Second Quarter Revenue: KRW 133 Trillion, Up 106% Year-on-Year
Data Center Segment Accounts for 92% of Total Revenue
"Surprise Performance Dispels Market Doubts"

Nvidia's revenue and operating profit have more than doubled compared to the same period last year, far surpassing expectations, fueling optimism that the current momentum for infrastructure investment in the artificial intelligence (AI) sector will continue. As major “hyperscaler” data center operators keep investing, Nvidia’s blowout earnings have reinforced the outlook for sustained large-scale investment in the coming months.

[Into the World of AI] Nvidia's Meteoric Growth... Unwavering Infrastructure Investment View original image

Nvidia reported sales of $96.22 billion (about KRW 133 trillion) in the second quarter of fiscal year 2027 (May–July), up 106% year-over-year, marking its 13th consecutive quarter of record sales. The company’s 2028 sales growth guidance is well above market expectations (45%), with actual customer demand surging by more than 100%, but the outlook is considered conservative due to supply chain constraints. Earnings per share (EPS) for the second quarter came in at $2.22, also beating the market estimate of $2.10.


By segment, the data center division—which includes AI chip sales—posted $89 billion in revenue, up 117% from the previous year, contributing 92% of total sales.


Nvidia projected, "Next year’s sales will grow by 70%," marking the first time the company has issued a one-year sales outlook, a move widely interpreted as a sign of confidence. This aggressive forecast is based on expectations that hyperscaler investment will continue to rise sharply.


Nvidia's strong results have dismissed concerns about a potential AI peak-out. Minseong Koh, an analyst at NH Investment & Securities, commented, "The core of Nvidia’s growth lies in the explosive surge in token processing volume due to the spread of agents that require 15 to 100 times more computing than humans," adding, "Thanks to the high compatibility of the Nvidia CUDA platform, the company is absorbing demand across the entire life cycle of both frontier and open-weight model training and inference, and sales expansion of the cutting-edge AI accelerator 'Vera Rubin,' which began shipping this month, is expected to boost future earnings."


Yeongwon Lee, an analyst at Heungkuk Securities, remarked, "These results align with Jensen Huang’s long-term forecast of a $4 trillion investment by 2030 and reaffirm optimism for continued AI investment growth despite many controversies," adding, "This also suggests that the favorable memory market could persist for an extended period, and the resulting strength in memory prices could start to weigh on Nvidia’s gross margin from next quarter onward."


[Into the World of AI] Nvidia's Meteoric Growth... Unwavering Infrastructure Investment View original image

There have been lingering doubts over the sustainability of AI investment, with concerns that hyperscalers would face funding difficulties as their cash positions dwindle. However, Nvidia’s better-than-expected results have dispelled much of this market skepticism.


The IT industry has especially focused on the rapidly rising demand for tokens that fuels AI investment. Since the emergence of agent AI, demand growth has accelerated. According to an OpenAI analysis, the gap in token usage between the top 10% of leading companies and regular companies that have adopted agentic AI ‘Codex’ has widened by as much as 17 times since April 2025. Analyst Lee said, "As AI evolves, token demand is surging, and investment in building the infrastructure to support this continues to grow structurally. Hyperscaler spending is outpacing expectations, and strong growth in memory semiconductor revenue is sustaining confidence in semiconductor demand growth."



Recently, as AI usage shifts from direct prompt input by humans to agent-based operation, it is expected that companies will operate a far greater number of AI agents than human employees. These agents run 24/7 in the background, collaborate with other agents, and dramatically increase computing consumption. Analyst Koh stated, "AI has started to perform productive and useful work; AI services are generating profitable tokens, and adding computing power translates into more tokens and earnings," adding, "The trend of continuous infrastructure investment by AI service providers is likely to persist for the time being."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing