Kospi forecasts range from 6,300 to 11,000 as uncertainty clouds year-end outlook

South Korea's stock market has entered a phase where direction is difficult to gauge. The Kospi has been pushed down to the 6,000 level, but the weight of opinion leans toward a rebound rather than a sustained downtrend.

However, uncertainty has grown over whether AI investment will continue and over variables including interest rates and oil prices, widening the range of year-end Kospi forecasts from as low as 6,300 to as high as 11,000.

According to a survey of heads of major domestic research centers, Shin Jung-ho, head of research at LS Securities, put his year-end Kospi band at 6,300–8,500. He identified the imbalance between investment and monetization within the AI ecosystem and geopolitical risks as causes of the recent correction, but said "expectations for the AI ecosystem still remain," and did not view the decline as a structural downtrend.

Cho Su-hong, head of research at NH Investment & Securities, also assessed the current range as one where a sharp rebound is possible in the short term even if further steep losses occur, but offered a more optimistic view by keeping his year-end upper target open to 11,000. He predicted that once big tech earnings and the continuity of AI investment are confirmed in late July, the market will refocus on fundamentals.

The two analysts largely agreed on the causes of the recent sharp decline. Shin pointed to the profit imbalance between AI investors and chipmakers, the shift in market attention from AI investment to monetization, and geopolitical risks.

Cho similarly analyzed that semiconductor peak-out concerns, the US-Iran conflict, domestic supply-demand disruptions and fears of rising US interest rates all played a role, saying "it was supply and demand, rather than any clear negative catalyst, that shook the market."

The two diverged on which sectors would lead a rebound. Shin said Samsung Electronics and SK Hynix could serve as the starting point, but predicted that buying interest would subsequently spread to sectors that had been overlooked — including defense, power machinery, materials and components, secondary batteries and cosmetics — in a broader "normalization."

Cho, by contrast, said concentration in sectors with confirmed earnings, such as AI infrastructure and semiconductors, was likely to persist for now. He also flagged energy, securities and premium consumer stocks as sectors worth watching by the same token.

On valuation, the two were broadly aligned. Shin described the current Kospi as historically undervalued, while Cho said that given the structural growth in corporate earnings, the 6,000 level is in effect a "rock bottom."

Both also leaned toward gradual stabilization rather than further foreign investor outflows, noting that foreign investors had already made considerable progress rebalancing their semiconductor holdings.

Debate continues in the market. The confirmation that AI investment momentum remains is clearly a positive, but cost pressures from rising oil prices and higher interest rates persist. There is also a risk that if funding costs rise while hyperscalers' cash reserves thin out, the enormous capital outlays could turn into a liability.

The concern is that if free cash flow at hyperscalers slows while borrowing costs rise further, the burden of massive AI infrastructure investment could grow significantly heavier.

In a recent report, KB Securities noted that as the US 10-year government bond yield has climbed back toward 4.7%, the investment thesis for AI-related stocks is shifting from "growth" to "cost of capital."

Alphabet (Google) reported cloud earnings that beat market expectations, yet the market reacted more sharply to the expanded capital expenditure guidance than to the earnings themselves. This suggests that the central concern is no longer AI infrastructure investment itself, but how long it can be sustained.

Kim Il-hyeok, a researcher at KB Securities, said "rising interest rates are a burden on industries and the broader market," explaining that "AI infrastructure investment is being viewed through the lens of capital cost burden rather than growth."

He added: "This shift in perspective was already confirmed when Amazon issued $25 billion in corporate bonds on July 7 and demand fell short of expectations." "From a tactical standpoint, it is time to trim exposure to growth sectors and increase allocations to defensive and financial sectors," he added.

th5@heraldcorp.com