Arnaud Yao of Amundi Asset Management: The stock market’s most core logic lies in the “mid-level”—AI’s first half drives price rises, while its second half drives price declines.

On July 9, Yao Yuan, Senior Investment Strategy Officer for Asia at Easton & Merrill Asset Management’s Investment Research Institute, shared Easton & Merrill Asset Management’s views on global macro trends and the capital markets for the second half of this year.

For the stock market, he believes the most core logic is actually the meso level—specifically, views on AI. If the AI trend or style rotates, the stock market will definitely be affected.

But AI is also an indispensable part of an investment portfolio, and the long-term outlook remains bullish. In the short term, investors need to stay prudent amid a series of changes across valuation, fundamentals, and technical factors. For assets that were previously more concentrated, hotter, and had heavier positioning, pay attention to position management and hedging.

As for gold, the allocation to gold in the average global investment portfolio is still below 3%. There should be room for both further allocation and upside. In the medium to long term, gold remains favored. However, in the short term, stay prudent, because transaction-style or cycle-driven factors are currently dominating the market.

AI is key

One important viewpoint Yao Yuan has is that in the second half of the year, the biggest key for the stock market is still AI. If AI’s style rotates out, that will definitely be the biggest challenge for the stock market.

It can now be clearly seen that the stock market’s most core logic is actually industry logic at the meso level, not macro logic. The most critical factor is the trend of the AI industry. If the AI trend or style changes, the stock market will definitely be most affected.

The fixed-income and FX markets are still trading under the logic of the real economy and have not seen particularly obvious disruptions from AI. The main drivers remain the real economy and the central bank’s monetary policy.

In the global equity markets, Yao Yuan is more bullish on emerging market stocks. He believes U.S. stocks are currently valued relatively high, so they are still neutral for now. Even if participation is needed, he suggests choosing equal-weight index exposure rather than market-cap weighted indexes, because market-cap weighted indexes have higher concentration.

For emerging markets, he is generally bullish on both stocks and bonds. The equity market fundamentals are also relatively solid, and valuations are somewhat more reasonable compared with mature markets. For bonds, whether they are hard-currency bonds or local-currency bonds, with geopolitical risk fading and energy prices falling, they will benefit in the middle.

The AI first half boosts prices, the second half boosts price declines

There are two prevailing views in the capital market. The first is that AI brings disinflation, because by improving future productivity, AI will ultimately lead to lower prices. The second is that AI brings inflation, which matches the data outcomes they observed over the past period: the products and services touched by AI are, in fact, seeing price increases.

Yao Yuan believes both viewpoints may be correct, but they differ in timing.

Disinflation is a phenomenon in the second half of AI development. If AI can truly lead to a major surge in productivity and a spike in supply, prices will certainly fall in the future.

But we are currently in the first half of AI development. Before productivity can be improved, the overall infrastructure must first be built. The demand for infrastructure is huge, and supply is the bottleneck. Therefore, prices for AI-related services and goods have surged.

Stay prudent on short-term trading

As for how to lay out AI investments, his overall view is that everyone should separate long-term allocation logic from short-term trading logic.

On the long-term allocation front, there is no change in the big trend. AI is very likely the most disruptive technology of this era—it will change the world, so you cannot stay on the sidelines; you need to get involved. AI is an indispensable part of an investment portfolio, and the long-term outlook remains bullish.

But in the short term, the trading logic really does require a cautious attitude. Not only valuations, market concentration, liquidity, and a series of technical indicators—after the rallies over the past four or five years, many indicators have already shown “overbought” and even crowded conditions.

On fundamentals, there is also a very clear problem this year: the free cash flow of large cloud service providers has fallen dramatically, and the cash flows of many cloud companies are now at negative levels. If internal funds are tight, it means future investment will increasingly depend on financing from the capital market. That means either everyone crowds into initial public offerings, or large-scale bond issuance.

This will lead to two outcomes: first, it drains market liquidity; second, capital expenditures become increasingly influenced by changes in the capital markets. If financing relies on equity, it is affected by sentiment in the stock market; if financing relies on bonds, it is affected by rising financing costs.

So, capital expenditures are currently the most important source of today’s AI boom. If the water source feeding the system is constrained, the downstream will certainly be affected by a series of domino effects. While the medium to long term is still favorable, in the short term, investors need to stay prudent given the various changes in valuation, fundamentals, and technical factors.

AI industry: China and the U.S. each have advantages

Now in the entire AI race, China and the U.S. are the two main competitors. Both China and the U.S. have their own advantages. China’s advantage is that the model development approach is open-source. Historically, open-source technologies usually outperform closed-source technologies, when considered from the perspective of market share share.

Second, China has an energy advantage, which allows token costs to be kept very low.

Third, because China is open-source, anyone can use it, and the underlying open-source code is visible to all. So if you obtain AI and want to do some customized adaptation—such as using AI for quantitative investing, or using AI for industrial optimization—the code can be rewritten by yourself.

If you purchase so-called closed-source products from OpenAI, it is a more standardized offering that is used for upper-layer optimization. So at the level of downstream diffusion and spread, China has advantages in certain areas.

Moreover, we have not fully entered the downstream application and diffusion stage yet. If we truly move into the downstream application and diffusion stage, the entire AI ecosystem would expand greatly. The AI being discussed now is still large language models, computing power, cloud, and storage.

But if AI truly becomes a technology that affects all aspects of the real economy and society—just like the internet—then any company that can leverage this technology well would effectively become an important AI investment target.

H-shares may see catch-up driven by a rebound

Regarding the performance of A-shares and H-shares, Yao Yuan believes the first half was affected by the energy crisis and changes in AI style.

From 2025 to the first half of 2026, AI style also has a very clear shift. In 2025, the U.S. “Magnificent Seven” generally performed well overall, with both software/hardware and large models doing well, so the “Magnificent Seven” performed strongly and H-share large-cap constituents also did well.

But from the end of 2025 to the first half of 2026—roughly seven or eight months—what we saw is that the most popular AI style was actually hardware: semiconductors and storage chips. These are not as prevalent in H-shares. By contrast, A-shares have a wider selection of hardware-related targets, so A-shares have outperformed H-shares this year.

However, in the CSI 300 and the SSE Composite Index, these constituents are not especially high in weight either. For example, in the STAR Market index or the ChiNext index, the weight of these constituents would be higher. So if you buy the ChiNext, you can expect better returns this year; but if you buy the main-board market, the overall performance is indeed somewhat unsatisfactory. The market’s constituent structure really does constrain the performance of these markets.

Looking ahead, will there be a style reversal in the second half of this year? The first is that energy crisis risk has basically been resolved. If there is no further large escalation beyond expectations, and energy prices remain at $70 to $80, it will shift from being a headwind to a tailwind.

Second is an AI style shift. In fact, signs of some style rotation have already been apparent in the past few weeks. Korean stocks saw a clear pullback; the U.S. “Magnificent Seven” itself is now showing signs of stabilizing; and Hong Kong tech indices have also rebounded clearly over the past one to two weeks.

So it is indeed possible that there will be style rotation, which could drive a catch-up rally for markets that previously underperformed.

Hong Kong stocks: neutral to mildly bullish

Their current rating for the entire China equity market is neutral. But this “neutral” includes both over-allocation to new quality productive forces and under-allocation to the old economy, such as consumption and real estate—it's the combined result of the two.

In the second half of 2026, H-shares have potential to outperform A-shares. The most important, core variable—besides global market impacts—is still micro fundamentals.

In the first half of 2026, Hong Kong stocks—especially heavyweight stocks and internet platform stocks—underperformed mainly because their earnings were significantly lower than expectations.

If in the second half of 2026, on earnings—after the effects of the earlier takeout food price war fade and a bottoming rebound appears, plus policy efforts later in the year that spur a recovery in domestic economic momentum—then if their earnings show signs of a bottoming rebound, it is reasonable to expect a round of catch-up for Hong Kong stocks.

So, for Hong Kong stocks, his view is neutral to mildly bullish.

Still bullish on gold in the medium to long term

For gold, when looking over a longer time horizon, gold has delivered three waves of bull markets over the past 60 years: first in the 1970s, second from 2000 to 2011, and third from 2018 to the present. These three waves correspond to complex geopolitical tensions, macroeconomic conditions, stagflation, and highly uncertain policy environments.

“Buy gold in chaotic times” is a market lesson validated by history. As an allocation-type asset, it should be allocated. Right now, the share of gold in the average global investment portfolio is still below 3%. Ray Dalio and others have urged increasing it to 5%, even 10% or higher. So there should be room for future allocation and upside. In the medium to long term, gold remains favored.

Gold faces pressure in the short term because there are many speculative and leveraged funds in the market that, in reality, have no strong “belief”—they’re just trading for a quick profit. If things turn bad, they quickly flee. That’s why the prior rally was so fast, and the correction was also very sharp, analyzed from a market perspective.

From a fundamentals perspective, when gold has moved on a long-term trend historically, it has both macro long-term trend support and short-term cyclical drivers.

For the big trend, concerns about geopolitics, inflation, and the fiat currency credit system lead to higher gold prices over the long run.

When long-term narratives are not the dominant force in the market, short-term cyclical drivers dominate instead.

Since 2024, gold has risen by 30%; in 2025, gold rose by 60%; and by the first three months of early 2026, gold rose another 30%. Over more than two years, it has mainly been long-term grand trends pushing gold prices higher.

But from the adjustment in March to now—over the past more than three months—short-term factors have again started to influence gold. A stronger dollar begins to weigh on gold, and rising U.S. real interest rates also start to weigh on gold. So gold has been switching back and forth between the long-term grand narrative and short-term trading logic.

As for when gold will jump back from the current short-term logic to the long-term logic, no one has a “crystal ball” to judge that. Therefore, still separate long-term allocation logic from short-term trading logic: hold long-term, and stay prudent in the short term, because transaction-driven or cycle-driven factors are currently dominating the market.

Their expectation for gold over the next 12 months is that it will reach $5,500 per ounce in 12 months, so it is still at a relatively high level, which also reflects the medium-to-long-term bullish view on gold.

Risk disclosure and disclaimer

        The market involves risk; investment is需谨慎. This article does not constitute personal investment advice, and it does not take into account any particular users’ specific investment objectives, financial situations, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article are consistent with their own specific circumstances. Invest at your own risk based on this.