AI Agent Proliferation Expected to Boost Semiconductor Inference Demand
If AI agents become commonplace, inference demand will explode, and improving these inference capabilities will require not only GPUs but also a significant role for CPUs.
The everyday use of AI agents is increasing semiconductor demand, with the role of CPUs becoming especially critical. Investors need to manage risk by adhering to principles of staggered buying and stop-loss orders.
If AI agents become commonplace, inference demand will explode, and improving these inference capabilities will require not only GPUs but also a significant role for CPUs.
Buying all at once during morning rallies is an incorrect allocation method; investors should plan their investments weekly and deploy cash in stages.
It is crucial to secure cash reserves for buying during afternoon corrections to respond flexibly to market volatility.
Staggered buying should be initiated during corrections or when the rate of ascent slows down.
This effect is expected to conclude around the first week of October.
However, this share buyback effect is scheduled to conclude by the first week of October, so attention should be paid to changes in supply and demand thereafter.
Individual investors are contributing to selling pressure by taking profits at their breakeven point when stock prices rise.
Unlike the relatively robust supply and demand structure of the US stock market, the South Korean stock market has an overwhelmingly high proportion of Samsung Electronics and SK Hynix's market capitalization.
Investing 20-30% of assets in a specific sector can be a very risky investment strategy, even if that sector only accounts for 1-2% of the total market capitalization.
To invest in materials, components, and equipment (SOBUJANG) companies, in-depth study of the specific stocks is essential.
It is important to understand the competitive landscape and technological capabilities of related companies, rather than simply investing because a particular stock has surged.
For example, knowing a company like SF Semiconductor and its competing groups like Hana Micron can increase investment success probability to 70%, and even up to 90% with deeper understanding.
Simply investing in recommended stocks is not advisable in the long term.
One of the essential conditions for stock investment is the strict execution of stop-loss principles.
It is crucial to mechanically sell when losses reach -15% to 20%.
Failing to stop-loss and letting losses accumulate to -30% to -40% can lead to personal distress and the loss of other investment opportunities.
By selling according to a predefined principle, further loss expansion can be prevented, and rational investment decisions can be made.
Recently, as HBM-related issues have been re-highlighted, this valuation gap is narrowing.
It is confirmed that Samsung Electronics was undervalued when considering market dominance and projected operating profit growth rates.
Recently, as HBM (High Bandwidth Memory)-related issues have been re-highlighted, the valuation gap between Samsung Electronics and SK Hynix is gradually narrowing.
This growth rate is over 50% higher than ChatGPT's 1.3 million downloads in its initial 10 days after launch.
Meta AI ensures service reliability by implementing a 'human loop' system that requires final user confirmation before payment.
The advanced inference capabilities of AI agents demand intensive data management from CPUs, leading to increased CPU demand.
Along with increased CPU demand, improvements in the speed and size of substrates (PCBs) for data processing are essential.
Substrate-related stocks such as Simmtech, TLB, Daeduck Electronics, Samsung Electro-Mechanics, and Korea Circuit are gaining attention due to new demand growth in the AI era.
The key issue at the US-China summit was not tariffs, but rather whether the US would maintain its semiconductor and secondary battery sanctions against China.
It is realistically difficult for the US to easily ease semiconductor sanctions against China.
Geopolitical variables such as Iran risk may have a greater impact on global stock markets than the US-China issue itself.
Major AI companies like OpenAI are increasingly contracting with consulting firms for external diagnostics.
As geopolitical risks increasingly impact the stock market, software consulting firms like Accenture are seeing a rebound driven by demand for AI safety.
This demonstrates the establishment of new consulting areas suited for the AI era, going beyond fears of a 'SaaS apocalypse'.
The general trend of AI investment involves software and intangible asset investments following hardware investments.
It is estimated that it will take about 2.5 years for companies to verify productivity metrics through AI and subsequently reinvest based on them.
Therefore, AI's contribution to productivity may be difficult to measure clearly in financial figures until 2027, with full-scale reinvestment due to productivity improvements expected from 2028.
Changes in production strategy are also emerging, such as reducing HBM stack height to increase overall supply for GPU procurement.
A reversal has occurred where the margin of general DRAM is becoming higher than HBM, impacting production strategies.
The strategy of reducing HBM stack height to increase overall supply for GPU procurement may be revised.
Reduced general DRAM supply due to production mix adjustments is expected to further intensify market supply-demand imbalances.
A return to past levels of 1.3 million won for Samsung Electronics and 220,000 won for SK Hynix is unlikely. Analysis suggests that a sharp decline in stock prices will not be repeated even if the US raises interest rates four times.
While an adjustment to around 1.6 million won is possible, the probability of a much larger decline and return to past levels is slim, and investors should scrutinize the underlying factors of stock prices.
Investment analysis reports often contain condensed information, necessitating an explanatory guide, and investors need to strive for a deep understanding.
Utilizing AI tools like ChatGPT to delve into report content repeatedly for learning is effective.
Rather than being fixated solely on superficial market figures, a culture of questioning and exploring the underlying causes must take root to elevate investors' analysis levels.
No major changes to AI-related large fundamental factors are expected during the Chuseok holiday period.
If the market corrects due to macro issues like the Iran-Israel war, a strategy of using this as a buying opportunity is necessary.
Consider rebalancing portfolios during the holiday period by liquidating weaker holdings and concentrating on stocks with growth potential.
Answers come from the transcript, with the exact spot cited.
Want the next article from 머니인사이드?
When 머니인사이드 publishes, we'll write it up like the one you just read and email it to you.
Exports Nearing $1 Trillion…Driven by K-Beauty and AI Semiconductors머니인사이드last week · 32:40 · 465.8K views · Created 2 days ago
Yield Surge Drives Sector RotationStockCharts TV37 minutes ago · 17:12 · 104 views · Created 18 minutes ago
Crew-13 Cleared for Launch to ISSNASA1 hour ago · 56:30 · 5.5K views · Created 1 hour ago