Newsletter commentary June 2026

In June, the market continued the sharp divergence seen in May. Stocks in the electronics and communications sectors, benefiting from the AI value chain, continued to rally strongly. Certain materials companies also benefited. Outside these areas, however, most sectors performed poorly, with many stocks suffering significant declines. The degree of market dispersion has become extreme and relatively rare.

For the first half of the year, among the Shenwan industry classification, the electronics sector rose by nearly 90%, while the commercial trade and retail sector declined by nearly 30%. The median A-share stock fell by nearly 16%, while the CSI 500 Index rose by approximately 21% and the CSI 300 Index gained around 8%. AI has effectively become the only dominant market theme, while most other sectors have acted as drags. Our portfolio remained relatively balanced and was under-invested in AI-related opportunities, which weighed on performance.

Objectively speaking, without the prosperity of the AI industry, the global economy would appear rather lacklustre and challenged. The operating model of the global economy over the past several decades is facing increasing pressure. Many countries carry high debt burdens, inflation has remained persistently elevated, and ordinary households have not felt meaningful improvement in living standards. Traditional sectors lack clear growth drivers, while many governments have limited policy tools to change the current situation. In reality, many governments either have few effective tools left, or can do very little.

Against this backdrop, only selected stocks in regions strongly linked to AI investment — including the US, China, Japan and Korea — have performed well. The extent of AI exposure has largely determined relative equity market performance.

Even the reopening of the Strait of Hormuz only created short-term volatility in related stocks, while oil prices quickly fell back toward pre-conflict levels. This also reflects the broader state of the global economy: demand remains weak, oil’s weight in the consumer basket is no longer comparable to that of the 1970s, and OPEC’s ability to control supply has weakened significantly. In addition, more oil supply is likely to enter the formal market over time. The rebalancing of crude oil, the most important bulk raw material, has already begun, and the medium-term equilibrium price may be lower than previously expected.

Another major variable is the potential change at the Federal Reserve. Viewed over several decades, the Fed may be standing at the beginning of a new phase. From Greenspan to Bernanke, Yellen and Powell, and potentially to Warsh, the size of the Fed’s balance sheet has expanded from around USD 900 billion to USD 6.9 trillion, while communication itself has become a policy tool.

Greenspan’s philosophy of “constructive ambiguity” marked the beginning of this evolution. Bernanke introduced and institutionalised forward guidance, elevating communication from a supporting tool to a core policy instrument. During the Yellen and Powell eras, the dot plot and forward guidance evolved into policy anchors heavily relied upon by the market. As a result, market research increasingly shifted away from fundamentals themselves toward interpreting the Fed’s intentions.

If Warsh were to take over, the starting point would be very different. Today, policy rates are around 4–5%, whereas Bernanke began from a zero-rate environment, when traditional tools were exhausted and communication had to compensate. That era may now be over. Warsh appears inclined to reorganise the Fed’s internal operating framework, including various working groups and younger reform-minded staff. If these initiatives prove effective, the impact could be significant and may represent a correction of the past two decades.

As the financial environment changes, the Fed’s objectives and tools may also evolve. For example, AI investment may become a new form of dollar recycling, potentially replacing the traditional mechanism of US trade deficits followed by capital inflows into US financial assets. This could have important implications for the pricing of many asset classes.

The first question for investors is: where are we in the AI cycle? There is little doubt that AI’s transformation of the world has only just begun. However, AI-related stocks may not move in perfect synchronisation with the actual development of the industry.

In the first half of the year, token usage experienced both volume growth and price increases. Price hikes and the shift from subscriptions toward usage-based pricing were important drivers of AI investment sentiment. However, more market participants are now beginning to question whether tokens are becoming expensive.

Unlike the internet industry, the marginal cost of token usage is not zero. AI therefore exhibits certain characteristics closer to manufacturing, which may make durable monopoly positions harder to sustain. In addition, model distillation is difficult to prevent, talent mobility remains high, and it may be challenging for any single model to maintain a long-term lead. If state-of-the-art models become too powerful, their use may also face restrictions. At the same time, constant catch-up from competitors means concerns over model commoditisation remain unresolved.

Trading behaviour has also changed. With information dissemination becoming more equalised, markets have shifted from layered trading behaviour to collective crowding, accelerating stock price movements.

Broadly speaking, AI development is moving from an arms race toward broader adoption; from spending regardless of cost toward a greater emphasis on cost-effectiveness. The industry is exploring various solutions. However, good products and good stock prices are not always the same thing. Margins across parts of the AI value chain have reached unprecedented levels. Whether such extraordinary profitability can be assigned long-term valuation multiples requires sustained sentiment and confidence. Recently, investors have begun to debate this issue more seriously.

Market expectations will also increasingly be constrained by the real world. Delays in any one part of the value chain may drag on the entire industry. So far, these constraints have mainly appeared on the supply-chain side. Going forward, pressures from production relationships and broader economic arrangements may also begin to emerge. Recently, more examples similar to Apple’s product price increases have appeared, and some non-AI semiconductor products have also started to rise in price. This diffusion may create cost pressure for downstream terminal products.

Financial constraints are another issue. As massive AI investment accumulates, and assuming depreciation periods of around six years, the required revenue growth will become increasingly demanding. Current AI revenue remains insufficient relative to the scale of investment. For AI revenue to grow meaningfully, it must be embedded across industries and aligned with existing pricing systems. This will require further technological progress or lower margins to expand the addressable market. The high margins seen during the arms-race phase may face new variables.

From an investment perspective, current bottlenecks in the AI supply chain are likely to trigger large-scale capacity expansion, and beneficiaries of this expansion cycle may attract increasing market attention. In addition, China’s AI development path may prove more sustainable, given its lower-cost approach and open-source ecosystem that benefits broader adoption. If so, domestic equipment and localisation opportunities may become important investment themes. China may also benefit from the spillover effects of overseas capacity expansion, while the replication of domestic AI solutions could create further opportunities.

At the same time, as AI has increasingly concentrated market attention and liquidity, other fundamentally sound companies have been crowded out. Looking ahead, with potential policy fine-tuning and a possible improvement in the broader economy, these previously overlooked companies that continue to grow may also present rebalancing opportunities.