The latest market adjustment phase is approaching its conclusion. The non-tech sectors led the decline in May-June, followed by the tech sector in June-July, creating a more complete correction structure.

Momentum and positive trading effects have contracted sufficiently, indicating conditions are ripe for this downswing to finish. The final stage warrants attention to potential capital diversion effects from the upcoming domestic memory leader's IPO.

Non-tech sectors corrected first in May-June, with tech following in June-July. We have consistently noted that while fundamental divergence exists, the divergence in capital flows is even more pronounced.

The "three-tiered capital divergence" – global funds returning to US equities and exiting other markets, a positive feedback loop in tech drawing funds from broad-based ETFs and traditional heavyweights, and industry ETFs being drained by single-theme tech products – created a lopsided dominance by single-theme tech products prone to a negative second derivative shift.

Now, with leading products restricting large subscriptions and retail investor buying interest waning, a short-term negative second derivative shift in capital flows is clear, disrupting the momentum inertia of the tech rally.

Concurrently, previously oversold rebound sectors have also retreated to relatively low levels, indicating that active capital is beginning to price in a market adjustment. The conditions for concluding the correction are in place: non-tech assets were already at low levels, and with the tech sector's correction, the adjustment structure is now more comprehensive.

Furthermore, the contraction in positive trading effects and momentum is nearing full extent, suggesting we are not far from the bottom of this cycle's adjustment. The final stage also requires monitoring the capital diversion impact from the domestic memory leader's listing.

While unrealized gains in tech holdings have fallen to low levels, there remains potential for adjustments involving capital outflows around the break-even point. If this materializes, it could mark the absolute bottom of this correction.

Pacing Slows as the Next Uptrend Requires Patience

The computing power inflation rally that began in mid-November 2025 has seen its breadth and depth reach considerable levels. In the short term, it is highly likely that the inertia of capital supply and demand has been broken, which could serve as a segmentation point for the tech rally's phase.

The pace of the tech rally is slowing, entering a stage of awaiting significant new industry catalysts. A restart of the rally will most likely mark a new phase, not a simple continuation of the previous stage's narrative.

With capital inertia broken, the tech rally's pace may slow accordingly. During the tech-dominated rally of May-June, "waiting for a new catalyst before deciding" often seemed wrong, as the market advanced along its established path without obvious new catalysts, perpetuating significant divergence.

However, with capital inertia now broken, we believe a new upswing requires a measured approach; now, "waiting for a new catalyst before deciding" is the correct strategy. We discuss three reasons.

First, since 2025, there have been three main inflection points where unrealized gains in the electronics sector turned negative: early April 2025, mid-November 2025, and late March 2026. Although the April '25 and March '26 adjustments had external geopolitical reasons, these points also served as windows for segmenting rally phases.

The initiation of a new upward cycle has consistently been premised on the realization of major industry trend catalysts. Before April '25 was the DeepSeek-driven rally; after was the beta rally from high overseas AI capital expenditure.

After the overseas AI capex beta rally reached a high in September '25, the Google AI supply chain rally began in mid-November, marking the start of the computing power inflation as the main theme.

Following the March '26 US-Iran conflict adjustment, Anthropic's profitable model propelled the computing power inflation rally to a new level. This time, with capital inertia broken, a new industry catalyst is also needed to restart the upward trend.

Second, the depth and breadth of the computing power inflation rally have become relatively saturated, which is another reason the tech rally may need to enter a new phase. As the computing power inflation theme deepens, the difficulty and elasticity of generating new computing power inflation through fresh technological route changes increase, while the probability of new computing power deflation rises.

Recently, cross-sector M&A activity by industrial chain firms has raised concerns about supply shocks. The fermentation of long-term technological route expectations can also cause significant short-term volatility, reflecting underlying fragility.

Third, pronounced structural divergence and a positive feedback loop in thematic capital flows may not align with the need for a stable, long-lasting rally. Restrictions on large subscriptions for tech-themed products to avoid concentrated trading may persist.

After the negative second derivative shift in capital flows, restoring the positive feedback loop requires time and support from new industry catalysts.

Unchanged Mid-Term Structural Outlook

The AI industry trend remains the main battlefield for major market cycles. A complete major cycle is characterized by the tech sector continuing to lead gains while a broader array of structures can also rise, creating a more diversified market.

For this diversification, brokerages are the preferred choice. Concurrently, within strategic resources, industrial metals and basic chemicals are favored over precious metals, while export/overseas expansion chain alpha and new consumption themes warrant attention.

The mid-term structural outlook remains unchanged, with the dominant style likely to persist throughout the major cycle. In the next phase, the AI industry trend will remain central, and computing power inflation – segments with rising utilization and supply bottlenecks supporting profits through volume and price increases, coupled with a rising valuation multiple – remains the typical model for finding high-elasticity investment cases.

A tech-led rally, accompanied by a richer set of fundamental improvement sources and a more diverse structure of incremental capital inflows enabling a broader array of sectors to rise, is the hallmark of a complete major cycle.

For this diversified rise, brokerages are the top pick. Non-bank financials are the only primary industry with "low PB, high ROE"; key institutional holdings have been sufficiently cleared out; profits are cyclical with the market, with venture capital and overseas expansion providing alpha.

Short-term volatility in brokerages mainly stems from the adjustment phase, where opinions are divided. Once the next upward phase restarts, the brokerage rally is highly likely to resume. The depth of the major cycle will determine the scale of the brokerage rally.

Additionally, in the short term, as the rise in US Treasury yields shifts from high inflation expectation drivers to real interest rate and term spread drivers, cyclical assets like industrial metals and basic chemicals may outperform precious metals during this stage.

Simultaneously, monitor mid-term opportunities in export/overseas expansion chain alpha and new consumption. Conditions for these sectors to deliver outperformance: the export/overseas chain alpha requires the short-term disturbance from the US-Iran conflict to pass and potential new trade friction boundaries to become clear, coupled with a reinforced recognition of the mid-term fundamental uptrend.

New consumption awaits the restart of consumer IPOs and the approval of new consumer-focused fund licenses.

Risk warnings include overseas economic recession exceeding expectations and domestic economic recovery falling short of expectations.