Newsletter commentary Feb 2026

Nonlinear Dynamics and Opportunities from Rapid AI Development

At the end of January and beginning of February, the so-called “Wash shock” created short-term volatility in precious metals, some industrial metals, and equities. The market quickly began to anticipate potential constraints, and prices started to stabilize. However, a change in leadership will inevitably bring a shift in the policy path, which will remain an ongoing narrative and periodically drive trading activity.

In the U.S. market, the Nasdaq saw a slight correction of about 2 points, while the S&P 500 remained stable. The previously dominant seven major tech stocks generally retreated, with Microsoft pulling back by over a quarter from its high. A-shares have also entered a more balanced and expansive phase: leading companies traditionally favored by institutions have shown divergence, with half of the top ten most heavily held stocks adjusting by over 10%, and only one rising by more than 10%. Capital is flowing out of crowded sectors in an orderly manner. Meanwhile, pro-cyclical and value sectors such as metals, construction, basic chemicals, petrochemicals, and steel are leading the entire market, as profit recovery and valuation repair resonate. The global market is transitioning from concentrated crowding to broad-based expansion.

After three years of rapid AI development, breakthroughs by Anthropic and Seedance 2.0 in programming and video generation, combined with active promotion from AI companies during the Lunar New Year, demonstrate that AI is accelerating its penetration across industries and households. The investment landscape for AI has evolved significantly. Some early movers are undergoing temporary adjustments, creating potential entry points. For example, Microsoft and several software companies have moderated from previous highs, while TSMC has risen nearly 50% during the same period. Capital is increasingly concentrating on storage, chip manufacturing, and power supply bottlenecks. The KOSPI index, up 48% this year following a 76% rise last year, further highlights strong growth opportunities.

The market is simultaneously assessing the pace at which AI investment returns can be realized and actively positioning in the “essential tools” segment of AI development. Innovation continues at a rapid pace, and while the early advantages of pioneers such as OpenAI, Microsoft, and NVIDIA are increasingly debated, the industry is still evolving at high speed. Despite NVIDIA’s gross margin reaching 80% and CSP ecosystem margins around 50%, compute token prices remain elevated. Storage sector valuations are approaching nearly 10 times historical averages. Companies locking in current prices for production over the next three years are actively expanding capacity, which is commercially rational and could generate substantial profits over the next three years. At the same time, rising storage prices have led to roughly a 20% reduction in production for certain consumer electronics.

If nonlinear dynamics increasingly shape economic activity, traditional long-term DCF models spanning 20 years or more may become less predictive. Emphasizing near-term fundamentals in investment decisions may therefore be more appropriate. Some AI-related power supply stocks in A-shares are priced for growth through 2030 at roughly 20x valuations—whether this represents foresight or resembles prior high-valuation examples (e.g., Haitan Soy Sauce at 80x, Aier Eye Hospital at 125x) will only be validated over time. In such a rapidly evolving environment, both selective participation and disciplined observation can be rewarding, while acknowledging the limits of our knowledge. Even realized gains may reflect survivor bias, and missed opportunities are a natural part of dynamic markets.

AI is beginning to impact virtually every aspect of society and the economy. At the macro level, AI investment is already contributing to short-term GDP growth and may influence employment structures over the medium to long term. Historically, each wave of technological progress has significantly enhanced productivity and ultimately benefited society at large. This current phase of AI, combined with subsequent advancements in robotics, has the potential to substantially boost productivity, with the resulting gains gradually spreading across more companies, individuals, and capital. The expansion in the market capitalization of large enterprises reflects this trend.

Human economic activity relies on interconnected income flows, and if wealth becomes overly concentrated among a few players, this feedback loop could be affected. Historically, the distribution of wages, profits, taxes, and depreciation has remained relatively stable, and significant short-term shifts could have far-reaching consequences. This may represent a transformation on a scale comparable to a millennium, with potential impacts far exceeding century-scale shifts in geographic or national asset allocation. Investment strategies must evolve in response, as entities traditionally responsible for resource allocation will face both new opportunities and challenges.

Rapid development in a single sector can temporarily reprice other industries, but such adjustments create opportunities for disciplined investors to capture relative value. Controllable AI development is challenging; for example, Anthropic declined the U.S. Department of Defense’s request to allow unrestricted weaponized AI, while OpenAI complied, demonstrating that competitive pressures are pushing the industry into new frontiers.

AI’s influence on investment is immediate: the democratization of information—regardless of accuracy—affects stock prices, and the feedback loop from information to price has shortened. The competitive dynamics between humans and between humans and machines are evolving, while humans continue to contribute unique value in deep reasoning and cognitive work.

Adapting to these changes is essential because they represent objective realities. Maintaining an open mindset toward multiple future scenarios is critical. In the short term, market opportunities this year have been driven not only by AI but also by traditional economic cycles. Some sectors have already undergone significant change compared with three to five years ago. Frequent nonlinear events demand greater breadth and speed in investment response. Achieving attractive risk-adjusted returns now requires both wider coverage and efficiency, while long-term forecasting remains challenging—but such an environment also offers substantial structural opportunities.

Rapid development in a single sector can also create temporary pressures for economic structural rebalancing. For example, AI’s fast-paced growth increases productivity within its own sector, altering relative prices in other industries. Extreme fluctuations may temporarily disrupt market signals, making decision-making more challenging for economic participants. For those who believe in AI’s transformative potential, such structural adjustments are a natural consequence. Controlling AI development remains difficult; for instance, Anthropic declined the U.S. Department of Defense’s request for unrestricted weaponized AI development, while OpenAI complied, illustrating that competitive pressures are driving the industry into uncharted territory.

AI’s impact on investment is immediate and direct: the widespread availability of information—regardless of accuracy—directly affects stock prices, and the feedback loop from information to price has shortened considerably. The competitive dynamics between humans, and between humans and machines, are evolving. At the same time, humans continue to provide unique value through deep logical reasoning in cognitive work.

Adapting to these changes is essential because they reflect objective realities. Maintaining an open mindset toward multiple potential future scenarios is critical. In the short term, market opportunities this year have not been driven solely by AI; traditional economic cycle dynamics continue to play an important role, and some familiar sector themes are likely to reassert themselves. Certain industries have already experienced significant transformation compared with three to five years ago. Frequent nonlinear events also demand greater breadth and speed in investment coverage. Achieving an attractive risk-adjusted return in stock selection now requires both wider exposure and higher efficiency. While precise long-term forecasting is becoming more challenging, these dynamics are also creating structural opportunities that merit close attention.