The key theme of this week's stock movements was AI. However, it wasn't a simple AI rally.

The market did not buy every company with "AI" attached to its name. Instead, it began to differentiate between companies where AI translates into actual demand, revenue, and profits, and those facing increased AI investment burdens or concerns about cannibalizing existing businesses.

The first question in this week's stock radar is this:

For whom is AI a benefit, and for whom is it a cost?

This question was clearly evident in the US market.

Stocks linked to power semiconductors and AI infrastructure showed strong momentum, while some software and cloud companies faltered due to concerns about slowing growth. The weakness in ServiceNow and IBM was not merely an issue of individual companies.

The market has begun to re-evaluate whether AI is a force that complements the existing software ecosystem or one that replaces certain areas.

Conversely, the strength in Texas Instruments and Intel offered a different message.

It signifies that strong expectations still surround companies connected to AI's impact on data centers, semiconductors, power management, advanced processes, and industrial demand. The market's interest remains sustained in segments where AI investment translates into actual facility investment and semiconductor demand.

Tesla was also an important subject of observation.

Even if earnings are better than expected, stock prices can fluctuate if large capital expenditures and free cash flow burdens are highlighted. This demonstrates that the market's evaluation criteria for growth stocks are changing.

The market no longer solely looks at growth stories. It now considers the costs associated with growth and how those costs pressure cash flow.

In the Korean market, semiconductors remain the central axis.

The semiconductor trend, represented by Samsung Electronics and SK Hynix, directly benefits from AI demand and the recovery of the memory market. What's crucial isn't just that stock prices have risen significantly, but whether earnings forecasts can continue to be revised upward. For semiconductor stock price increases to be justified, AI demand must translate into actual orders, price increases, and margin improvements.

The next area to watch is the "diffusion stocks." After semiconductors move first, the market may turn its attention to sectors like automobiles, shipbuilding, machinery, transportation, refining, chemicals, and IT services.

However, this diffusion does not happen automatically. Each industry must demonstrate its own basis for profit improvement. A rotation where other sectors rise simply because semiconductors have risen is unlikely to be sustained.

The name for this week's stock movements is "AI Selective Filter."

The fact that AI is the market's central narrative remains unchanged. However, within AI, beneficiaries, cost-burdened stocks, and stocks facing replacement risk have begun to diverge. The earnings season is accelerating this differentiation.

This week, the market is asking stocks this:

"Show us what you can earn with AI, not just that you are related to AI."

Next week's checkpoints are threefold:

First, the earnings and capital expenditure plans of AI infrastructure companies. We need to verify if revenue growth is exceeding cost increases.

Second, whether semiconductor earnings forecasts will be further revised upward. Leading indicators in the Korean market are still semiconductors, and if earnings forecasts stagnate, stock prices may face pressure.

Third, the basis for profit improvement in diffusion stocks. For automobiles, shipbuilding, industrials, and IT services to follow semiconductors, they need a foundation for profit improvement, not just supply and demand dynamics.

※ This article is an explanatory commentary on market trends, not an investment recommendation. Actual stock prices and fund flows may vary depending on exchange rates, oil prices, geopolitical variables, corporate earnings, and policy changes.