year 2025 exceeded $75 billion annually.

This model involves attaching AI to Office and cloud services, layering prices on top of the existing paid customer base. It can be considered a relatively stable paid model among US companies.

Alphabet, Google's parent company, and Amazon, a US e-commerce and cloud company, are both experiencing significant revenue expansion and recovery cost burdens simultaneously.

Alphabet's Google Cloud revenue in the fourth quarter of 2025 was $17.664 billion, a 48% increase year-over-year, and its annual revenue for 2025 exceeded $402.8 billion.

Amazon announced that as of the fourth quarter of 2025, Amazon Web Services (AWS), Amazon's cloud business, had an annualized revenue of $142 billion, and as of the first quarter of 2026, its annualized AI revenue surpassed $15 billion.

However, Amazon added 3.9 gigawatts (GW) of power capacity in 2025 alone and plans to double its total power capacity by the end of 2027. While it is earning money quickly, the burden of upfront infrastructure investment is also increasing proportionally.

In summary, the US AI industry is now moving beyond "growth competition" into "recovery competition."

Structures that attach AI to existing cash-generating engines, like those of Meta and Microsoft, are relatively capital-efficient. On the other hand, structures that simultaneously expand search, cloud, and AI infrastructure, like those of Alphabet and Amazon, are large in scale but require more time for recovery.

The question for the US is "How big an industry will AI become?" and the subsequent question is "When will that industry truly start to yield profits?"

Korea: Two Paths – Efficiency Amplification and Infrastructure Transition

The picture in Korea is different.

There are not many companies in Korea that show AI's standalone revenue as a separate item as prominently as in the US. Instead, AI is currently operating by increasing the productivity, conversion rates, and per-customer revenue of existing businesses.

Naver, a leading domestic platform company, is the most symbolic example.

Naver recorded its highest-ever annual revenue of 12.03 trillion won and operating profit of 2.2081 trillion won in 2025. The company cited the expansion of AI-based personalized services as a growth driver.

This means that Naver's AI acts as a structure that enhances existing revenue-generating power by increasing the conversion efficiency of search, advertising, and shopping, rather than being an independent product.

Samsung SDS, an enterprise IT service company, is a sector in Korea showing relatively quick monetization.

In the fourth quarter of 2025, revenue was 3.5368 trillion won and operating profit was 226.1 billion won. Annual revenue was 13.9299 trillion won, and operating profit was 957.1 billion won.

The company is expanding its adoption in the public sector, finance, and enterprise fields, spearheaded by FabriX (an enterprise generative AI platform) and Brity Copilot (an AI collaboration tool).

Given that actual paid adoption is preceding buzz, it can be seen as a representative case of Korean enterprise-to-enterprise (B2B) AI.

SK Telecom and KT are closer to the infrastructure and transition type.

SK Telecom reported that its AI Data Center (AIDC, AI-dedicated data center) revenue for 2025 was 519.9 billion won, a 34.9% increase year-over-year. This was attributed to increased utilization rates of the Gasan and Yangju data centers and the acquisition of the Pangyo data center.

KT reported that kt cloud's revenue in 2025 was 997.5 billion won, a 27.4% increase year-over-year, with the growth driven by increased demand for AI cloud and greater utilization of global customer data centers.

While both companies have a clear direction, their capital recovery periods may be longer than Naver or Samsung SDS, as they involve upfront investments in data centers, power, and GPUs (semiconductors for large-scale AI computations).

Comparison of Korean and US Characteristics

The difference between Korea and the US can be summarized in one sentence.

The US is fostering AI as an independent industry, competing on economies of scale and speed of recovery, while Korea is first utilizing AI as an efficiency amplifier for existing industries.

Therefore, in the US, it is important "how large a revenue AI itself generates," as seen with Meta, Microsoft, Alphabet, and Amazon. In Korea, it is more important "how much AI boosts the growth rate and operating profit margin of existing revenues," as seen with Naver, Samsung SDS, SK Telecom, and KT.

Industrialization comes first in the US, and application comes first in Korea.

Market Analysis

The conclusion from a stock market perspective is also clear.

Currently, the most efficient AI monetization still appears first in structures that add AI to existing cash-generating engines. Meta and Naver are close to this model.

The fastest-growing revenue areas are enterprise AI and cloud infrastructure. Microsoft, Alphabet, Amazon, Samsung SDS, and SK Telecom are within this trend.

On the other hand, the area that will be verified last is infrastructure-based AI, which requires large upfront investments. This is because even if revenue increases, the recovery period may be long.

Ultimately, the outcome of this AI cycle will be determined not by "who has built the smartest model," but by "who can increase the revenue per unit and turnover rate of existing businesses with the least additional cost."

※ The ranges above are reference calculation values to aid in understanding the article, not investment recommendations. Actual stock prices may vary depending on various factors such as performance, interest rates, regulations, competition, supply and demand, and capital expenditure scale.